Predicting Solar Power Production:

$1,995 Non-Member Price Free for members only Predicting Solar Power Production: Irradiance Forecasting Models, Applications and Future Prospects Steven Letendre, PhD Miriam Makhyoun Mike Taylor Green Mountain College Vermont Solar Electric Power Association Washington, DC March, 2014

Acknowledgments The authors would like to thank the following individuals for taking the time to share their expertise and insights on solar forecasting: Jon Black, ISO New England; John Tedesco, Green Mountain Power; Steve Steffel, Pepco Holdings; Joshua Stein, Sandia National Laboratory; Richard Perez, SUNY Albany Atmospheric Sciences Research Center; Drake Bartlett, Xcel Energy; Jim Blatchford, CAISO; Dallas Cornier, SDG&E; Jack Peterson, Southern California Edison; Jan Kleissl, University of CA San Diego; Sue Ellen Haupt, NCAR; Manajit Sengupta, NREL; Tom Hoff, Clean Power Research; David Edelson, New York ISO; Marc Keyser, MISO; William Holmgren, University of Arizona; Mark Ahlstrom, WindLogics Inc.; and John Zack, AWS Truepower. While we appreciate the many insights provided to us, any errors or omissions in the final report are the sole responsibility of the authors. Front cover photo of control room has been provided courtesy of ISO New England. 2 Glossary of Terms Area Solar Forecast: A forecast of two or more solar energy plants. DG Solar: Solar energy systems deployed within utility distribution networks. Direct normal irradiance (DNI): The amount of solar radiation received per unit area by a surface that is always held perpendicular (or normal) to the rays that come in a straight line from the direction of the sun at its current position in the sky. Diffuse Horizontal Radiation (DHI): The amount of radiation received per unit area by a surface (not subject to any shade or shadow) that does not arrive on a direct path from the sun, but has been scattered by molecules and particles in the atmosphere and comes equally from all directions. Global Horizontal Irradiance (GHI): The sum of DNI and DHI. Global NWP: Predictions of global weather patterns. Insolation: Short for incident or incoming solar radiation, which is a measure of solar radiation energy received on a given surface area and recorded during a given time. Mesoscale NWP: Predictions of regional weather patterns. Nowcast: A detailed description of the current weather along with forecasts up to 30 minutes. Numerical Weather Prediction (NWP): The use of mathematical models of the atmosphere and oceans to predict the weather based on current weather conditions. Persistence Forecast: Forecasts based on extrapolating current conditions into the future. Point Solar Forecast: A solar forecast for a single solar energy plant. Variable Generation: Sources of power that vary based on weather conditions, including wind and solar. solar electric power association

Contents Contents Acknowledgments Contents List of Figures...ii List of Tables...ii Executive Summary...iii 1. Introduction... 1 2. The Solar Resource... 3 2.1 Fundamentals of solar irradiance... 3 2.2 The variable nature of solar energy... 4 3. Forecasting Solar Irradiance... 7 3.1 Physical methods to forecast solar irradiance... 8 3.1.1 Numerical weather prediction (NWP)... 8 3.1.2 Cloud imagery... 9 3.2 Statistical/hybrid methods to forecast solar irradiance...10 3.3 Understanding solar forecast errors and variability...11 3.3.1 Point versus area solar energy plant predictions and PV power output variability...14 4. Using Solar Power Production Predictions for Utility Planning and Operations...17 4.1 Transmission-level applications of solar energy predictions...17 4.1.1 CAISO s approach to using solar power predictions...21 4.2 Distribution-level applications of solar energy predictions...24 4.2.1 Case Study: Hawaiian Electric Company...25 4.2.2 Case Study: Tucson Electric Power...27 i 5. Solar Forecast Provider Survey Results...29 6. Future Developments and Recommendations...33 6.1 Current solar forecasting research projects...33 6.2 Conclusions and Recommendations...35 Works Cited...38 i Predicting Solar Power Production

List of Figures ii Figure 1: Modern Solar Forecasting System Schematic...iv Figure 2: National Solar Radiation Data Base (NSRDB) stations by class... 4 Figure 3: U.S. solar resource map... 4 Figure 4: Time scale relevant to operating power systems... 5 Figure 5: Illustration of solar plant power prediction system... 7 Figure 6: Total sky imager image...10 Figure 7: Satellite cloud images...10 Figure 8: Annual Root Mean Squared Error (RMSE) as a function of forecast time horizon...12 Figure 9: Surface Radiation Network (SURFRAD) Stations Locations...12 Figure 10: Quantifying Solar Variability...15 Figure 11: Forecast Solar vs. Actual Output, Northern California...22 Figure 12: Forecast Solar vs. Actual Output, Southern California...22 Figure 13: Forecast Solar vs. Actual Output, Northern California...22 Figure 14: Forecast Solar vs. Actual Output, Southern California...22 Figure 15: California Independent System Operator Day-Ahead Market Timeline...23 Figure 16: California Independent System Operator Hour-Ahead Market Timeline...23 Figure 17: California Independent System Operator Net Load 2012-2020...23 Figure 18: Distribution Impacts of Distributed Generation Solar Hawaiian Electric Company...26 Figure 19: Impact on System Load for Hawaiian Island...26 Figure 20: Prototype of Variable Generation Display for System Operators...26 Figure 21: Prototype Web Interface for Variable Generation Tucson Electric Power Service Territory...27 Figure 22: Spatial Resolution Capability of Solar Forecast Services...30 List of Tables ii Table 1: Table 2: Table 3: Table 4: Table 5: Table 6: Table 7: SEPA Recommendations to Advance Solar Forecast Market Development...iii Annual Error Metrics Day-Ahead and Hour-Ahead Forecasts for Surface Radiation Network (SURFRAD) Locations...13 Event Example Up-Ramp or Down-Ramp of 20% Capacity in 1 Hour...13 Solar Forecast End-Users and Potential Applications...18 Solar Forecasting Service Providers Invited to Complete Online Survey...29 Number of Forecast Providers Active in Each Area...30 Solar Forecasting Providers Information Needs...31 solar electric power association

Executive Summary Executive Summary Solar energy represents a vast, renewable resource that can be tapped to meet society s growing demand for electrical energy. Deployment of systems that convert solar irradiance into useful electrical energy has accelerated in the past decade in the U.S. and globally, particularly for photovoltaic (PV) systems the direct conversion of irradiance to electricity. Utility companies play an integral part in managing this growth and integrating solar generating plants into the existing grid infrastructure and utility operations. Grid integration studies of variable generation resources, including wind and solar, conclude that it is technically feasible for these resources to provide a significant portion of the nation s energy needs at manageable costs. These studies recommend operational strategies and market structure changes needed to address increased levels of uncertainty that high-penetration of renewable resources presents to utility companies and grid operators due to their intermittent nature. The use of advanced forecasting of variable generation is one of these essential strategies. Forecasts of future solar energy system output allows grid operators and utilities to proactively manage variable output, and thus integrating solar resources into the existing grid at lower costs to society. This report provides a review of solar forecasting approaches and how they are being used by grid operators, utility companies, and other market Table 1: SEPA Recommendations to Advance Solar Forecast Market Development I. Development of Forecasting Standards/ Guidelines II. Economic Assessment of Solar Forecasting Value Relative to Costs III. Distributed Solar Working Group and Workshops IV. Expanded Engagement of Policymakers and Regulators The solar forecasting industry lacks a set of standards or industry guidelines, which acts as a barrier to greater use and understanding of the value of solar forecasting to utility planning and operations. Guidelines or standards with regard to defining forecast time horizons relevant to utility and market operations, measuring and reporting forecast error metrics, and data and communication requirements for modern solar forecasting systems are needed. There is a lack of information within the literature on the economic value that solar forecasts could provide to grid operators and utilities. While the value of solar forecasting qualitatively seems quite large relative to the costs associated with producing solar forecasts, there are currently no quantitative analyses to support this notion. Solar forecasts provide value to numerous stakeholders; an attempt should be made to quantify these values to provide the economic basis for expanded use. Furthermore, these analyses could help to identify the incremental value associated with improved forecast accuracy. While there are various groups working on different aspects of solar deployment and variable generation forecasting, there is a need for a focused effort on DG solar. A working group comprised of relevant stakeholders should be convened to identify the critical issues and begin a collaborative process to address the unique challenges of DG solar. Periodic workshops should be held to share research and current practices used to understand and manage the impacts of DG solar and the role of solar forecasting. Policymakers and regulators need to gain a greater understanding of the value that solar forecasting can bring to meeting policy objectives, including renewable portfolio standards. As grid-connected solar deployment expands and forecasting needs increase, utilities, regulators and forecast service providers need to investigate the best mechanisms in terms of reporting system characteristics providing the needed data to produce accurate solar forecasts, standards in terms of allocating the costs incurred by grid operators and utilities associated with solar forecasting, and promoting market structures and scheduling timelines that leverage modern solar forecasting services. iii iii Predicting Solar Power Production

iv iv participants for planning and operations. This report is based on a survey of the literature and interviews with experts from three broad stakeholder groups (ISO/RTO managers, utility managers, and research scientists). Seventeen companies globally were identified that potentially provide solar forecasting services and asked to complete an online questionnaire. Thirteen companies completed the questionnaire, all of whom work with utilities or related organizations. These companies use state of the art solar forecasting systems to provide customized forecasts to meet their clients particular needs. The research and interviews conducted for this report suggest several initiatives that should be undertaken to accelerate the adoption of solar forecasting in response to the anticipated growth in solar energy system deployment in the coming decade. These initiatives can be grouped into the following four broad categories presented in Table 1. The report begins with a review of the variable nature of solar and is then followed by an examination of the various ground-level solar irradiance forecasting approaches and models used for each solar production forecast time horizon. Next the various ways in which utilities and grid operators can use forecasts of variable generation are explored, including specific uses of solar forecasts at the transmission and distribution levels. Given that California is leading the nation in the adoption of solar energy, extended coverage of the California Independent System Operator s (CAISO s) uses of a central solar forecast is provided. Extended coverage of the Hawaiian Electric Company (HECO) and Tucson Electric Power (TEP) is also provided to highlight efforts at the distribution level to integrate solar forecasting into energy management systems. HECO has the highest PV penetration levels of any utility company and is on the leading edge of exploring how to use solar forecasts to manage high penetration of distributed PV. TEP is a power company and balancing authority in the Southwest that has experienced significant growth in utility-scale and distributed PV deployment. Next the results of the solar forecast provider survey are presented and the report Although the solar forecasting industry is in the early stages of market development and acceptance, several companies currently offer solar forecasting services concludes with a section on future developments and recommendations. Although the solar forecasting industry is in the early stages of market development and acceptance, several companies currently offer solar forecasting services to various end-users as part of other services they provide to the renewable energy industry and other weather-sensitive industries. Modern solar forecasting services use sophisticated physical numerical weather prediction models (NWP), cloud imagery analyses, and statistical manipulations to provide state of the art solar forecasting services covering a range of forecast time horizons (see Figure 1). A point solar forecast is prepared for a particular solar power facility, while an area forecast represents the predicted aggregate output of two or more solar power plants within a defined geographic region. One study finds that solar irradiance forecast error measured in root mean square error, a measure of the standard deviation of the differences between predicted values and observed values, can range from 100 watts 200 watts per square meter depending on the forecast method and time horizon. To place this in context, the solar resource is approximately 1,000 watts per square meter for a surface perpendicular to the sun s rays at sea level on a clear day. Irradiance forecast errors vary depending on the overall weather conditions; forecasts on clear sky days are more accurate than Figure 1: Modern Solar Forecasting System Schematic DATA (NWP, CLOUD MOTION, WEATHER, SOLAR PLANT) SOLAR PLANT PREDICTION ENGINE (PHYSICAL & STATISICAL MODELS) SOLAR PLANT SIMULATION SOLAR PLANT POWER PREDICTION solar electric power association

Executive Summary forecasts on partly cloudy days. Furthermore, forecast errors are lower for area solar power forecasts relative to point forecasts due to the fact that aggregation of geographically disperse systems reduces the random impact clouds have on solar generation system output. There is no standardized approach, however, to calculating and reporting forecast error for solar system predictions thus making comparisons between alternative forecasts difficult. While much can be learned from the experiences gained using wind forecasts in the past decade for utility planning and operations, the distributed nature of solar deployment and the impact that clouds have on solar resource variability create unique and different challenges. By and large distributed (DG), behind the meter solar is invisible to balancing authorities and utility companies meaning they do not have real-time production data to provide situational awareness. The demand-side nature of DG solar shows up through shifting load patterns. While utilities have decades of experience with load forecasting, DG solar results in a new element of variability and uncertainty to system load not captured by existing load forecasting methods. This is a new paradigm for utilities and much work is needed to integrate solar forecasting into existing load forecasting practices. The consistent patterns of load seen over time are changing due to DG solar and thus load forecasting techniques will need to be modified in response, including greater emphasis on intra-day and intra-hour load forecasting. Utilities and grid operators need to understand total system load, including the load being served by DG solar, to provide situational awareness. This is critical to proactively prepare for the net load impacts of DG solar variability resulting from changing weather patterns and emerging cloud formations. Systems and models are being developed today to make DG solar visible to grid operators and utilities. Clean Power Research is currently providing the California System Operator (CAISO) with forecasts of the 130,000+ DG solar energy systems in the State on an experimental basis. DG solar forecasts require data on DG solar systems, most of which is being collected through the net metering approval and interconnection processes, but might not yet be available for forecasting organizations to access. Solar Forecasting Service ProviderS Clean Power Research (U.S.) 3 Tier (U.S.) AWS TruePower (U.S.) JHtech--Solar Data Warehouse (U.S.) Windlogic, Inc. (U.S.) Global Weather Corporation (U.S.) Green Power Labs (Canada) Iberdrola Renewables, LLC (Spain) irsolav (Spain) Meteologica SA (Spain) NNERGIX Energy Management, S.L. (Spain) Meteocontrol Energy and Weather Services (Germany) WEPROG (Germany) Reuniwatt (France) ENFOR A/S (Denmark) Datameteo (Italy) DNV-GL (Garrad Hassan, Inc.) (Norway) In California the CAISO, utility companies and other market participants are using solar forecasting services for a variety of applications. Similar organizations in other regions including New Jersey, Hawaii, Arizona, and Colorado are beginning to experiment with using forecasts of solar plant production for operations and planning purposes. Grid operators and utilities in other regions are actively researching forecasting needs and potential solutions. Regional grid operators and balancing authorities can use forecasts of variable generation to determine the need for operating reserves and for scheduling utility-scale solar plant generation. An emerging use of variable generation forecasting is to predict significant ramping events caused by sudden changes in output from wind and solar generators. Utilities and energy traders can use forecasts of variable generation to develop bidding strategies for hour-ahead and day-ahead markets. Solar energy forecasts of distributed, behind the meter solar systems can be used to create more accurate load forecasts, as grid operators must prepare to meet system load v v Predicting Solar Power Production

net of DG solar. Distribution utilities can also use area forecasts of distributed solar to gain situational awareness to address potential reliability concerns within the distribution networks they manage. In addition, forecasts of variable generation, both central utility scale and distributed systems, can be used for planning purposes. In practice today, solar forecasting is used primarily at the system level to schedule energy production from large utility-scale solar plants. Developing tools that use forecasts of DG solar is still in the research and development stage in the U.S. Based on questionnaire responses, some forecast providers report challenges in obtaining timely and reliable data from their utility clients to prepare a solar forecast and in some cases a lack of the necessary information technologies to efficiently integrate with their forecasting systems. In addition, communications between forecast providers and end users can be challenging in terms of clearly articulating forecasting needs and the value that solar forecasting can bring to utility planning and operations. All solar forecasting service providers responding to the online questionnaire indicated ongoing research and development efforts to enhance their solar forecasting services. Predicting the output of solar energy plants for various forecast time horizons can be a valuable tool allowing grid operators and utilities to reduce the costs of integrating solar sources of generation into the existing grid. This report represents a snapshot in time, as we anticipate rapid developments in forecasting methods and uses in the coming years. Types of solar forecast Type vi Day-ahead, hour-ahead, & intra-hour Central utility-scale vs. DG solar Probabilistic vs. Deterministic Point vs. Area forecasts Output vs. Rate of change Forecasts based on time horizon. Large-scale vs. Small scale solar installations. A range of possible solar system output based on probability or a specific output value. Forecasts for one solar plant versus a forecast for an aggregate of geographically dispersed solar energy systems. Forecast of power output versus expected change in output over the forecast time step. vi solar electric power association

SECTION 1 1. Introduction Deployment of solar energy technologies nationally and globally is accelerating. In October of 2013 solar reached an interesting milestone as being the sole new source of generation added to the nation s power grid that month at 530 MW. Analysts project that 2013 saw a new record in solar energy deployment in the U.S., with a projected 4,300 MW-dc of PV to come online representing a 27% growth over 2012 installation totals, while cumulative PV capacity is projected to surpass 10 GW-dc (Kann et al., 2013). Furthermore, it is projected that 2013 will be a record year for concentrating solar power (CSP) as over 800 MW-ac is expected to be commissioned (Kann et al., 2013). Economies of scale in manufacturing and continuous technical improvements should result in future cost reductions further strengthening solar energy s position as an expanding resource to meet society s demand for electricity. The variable nature of solar energy production must be actively managed when the level of grid penetration reaches a certain threshold, either on a particular feeder or transmission line or at the system level. As the case with wind power, utilities and grid operators in the U.S. began to use wind power forecasting to manage the variable nature of the resource as installed wind capacity expanded exponentially in the past decade. According to data from the American Wind Energy Association, today there is over 60 GW of installed wind capacity nation-wide. Although the amount of installed solar when forecasting may provide value is not known with certainty and will be different for various utilities and balancing authorities, Pelland et al. (2013) suggest that as a rule of thumb for when solar power forecasts appear to be needed is when yearly solar energy production reaches a level of 1% 2% of annual energy demand. Solar forecasting as a decision support tool for utilities, balancing authorities, and power market participants is in the early stage of market development and acceptance. The wind forecasting industry is more developed and many different user groups reap benefits from integrating wind power forecasts into their decision frameworks. The CAISO was the first ISO/ RTO to start using wind power forecasting for system operations when it introduced forecasting as part of its Participating Intermittent Resource Program (PIRP) in 2004. Three other ISOs/RTOs Midcontinent ISO (MISO), New York ISO (NYISO), and Electric Reliability Council of Texas (ERCOT) all introduced centralized wind power forecasting in 2008; PJM Interconnect (PJM) began implementing its wind forecasting system in 2009, South West Power Pool (SPP) in 2011 and the ISO New England (ISONE) recently instituted a central wind forecasting system at the end of 2013. Many of the power market reforms and grid management strategies designed to leverage wind farm forecasts to address the variable nature of the resource are directly applicable to solar. However, solar energy has its own unique challenges in terms of forecasting methods and applications. Much of the variability in solar output is predictable, given that the sun changes position throughout the day and throughout the seasons. Rapid changes in ground-level irradiance are driven by clouds, which are inherently difficult systems to model; as discussed later in this report current research and development efforts are focused on improving this aspect of solar forecasting. Another significant difference with solar relative to wind is the distributed nature of solar energy deployment. Much of the installed solar to date is embedded within distribution networks and is largely invisible to grid operators, and therefore DG solar systems do not currently have the ability to be curtailed if necessary like larger-scale wind plants. Systems and models are currently being developed to make DG solar visible to grid operators, and new inverter technologies and smart meters may provide grid mangers with tools to curtail load and or solar output to address reliability concerns from DG solar variability. This report provides a review of solar forecasting approaches and how they are being used by grid operators, utility companies, and other market participants for planning and operations. This report 1 1 1 Predicting Solar Power Production

is based on a survey of the literature and interviews with experts from three broad stakeholder groups (ISO/RTO managers, utility managers, and research scientists). Seventeen companies globally were identified that potentially provide solar forecasting services and asked to complete an online questionnaire. Thirteen companies completed the questionnaire, all of whom work with utilities or related organizations. These companies use state of the art solar forecasting systems to provide customized forecasts to meet their clients particular needs. The report begins with a review of the variable nature of solar and is then followed by an examination of the various ground-level solar irradiance forecasting approaches and models used for each solar production forecast time horizon. Next the various ways in which utilities and grid operators can use forecasts of variable generation are explored, including specific uses of solar forecasts at the transmission and distribution levels. Given that California is leading the nation in the adoption of solar energy, extended coverage of the CAISO s uses of a central solar forecast is provided. In addition, extended coverage of Hawaiian Electric Company (HECO) and Tucson Electric Power (TEP) is presented, as they have some of the highest levels of PV penetration and are beginning to experience system impacts from solar and other sources of variable generation. As a result, they are on the leading edge of exploring how to use solar forecasts to manage high penetration levels of distributed PV. 2 2 2 solar electric power association

SECTION 1 2. The Solar Resource It is widely recognized that the solar resource is vast. Each hour the sun showers the earth with enough potential energy to more than satisfy global energy demand for an entire year. While humans have developed a number of simple and highly technical systems to convert solar energy into useful forms, solar energy still remains a small part of our overall energy economy. Here we cover the fundamentals of solar irradiance and analyze the variable nature of the resource, which is a key barrier to greater use of the vast solar resource. 2.1 Fundamentals of solar irradiance The quantity of fossil fuel sources of energy are assessed based on sophisticated geologic surveys that quantify the stock of recoverable reserves. Solar energy, in contrast, is quantified based on an assessment of energy flows over various time frames and at particular locations. Typically this flow is quantified as kwh per square meter per time period (day, month, or year) for a particular location on earth. Solar irradiance (irradiance is the power of electromagnetic radiation per unit area incident on a surface) consists of two primary components: direct and diffuse. Direct normal irradiance (DNI) is the amount of solar radiation received per unit area by a surface that is always held perpendicular (or normal) to the rays that come in a straight line from the direction of the sun at its current position in the sky. Diffuse Horizontal Radiation (DHI) is the amount of radiation received per unit area by a surface (not subject to any shade or shadow) that does not arrive on a direct path from the sun, but has been scattered by molecules and particles in the atmosphere and comes equally from all directions. Global Horizontal Irradiance (GHI), also referred to as total solar radiation, is the sum of DNI and DHI. The relevant irradiance metric for concentrating solar technologies is DNI, while GHI is relevant for all nonconcentrating solar technologies and applications. 1 To properly analyze the potential of solar power, understanding the sun s trajectory across the earth s sky is critical. The amount of potential solar energy striking 1 This report focuses primarily on GHI forecasting as DNI forecasting is notoriously difficult and is less critical as many concentrating solar power plants integrate energy storage or natural gas combustion backup to address the variability challenge. Forecasting GHI is an important tool to address the variability of utility-scale and DG solar photovoltaic plants, which is the focus of this report. Predicting Solar Power Production a collector surface varies as the solar azimuth angle changes throughout the day. The accepted convention for the solar azimuth angle for solar energy applications is clockwise from due north, so east is 90, south is 180 and west is 270. In addition, the solar zenith angle, or solar elevation, defines how high the sun is from the horizon, which changes throughout the year with the seasons. Solar energy collectors can be designed to track the sun seasonally based on changes in the solar zenith angle or daily based on the changing solar azimuth angle (single axis tracker) or both (double axis tracker). Concentrating solar energy technologies utilize duel-tracking systems to position the solar collector devices perpendicular to the sun s rays to maximize DNI exposure. The National Renewable Energy Laboratory (NREL) maintains the National Solar Radiation Database, (NSRDB). The updated 1991-2010 NSRDB contains hourly solar and meteorological data for 1,454 locations in the U.S. and its territories. Figure 2 identifies the 1,454 different NSRDB locations by station classification. The station classification is based on data quality and completeness. Class I and II sites have no missing data, with the former having less than 25 percent of the data for the period of record exceeding an uncertainty of 11 percent (Wilcox, 2012). NREL produces typical meteorological year (TMY) datasets derived from the 1961-1990 and 1991-2005 NSRDB archives, with the latest version referred to as TMY3. The TMY3 data sets contain hourly values of solar radiation and meteorological elements for a one-year period; these hourly insolation values are 3 3 3

Figure 2: NSRDB stations by class Figure 3: U.S. solar resource map 4 4 (Source: http://rredc.nrel.gov/solar/old_data/nsrdb/1991-2010/figure1.html) based on averaging data from multiple years and thus represent typical conditions. Their intended use is for computer simulations of solar energy conversion systems and building systems to simplify performance comparisons of different system types, configurations, and locations. Given that they represent typical rather than extreme conditions, NREL suggest that they are not suited for designing systems to meet the worst-case conditions occurring at a particular location. Figure 3 presents the solar resource map for the U.S. derived from TMY3 data files. The annual average daily solar energy striking a flat-plate fixed collector in the U.S. ranges from an annual average of over 6.5 kwh/m 2 /day in the Southwest, to approximately 3.5 kwh/m 2 /day in the Pacific Northwest. Solar is an abundant energy resource, but not as concentrated relative to fossil fuel energy sources. In the early 1990s solar resource assessment techniques were developed using satellite-derived cloud cover (Source: http://www.nrel.gov/gis/images/eere_pv/ national_photovoltaic_2012-01.jpg) data. Through well-tested algorithms cloud cover data from geostationary weather satellites can be used to estimate ground-level irradiance at a particular location at any given time; this method was adopted as part of the 2010 NSRDB update. Satellite-based resource assessment techniques have been rigorously validated using ground-level irradiance measurements (Perez and Seals, 1997). The availability of satellite-derived solar resource data providing time and location specific irradiance data freed analysts from the limitations of average solar resource data, thus allowing detailed analyses of solar energy s potential contribution as a grid-connected resource (Letendre and Perez, 1996). These techniques have been further refined and developed by NREL to create simulated solar production datasets for the continental U.S. for use in large-scale solar integration studies. As discussed below, the use of satellite cloud images for irradiance forecasting is a fundamental technique used today for solar power predictions up to 6 hours ahead. 2.2 The variable nature of solar energy 4 As described above the solar resource striking a collector surface, either fixed or tracking, is variable throughout the day and throughout the year. To a large degree the variability associated with the position of the sun in the sky impacts solar power generating systems in a relatively uniform manner and can be readily anticipated. The stochastic, short-term variability component of irradiance striking the earth driven by clouds and weather conditions creates uncertainty in output from solar energy systems, which is of greatest concern to grid operators and utility companies. Changes in the sun s position during the day can result in solar electric power association

SECTION 2 a 10% - 30% change in output over a 15-minute interval for single-axis tracking photovoltaic (PV) plants. In contrast, changes in insolation at a particular point can exceed 60% of the peak insolation in a matter of seconds (Mills et. al, 2009). The impact on a PV plant s output from passing clouds, however, depends on the size of the system, cloud speed, cloud height, and other factors (Mills et al., 2009). Larger arrays will experience less variability from passing clouds as compared to smaller arrays. For PV arrays with a rated capacity of 100 MW, the time it takes to shade the system will occur over minutes not seconds. Electricity is a unique form of energy that must be both generated and consumed simultaneously. Although electricity can be stored for later use, the current grid has very little capacity to store electricity and roundtrip efficiency losses associated with energy storage are significant. Thus, grid operators have implemented systems and procedures that consistently match power supply with power consumption to achieve reliability standards established by national and regional electric power reliability organizations. Figure 4: Time scale relevant to operating power systems SYSTEM LOAD (MW) A A 0 4 8 12 16 20 24 TIME (HOUR OF DAY) SECONDS TO MINUTES REGULATION B B TENS OF MINUTES TO HOURS LOAD FOLLOWING C C DAY SCHEDULING Variability is an inherent characteristic of the electric power system given that the demand for power varies throughout the day and the seasons. Although there is uncertainty associated with load forecasting, experience has afforded the industry a level of comfort around load forecast errors and their trends. Furthermore, system operators must also manage the uncertainty associated with generators and transmission assets that may not perform as anticipated due to technical malfunctions or weather-related factors. System operators and planners rely primarily on operating reserves, which are additional generating units scheduled beyond what is anticipated to meet forecast loads, to address the variability and uncertainty of power systems (Mills et al., 2009). Figure 4 illustrates the time scales relevant to operating power systems and the strategies used to manage the variability and uncertainty associated with each distinct time interval. Grid operators have decades of experience operating power systems with conventional, dispatchable power plants fueled with coal, natural gas, nuclear or hydro resources. Dispatchable resources tend to use fuels that can be stored or otherwise predictably relied upon over relatively short time frames, and technologies that have relatively known, quantifiable outage rates associated with them. Undoubtedly, greater use of intermittent generation including wind and solar will add a new dimension of variability to the power system. In response to rapid growth in wind and solar generation on regional grids and state-level renewable energy mandates, the U.S. DOE and the NREL teamed-up to analyze the impacts of increasing wind and solar energy on the western and eastern electric grids. Initiated in 2007, the Eastern Wind Integration and Transmission Study (EWITS) evaluated the power system impacts associated with 20% - 30% wind energy penetration on the Eastern Interconnection. The Western Wind and Solar Integration Study (WWSIS) analyzed the operational implications associated with up to 35% wind and solar energy penetration on the Western Interconnection. The high renewable 5 5 5 (Source: Michael Milligan, NREL, presentation given at PV Variability Workshop) Predicting Solar Power Production

6 6 6 energy penetration scenarios in these studies were evaluated in terms of integration costs as well as potential transmission constraints and operational changes that might arise. Phase II of the WWSIS was completed in September of 2013. Bird and Lew (2012) provide a summary of the lessons learned from the Phase I EWITS and WWSIS studies, and others, on integrating variable generation into the U.S. bulk power system. The key findings are: Higher penetrations of variable renewable generation are manageable: The Eastern and Western studies found that by 2025, with operational changes and expanded transmission access, high penetrations up to 30% or 35% of variable renewable generation are technically feasible to integrate so that load and generation can be balanced for all hours of the year. Larger balancing areas and geographic diversity of renewable resources are preferred to minimize impacts: Both the Eastern and Western studies as well as other recent studies show that large operating areas and sufficient transmission access are effective measures for managing wind energy. Larger operating areas provide benefits of reducing the variability of the wind and solar when it is spread over a larger geographic area and providing access to more resources for balancing the system. Sub-hourly scheduling and planning are preferred: Both studies reinforce the idea found by other integration studies that sub-hourly scheduling and dispatch can aide in the integration of higher penetrations of wind and solar on the grid. Using advanced forecasting can reduce costs: Forecasting is an important tool for reducing the uncertainty and managing the variability associated with wind and solar generation. While forecast accuracy is increased closer to the time of actual generation, incorporating day-ahead forecasting into the process of committing generation units can help operators mitigate the uncertainty of wind generation. Demand response can be cost-effective: Demand response is one mechanism for providing systems with operational flexibility. It can be useful for assisting with the integration of variable renewable generation by helping to address large loss of generation events. Integration costs are manageable: Systems can incur costs due to increased balancing resources needed to manage variable renewable energy generation on the system. The costs can include the need for additional operating reserves, changes in dispatch, and new transmission. Renewables can play a role in resource adequacy capacity value: While wind and solar are primarily energy resources, they can contribute to resource adequacy of the system. However, in contrast to conventional generators, only a portion of the capacity of wind and solar facilities can be relied on to be available to meet peak demand. (Bird and Lew, 2012, pages 4 8) Phase II of the WWSIS evaluated the increased costs associated with increased cycling and ramping of conventional generation resources in response to high penetration of wind and solar, and the associated emissions impacts. The study concluded that increased cycling costs assuming 33% wind and solar penetration are small relative to the reduced fuel costs, which were estimated to be $0.14 $0.67/MWh and $28 $29/ MWh respectively. In terms of emissions impacts the study found that cycling had very little (<5%) impact on the CO 2, NO X, and SO 2 emissions reductions from wind and solar and that wind- and solar-induced cycling can have a positive or negative impact on emissions rates depending on the generation mix and wind and solar penetration level. The high renewables penetration studies to date highlight the importance of advanced forecasting of variable generation, which allows grid operators to anticipate and actively manage the uncertain nature of wind and solar plants. Experience to date in the wind sector suggests that significant economic benefits are realized from integrating forecasting into scheduling systems and market operations. The economic benefits of forecasting are a result of reduced imbalance charges and penalties, competitive knowledge advantage in real-time and day-ahead energy market trading, and more efficient project construction, operations, and maintenance planning. solar electric power association

SECTION 3 3. Forecasting Solar Irradiance Solar irradiance forecasting provides a critical input to predicting a solar power plants output at various points in the future. As discussed above, solar plant predictions provide grid operators, utilities, and market participants data for use in decision support tools, including scheduling reserve capacity or developing bidding strategies for hour-ahead and day-ahead wholesale power markets. These different uses of solar power predictions require different types of forecasts. For example, a forecast may apply to a single PV system (point forecast) or refer to the aggregation of large numbers of distributed PV systems spread over an extended geographic area (area forecast). A forecast that focuses on the rate of change in solar power output may be needed for decision support tools designed to predict significant ramp events on regional grids. Solar power prediction methods are generally characterized as physical or statistical, however in practice the lines between these approaches is blurred. Physical approaches explicitly model physical atmospheric phenomenon as part of the irradiance prediction using numerical weather prediction (NWP) models or sky images. Statistical approaches predict irradiance from training and statisticallyderived values. For example, a physical approach may use developing cloud vector-based forecasts via interpolation of recent consecutive sky images and a statistical approach may use current and historical power output alone to predict future output. NWP model outputs can also be fed into statistical models to improve the forecast, but often it is the case where physical NWP models have to be developed and run at higher spatial and temporal resolutions than the general NWP models. The solar plant simulation component of a forecasting system is often a separate module that uses the required meteorological forecasts (e.g., irradiance, ambient temperature, and wind speed) as inputs, and can also be in the form of a physical model with known specifications that can be modeled or a statistically-derived one. Solar forecast providers in practice draw from multiple forecasting methods to tailor solar plant power predictions to end-user needs. A persistence forecast is often used as the benchmark for assessing the skill of various solar plant power prediction models. A persistence forecast simply extrapolates current plant output into the future, with perhaps adjustments based on changing sun angles. Figure 5 illustrates the key components of an advanced solar plant power prediction system. The solar plant simulation component is required to predict solar Figure 5: Illustration of solar plant power prediction system 7 SOLAR PLANT PREDICTION ENGINE NWP DATA CLOUD MOTION DATA WEATHER DATA SOLAR PLANT DATA PHYSICAL MODELS (IRRADIENCE/MODULE TEMPERATURE) STATISTICAL MODELS NORMALIZED POWER SOLAR PLANT POWER PREDICTION 7 7 SOLAR PLANT SIMULATION 0 2 4 6 8 10 12 14 16 18 20 22 24 TIME Predicting Solar Power Production

plant power production using a set of mathematical equations that combines solar resource information with solar system specifications to produce power output predictions. The majority of solar energy plants receive incoming irradiance on a tilted plane. As a result, the forecasted GHI has to be converted according to the orientation and tilt of the solar plant collectors to model the irradiance actually utilized by the solar energy plant. Solar energy power predictions can be provided to end-users either in a deterministic or probabilistic format, with the former being a specific value for solar energy production at a particular time step (15-minutes or one hour) and the latter being a range of values based on probability theory. While still in the research phase, forecasts could also provide information on the expected variability of output for the time period for which energy output is being forecasted. For example, an hour ahead forecast may predict 350 MW on average for the next hour for a 500 MW plant; this does not indicate how variable that output will be during the forecast hour. 3.1 Physical methods to forecast solar irradiance The numerical weather prediction (NWP) method draws from one or more of the several existing global weather models to forecast plane of array irradiance and provide the necessary meteorological data to estimate back of module temperature, and is generally accepted as the most accurate method for forecast horizons of 6 hours and beyond. Cloud imagery techniques, using either satellite cloud images or images from terrestrial-based sky imaging devices, predict cloud motion into the future and then apply established algorithms to predict plane of array irradiance; these methods are generally understood to produce the most accurate forecasts in the 0 6 hours ahead time horizon. Solar forecast service providers integrate various physical and statistical methods (dictated by the forecast time horizon) to forecast plane of array irradiance. Post processing of physical data is used to eliminate systematic errors and tune the solar plant simulation model based on a variety of relevant data including historic solar plant production data; these post-processing methods are referred to as model output statistics (MOS). 8 8 8 3.1.1 Numerical weather prediction (NWP) NWP uses the power of computers to make weather forecasts using complex computer programs run on supercomputers that generate predictions of many atmospheric variables such as temperature, pressure, wind, and rainfall. Three-dimensional models of the atmosphere and oceans are used to predict the weather based on current weather conditions; initialization is the term used to describe the process of entering observation data into the model to generate initial conditions. A number of global and regional forecast models are run in different countries worldwide, using current weather observations relayed from devices known as radiosondes on weather balloons, radar, weather satellites and ground station measurements as inputs. NWP model runs are typically initiated two to four times per day, for example at 0, 6, 12 and 18 UTC (Universal Time Coordinated). In order to limit computational requirements, global NWP models have relatively coarse resolutions with grid spacing on the order of 40 km to 90 km. Mesoscale NWP models cover a limited geographical area with higher resolution, which attempt to account for local terrain and weather phenomena in more detail than global models. The weather forecasts that we rely on daily are the results of one or more of these NWP models. Perez et al. (2013) combined and evaluated three independent validations of GHI multi-day NWP forecast models that were conducted in the U.S., Canada, and Europe. Two NWP models were used in each of the three validation efforts analyzed the European Centre for Medium-Range Weather Forecasts (ECMWF) global model and the Weather solar electric power association

SECTION 3 Research Forecasting (WRF) mesoscale model initialized with Global Forecast System (GFS) forecasts. They concluded that the GFS-based WRF model forecasts do not perform as well as global forecastbased approaches such as the ECMWF model and that the simple averaging of the various models outputs perform better than individual models. Combining multiple NWP forecasts is a standard approach for improving forecast performance, which is referred to as an ensemble forecast. Ensemble forecasts can also be produced by varying the initial conditions or physical parameterizations within a single NWP model to yield a probabilistic forecast. Perez et al. (2013) also found that post-processing using MOS techniques significantly improve forecast performance for mesoscale models in the U.S. Statistical tools applied to forecasts generated using the Canadian Global Environmental Multiscale Model (GEM) based on GFS forecasts achieved similar performance as the ECMWF forecasts in North America and in Central Europe. 3.1.2 Cloud imagery The second broad category of physical solar power prediction approaches uses cloud images to predict cloud movement, and the associated impacts on plane of array irradiance and back of module temperature into the future at a particular geographic site. This approach reduces forecast error over NWP forecasts in the 0 6 hour-ahead (intra-day) timeframe. Higher spatial resolution (10-100 meters) and shorter sampling rates (30 seconds) of cloud images are achieved using total sky imaging (TSI) devices relative to satellite cloud images (1 km and 15 minutes respectively). TSI-based forecasting is used to predict solar plant power output in real time (nowcast) up to 10-30 minutes ahead (intra-hour) through advanced image processing and cloud tracking techniques. Due to the limited view capabilities, TSI devices are unable to detect clouds that will impact a site beyond the 30 minute time horizon. TSI devices take an overhead image of the surrounding sky (Figure 6), which provides the images that are then processed to generate a forecast. In general this approach assumes the opacity, direction, and velocity of movement of the clouds in the future is consistent with the initial conditions observed through the TSI device. Irradiance is predicted based on current cloud shadows and then the cloud shadows are relocated forward in time based on cloud velocity and direction to generate the forecast (Pellend et. al, 2013). One total sky imager, depending on cloud height, can deliver an image for 5-10 square miles under cloudy sky conditions and 15 square miles in totally clear sky conditions. Thus, one imager can be used for a multi-mw solar farm, but several TSI devices would be needed for a multi-hundred MW farm. Solar power predictions using a TSI-based system are the best technique to predict short term ramps for individual solar power plants. However, the need for managing ramps on a large grid system decreases with geographic distribution (discussed in greater detail below). Thus, total sky imagers may have limited applicability to island regions, small balancing authorities, and perhaps to independent solar power producers that must meet certain ramp rate parameters. Given that TSI devices are not ubiquitous, cloud images obtained from geostationary satellites (Figure 7) provide an alternative way to obtain cloud images, which are available for the entire globe albeit with lower spatial resolution and longer sampling rates than TSI images. Clouds reflect light into the satellite leading to detection and the ability to calculate the amount of light transmitted through the cloud, thus generating cloud images with the opacity characterized that are processed in a similar manner to cloud images generated using TSI devices. The direction and speed of clouds can be assessed using subsequent satellite images to predict future impacts on plane of array irradiance for a particular solar plant site. The lower spatial and higher temporal resolution for satellite images causes satellite forecasts to be less accurate than sky imagery on intra-hour time scales, but extensive comparisons or combinations of the two approaches have not been conducted (Pellend et al., 2013). The satellite imagery technique for solar power predictions is considered the best forecasting technique from 1 to 6 hours in advance of the forecast time horizon. 9 9 9 Predicting Solar Power Production

Figure 6: Total sky imager image Figure 7: Satellite cloud images (Source: http://geneq.com/catalogues/tsi-880_2.gif) (Source: Perez et al., 2010a) 3.2 statistical/hybrid methods to forecast solar irradiance 10 10 10 A persistence forecast is the most basic statistical approach to predict the future output from a solar plant; past time series data is used to forecast solar plant power output in the future, with minor adjustments based on the sun s position in the sky. Predicting solar plant production using purely statistical methods is not generally part of modern solar power prediction systems. However, hybrid approaches make use of advanced statistical techniques to correct for known deficiencies associated with different forecasting methods through adjustments for model biases or automated learning techniques. As mentioned above, MOS uses statistical correlations between observed weather elements and climatological data, satellite retrievals, or modeled parameters to obtain localized statistical correction functions. This allows for the enhancement of low-resolution data by considering local effects including topographic shading, or for correcting systematic deviations of a numerical model, satellite retrievals, or ground sensors. One of the major disadvantages of statistical methods is the requirement for accurate historic datasets used to develop statistical correlations separately for each location. This means that MOS-based forecasts are not immediately available for larger areas or for locations without prior measurements, such as most non-urban solar power plants (Kleissl, 2010). Purely statistical approaches to solar power predictions do not attempt to simulate future production using irradiance and module temperature; their starting point is a training dataset that contains a variety of inputs that could include NWP outputs, ground station or satellite data, historic solar plant production data etc. The dataset is used to train models such as autoregressive or artificial intelligence models that produce a forecast of solar plant output over a specified future timeframe based on past inputs available at the time when the model is run (Pelland et al., 2013). The line between physical and statistical methods is blurred in today s modern solar power prediction systems. Best practices adopt a hybrid approach whereby physical models produce irradiance and module temperature forecasts used in a solar plant simulation model, the output of which is then subject to statistical post-processing to improve accuracy. Past solar plant output is an important source of data to train MOS post-processing models thereby improving future forecast skill. The best statistical approaches make use of the knowledge from physical models to select input variables used to train a statistical solar electric power association

SECTION 3 forecasting model. As discussed later in this report, IBM was awarded a grant from the U.S. DOD in 2012 to advance the state of the art in statistical applications to improve solar forecasting. The ability to leverage new computational techniques including machine learning and the rapid statistical processing of large amounts of data using super computers will likely have a significant impact on the quality of solar power prediction models in the coming decade. 3.3 understanding solar forecast errors and variability Ultimately, a forecast s accuracy is judged by the difference between the solar plant output prediction and the actual plant output over some time interval (hourly) and some specified time period (year). Much research has been done to characterize the error of various forecasting approaches by comparing solar irradiance forecast results with ground-measured irradiance to assess the accuracy of various solar forecasting approaches over multiple time horizons. The three standard error metrics reported in the literature include the root mean square error (RMSE) measuring a model s overall performance (especially where extreme events are a concern), the mean absolute error (MAE), another measure of overall performance, and the mean bias error (MBE) a measure of a forecast s bias, meaning its tendency to over or under predict. These error metrics for solar irradiance are given in watts per square meter, but can also be presented in relative (%) terms; PV forecasts are commonly normalized by dividing by the rated capacity of the PV system for which the forecast was generated. In terms of irradiance forecasting, the RMSE represents the square root of the sum of the squares of the difference between hourly forecasted irradiance and hourly measured irradiance over a one year period. The MAE is defined in the context of irradiance forecasting to be the sum of the absolute values of the differences between hourly forecasted irradiance values and hourly measured irradiance values over a one year period divided by the number of observations (8,760). The MBA is calculated similarly to the MAE with the exception that the differences between hourly forecasted irradiance values and hourly measured irradiance values are simply summed and not converted to absolute values, thus this error metric will indicate whether forecast tends to over or under predict the solar resource. The three standard error metrics reported in the literature include the root mean square error (RMSE) measuring a model s overall performance (especially where extreme events are a concern), the mean absolute error (MAE), another measure of overall performance, and the mean bias error (MBE) a measure of a forecast s bias, meaning its tendency to over or under predict. Climate conditions and forecast horizons have been shown to directly impact the accuracy of irradiance forecasts. Clear skies are relatively easier to forecast resulting in lower forecast errors than partly cloudy and overcast weather conditions. Forecast accuracies also decrease as the forecast time horizon increases. The designation of the best solar forecasting approach given forecast time horizon identified earlier in the report is based on research that characterizes the error of different forecasting approaches for various forecast time horizons. Figure 8 is an important graph produced by Perez et al. (2010b) demonstrating that satellite cloud motion forecasts perform better than NWP forecasts in the 1 6 hour time horizon, beyond which NWP produces a superior forecast. In addition, Figure 8 illustrates how forecast error increases as the forecast time horizon extends into the future. The forecast algorithms evaluated in Figure 8 were used to forecast hourly GHI, which were validated against ground measurements for seven climatically distinct locations in the U.S. for one year. The short term forecasts that extend up to 6-hours ahead are based upon cloud motion derived from consecutive geostationary satellite images. The medium term NWP forecasts extend up to 6-days ahead and are modeled from gridded cloud cover forecasts from the U.S. 11 11 11 Predicting Solar Power Production

12 12 12 Figure 8: Annual RMSE as a function of forecast time horizon RSME (W/M -2 ) 7-SITE COMPOSITE 300 250 200 150 100 50 1 2 3 4 5 6 2 3 4 5 6 7 HR HR HR HR HR HR DAY DAY DAY DAY DAY DAY FORECAST TIME HORIZON MEASURED PERSISTENCE CLOUD-MOTION FORECAST NDFD SATELLITE REFERENCE (Source: Perez et al., 2010b) Figure 9: SURFRAD Station Locations DESERT ROCK, NV FORT PECK, MT BOULDER, CO SIOUX FALLS, SD BONDVILLE, IL GOODWIN CREEK, MS (Source: http://www.esrl.noaa.gov/gmd/grad/surfrad/sitepage.html) PENN STATE, PA National Digital Forecast Database (NDFD). These two methods are compared against two benchmark forecasts, including a simple persistence forecast and a satellite nowcast (irradiance estimates at the time the satellite images were taken). Hourly forecasts were tested against irradiance data from the seven Surface Radiation Budget Network (SURFRAD) stations across the U.S. as depicted in Figure 9, which cover several distinct climatic environments. Table 2 contains RMSE metrics for hour-ahead and day-ahead solar forecasts for each of the seven SURFRAD stations (Perez et al., 2010b). The RMSE statistics and the mean observed GHI are presented in watts/m 2. The table also includes a clearness index in percentages for each location, with higher percentages indicating relatively clear skies in comparison to lower values. The data presented here is based on the entire validation period of one year. As observed from the data presented in Table 2, hour-ahead solar forecasts have lower errors than day-ahead forecasts. The standard error metrics used to assess solar forecasts provide different perspectives on forecast quality and their published values depend on the time span of the validation analysis, use of sunlight hours only, and data aggregation techniques which could make comparisons between forecasts meaningless (Kostylev and Pavlovski, 2011). Hoff et al. (2012) evaluate different approaches to reporting relative error metrics for irradiance forecasting offering recommendations for the most appropriate approach using a subjective grading scale including whether: The method is commonly accepted; Is simple to understand; Depends on the 24-hr. versus daytime only distinction; Depends on the data selection threshold (the value below at which data is excluded for night time hours); and Is affected by outliers. Based on these criteria, they conclude that the relative error metric of MAE divided by the average measured irradiance values over the validation period provides the best practical measure of relative dispersion error for solar resource forecasts. They also suggest that the RMSE divided by solar plant capacity is desirable because it is commonly accepted and is simple to understand; furthermore the wind industry has already adopted this method. An alternative approach to assessing the performance of a forecasting system is to use categorical statistical measures, which assess whether or not the forecast accurately predicts specific events. Table 3 was taken from a recent DOE presentation outlining such a framework. Here the event under consideration would be ramps (up or down) equal to 20% of the rated solar capacity occur in a given hour. Simply solar electric power association

SECTION 3 Table 2: Annual Error Metrics Day-Ahead and Hour-Ahead Forecasts for SURFRAD Locations Location Mean Observed GHI Clearness Index RMSE Hour-Ahead Forecast RMSE Day-Ahead Forecast DESERT ROCK, NV 498 w/m 2 90% 80 w/m 2 125 w/m 2 FORT PECK, MT 357 w/m 2 75% 94 w/m 2 148 w/m 2 BOULDER, CO 369 w/m 2 71% 120 w/m 2 188 w/m 2 SIOUX FALLS, SD 364 w/m 2 76% 68 w/m 2 140 w/m 2 BONDVILLE, IL 349 w/m 2 69% 85 w/m 2 151 w/m 2 GOODWIN CREEK, MS 397 w/m 2 76% 80 w/m 2 149 w/m 2 PENN STATE, PA 323 w/m 2 66% 86 w/m 2 141 w/m 2 (Sources: Perez et al., 2010b) Table 3: Event Example Up-Ramp or Down-Ramp of 20% Capacity in 1 Hour Contingency Table Observation YES No TOTAL Forecast Yes Hit False Alarm Forecast Yes No Miss Correct Null Forecast No Total Observation Yes Observation No Total Sample Size Example: Accuracy = (Hits + Correct Negs)/Total 13 adding up the number of times the forecast predicted the correct outcome and divide by the total number of predictions provides an assessment of the forecasting system in terms of the percentage of times the forecast predicted the outcome correctly. Ultimately, the appropriate error metric must meet the needs of the end-user and be relevant for the particular use of the solar power prediction. Grid operators need metrics that accurately reflect the costs of forecast errors, while solar resource research scientists require indicators of the relative performance of different forecast models, and of a single model under different conditions. Solar forecast user groups should engage in a collaborative process to provide guidance to forecast providers in terms of what they need to know relative to the quality of the forecast services being provided. 13 13 Predicting Solar Power Production

3.3.1 Point versus area solar energy plant predictions and PV power output variability 14 14 14 As discussed in the next section, there are a variety of solar forecast end-users and applications. Certain end-user groups, in particular balancing authorities, are concerned with the aggregate impact that the fleet of solar power systems within their geographic footprint has on system reliability and market efficiency. The diminished variability of the output from multiple wind turbines within a wind farm relative to a single wind turbine is well understood and documented. Variability in power production further declines when a broader geographic perspective is adopted including multiple geographically dispersed wind farms. The underlying weather changes that impact wind speeds do not impact geographically dispersed wind farms uniformly or at the same time, thus reducing the aggregate impact on variability of concern to grid operators. Similar to wind plant output, the variability of multiple solar power plants when viewed in aggregate is much less than a single solar power plant. There is currently no standard for characterizing PV power output variability. Perez et al. (2011) developed a set of empirical models capable of extracting metrics quantifying the short-term variability of the solar resource based upon site/time specific satellitederived hourly irradiance data. The model is derived from over 92,000 experimental hourly data points at twenty-four sites in the U.S. The model returns four metrics characterizing intra-hourly variability; these include the standard deviation of the global irradiance clear sky index (ratio of actual insolation to the clear sky insolation), and the mean index change from one time interval to the next, as well as the maximum and standard deviation of the latter. The variability time scales addressed in the model are 20 seconds, 1-minute, 5-minutes and 15 minutes. The variability within a given time interval (e.g., one-minute changes over one hour) is reduced when multiple geographically dispersed solar energy systems are aggregated. Hoff and Perez (2010) find that the power output variability (one minute time interval) from 1,000 MW of dispersed 4-kW residential PV systems corresponds to the output variability associated with 0.2 percent of what it would be if the PV capacity was concentrated in a single location. Hoff and Perez (2011) develop a method to analyze irradiance correlation coefficients between two locations as a function of distance, time interval, and other parameters. The left side of Figure 10 illustrates measured 10-second irradiance data, while the right side of the chart illustrates the change in irradiance using a 10-second time interval for a network of 25 weather monitoring stations in a 400-meter grid located in Cordelia Junction, CA on November 7, 2010. The light gray lines represent irradiance and variability for an individual station, while the red lines correspond to average irradiance and variability across 25 locations. Based on this analysis, variability is reduced by more than 70% when comparing the single-site measurements with the average of the measurements from the 25 sites. Correlation analyses from 70,000 station pair combinations demonstrates that irradiance correlation coefficients between two sites decreases predictably with distance and that the decrease occurs more slowly with longer time intervals (Hoff and Perez, 2011). This provides statistical evidence that variability of solar power predictions between plants becomes increasing unrelated with increased distance and thus aggregation reduces the overall variability of multiple PV systems output. Several studies demonstrate that error metrics for single point irradiance forecasts are much higher than an average of forecasts for multiple points, commonly referred to as an area forecast. For example, Lorenz et al. (2009) demonstrate that the day-ahead irradiance forecasts for single stations in Germany show a RMSE of 36%, where as an area forecast for the entire country results in a RMSE of 13%. In addition, Pelland et al. (2011) found that the RMSE for the forecast of the average irradiance of 10 ground stations across Canada and the U.S. was approximately solar electric power association

SECTION 4 Figure 10: Quantifying Solar Variability IRRADIANCE 10-SECOND CHANGE IN IRRADIANCE IRRADIANCE (W/M^2 ) 1000 750 500 250 0 6:00 12:00 18:00 PACIFIC STANDARD TIME IRRADIANCE CHANGE (W/M^2 ) 500 250 0-250 -500 6:00 12:00 18:00 PACIFIC STANDARD TIME 1 LOCATION 25 LOCATIONS CLEAR SKY 10 SEC. DATA FROM 400M X 400M GRID AT CORDELIA JUNCTION, CA (NOV. 7, 2010) 1 LOCATION (α = 31) 25 LOCATIONS (α = 9) 10 SEC. DATA FROM 400M X 400M GRID AT CORDELIA JUNCTION, CA (NOV. 7, 2010) (Source Hoff and Perez, 2011) 67% lower than the RMSEs of the irradiance at each of the individual ground stations. The research to date on solar plant variability demonstrates that smoothing (reduced variability) occurs both within PV plants and for an aggregate of solar power plants within a larger geographic region. Furthermore, forecast error for one individual irradiance sensor will be higher than that for an individual solar energy plant occupying some defined area depending on system size. Forecast error will likewise be reduced for an aggregation of several geographically dispersed solar plants. The relationship between PV power output variability and forecast error is through the impacts that clouds have on ground level insolation. Clouds lead to PV power output variability and similarly make forecasting irradiance more challenging. Improved characterization of these dynamics will help grid operators anticipate and better manage the variable nature of solar energy production. 15 15 Predicting Solar Power Production

16 16 solar electric power association

SECTION 4 4. Using Solar Power Production Predictions for Utility Planning and Operations Table 4 lists the various solar forecast end-users and potential applications of area forecasts and point forecasts. Here we define area forecasts as a composite of two or more geographically dispersed solar plant power predictions, whereas a point solar forecast is a prediction of future output for one specific solar plant. While solar plant predictions will be used in many of the same ways that wind forecasts are being used today, the distributed nature of much of the solar being deployed will result in new end-users and applications. Area forecasts of solar energy power predictions can be produced for entire balancing areas, nodes within wholesale markets, distribution systems, or even feeder-level forecasts within distribution systems depending on end-user requirements. The end-users and the applications of solar forecasting services will be different for each region depending on the mix of central utility-scale solar versus distributed solar, and whether or not electricity markets are unregulated or regulated. In regions with competitive wholesale power markets, predicting intra-hour, intra-day, and day-ahead power production from single or multiple solar power plants may be used for reliability planning and efficient market operations. In those regions with vertically integrated utilities that remain regulated monopolies, forecasts of solar plant power production may be used for reliability planning and scheduling of generation. Distributed solar installations are essentially invisible to grid operators as PV penetrations grow increasingly modify the load that appears on the system. Load serving entities in competitive markets and vertically integrated utilities can use area solar power predictions to forecast system load net of distributed solar production. Having the ability to distinguish at any given moment the load being served by the grid versus the load being met by DG solar provides critical situation awareness for managing distribution systems. 4.1 transmission-level applications of solar energy predictions Experience to date using solar energy power predictions is limited. As mentioned above, the level of solar energy penetration on the nation s grid is relatively small and thus the need for solar forecasting is limited. This is changing rapidly, however, and today there are certain regions that have experienced high levels of solar penetration and thus solar forecasting service providers are gaining new customers. One of the key user groups of variable generation forecasting services are ISOs/RTOs and other balancing authorities. Balancing authorities (BA) coordinate generation and transmission across a wide geographic area, matching generation instantaneously to the demand for electricity. BAs forecast day-ahead load and schedule generation to assure that sufficient generation and reserve capacity is available if demand rises beyond the forecast or a power plant or power line drops offline. In addition to coordinating electric generation and transmission, ISOs/RTOs provide non-discriminatory transmission access, facilitate competition among wholesale electricity suppliers, and conduct regional planning to ensure a reliable grid for the future. According to the ISO/RTO Council, two-thirds of electricity customers in the U.S. are served by ISOs/RTOs; the remaining one-third of electricity customers are within regions served by vertically integrated utilities. These utilities often serve as the balancing authority or are members of regional balancing authorities, and are generally smaller in size relative to the larger regions served by ISOs/RTOs. As discussed above, larger balancing areas provide advantages in terms of managing the variability of wind 17 17 Predicting Solar Power Production

Table 4: Solar Forecast End-Users and Potential Applications User Group ISOs/RTOs & Balancing Authorities Application Day-ahead reliability planning Hour-ahead reliability management Security constrained unit commitment Real time dispatch Load forecasting Ramp event prediction Transmission security planning and outage coordination Distribution Utilities (either serving load as LSE or not serving load in unregulated or regulated markets) Distribution system planning Distribution management systems Outage management systems Smart grid infrastructure management Load forecasting 18 18 Scheduling Coordinator (participant in competitive wholesale market) Scheduling Coordinator (regulated market) Load Serving Entity (participant in competitive wholesale market) Load Serving Entity (regulated markets) Energy Traders Research Labs Project Developers Day-ahead scheduling in competitive markets Hour-ahead scheduling in competitive markets Day-ahead scheduling Hour-ahead scheduling Day-ahead load bids Hour-ahead load bids Day-ahead load forecasts Hour-ahead load forecasts Day-ahead, hour-ahead, and intra-hour bidding strategies Day-ahead, hour-ahead, and intra-hour simulations for variable generation integration studies Day-ahead, hour-ahead, and intra-hour simulations for project pro forma solar electric power association

SECTION 4 and solar generation due to the geographic smoothing of multiple variable generation plants and greater access to a variety of generation resources for system balancing. The California Independent System Operator (CAISO) is the only ISO/RTO that currently subscribes to a centralized solar forecasting service. This is in contrast to the fact that all seven ISOs/RTOs in the U.S. are using centralized wind forecasts to better integrate wind power into system operations and markets. This includes the following ISOs/RTOs along with the year they adopted wind forecasting: MISO/2008, NYISO/2008, PJM/2009, ERCOT/2008, CAISO/2004, ISONE/2013, and the SPP/2011. Although SPP is behind the other ISOs/RTOs in terms of market development, and currently only operate a real-time imbalance market, the wholesale energy markets operated by these regional ISOs/RTOs have seen a considerable degree of convergence in market design in recent years. They share several features in common that are advantageous for integrating variable generation resources such as wind into system and market operations. These include: large balancing authority areas, centralized unit commitment and scheduling, frequent dispatch of resources in real-time (typically every 5 minutes), and locational marginal pricing. Like all balancing authorities, ISOs/RTOs must determine operating reserve requirements to meet reliability standards established by the North American Electric Reliability Corporation and the regional electric reliability organizations. Operating reserves can be categorized into two general categories: regulating reserves and contingency reserves. Regulating reserves are held by a BA to respond immediately to instantaneous mismatches between supply and demand. Generators on automatic generation control (AGC) can move up or down depending on the frequency deviations on the network. Contingency reserves are held to respond to forced outages of power generators or transmission lines do to an unforeseen contingency like a mechanical failure. Increasing penetration of variable generation will likely increase the need for operating reserves; however, the degree to which these requirements will increase is uncertain. Day-ahead and hour-ahead forecasts of variable generation, as well as an understanding of the degree of uncertainty implicit to these forecasts, can be used by balancing authorities in the processes they use to determine the level of operating reserves necessary to meet reliability standards. Forecasts of variable generation can play an important role in market operations. Until recently, wind power scheduling in ISOs/RTOs in the U.S. was conducted on an hour-ahead basis using forecasts provided in that timeframe. In regions with high and increasing levels of variable generation, day-ahead forecasting is becoming increasingly used for scheduling purposes. In most ISOs/RTOs it is optional for wind resources to bid into the day-ahead market. In MISO and PJM, wind facility owners are required to provide day-ahead forecasts of the predicted hourly output if they are considered capacity resources. The MISO implemented a Dispatchable Intermittent Resources (DIR) tariff to improve MISO s remote control over wind facilities. DIR allows MISO to remotely reduce the overall output of a wind facility, instead of simply shutting the wind facility down, to avoid or minimize system congestion. This approach minimizes the amount of wind facility curtailments, reduces transmission line congestion, facilitates lower electric production costs and provides savings to ratepayers whose electric providers purchase electricity through the MISO wholesale market. Botterud et al. (2009) emphasize the importance of integrating wind power forecasts into day-ahead market clearing processes, as wind power will have an increasing impact on the marginal price of electricity. The cost of wind generation should be properly reflected in day-ahead market clearing prices, which is where most of the energy purchased and sold in power markets is settled. Wind forecasts are being used in many regions as part of the real-time market operations and dispatch. For example, the NYISO incorporates its real-time wind forecasts (updated every 15 minutes for the next 15- minute interval looking out 8 hours) into real-time unit economic dispatch. Wind bid curves are submitted 70 minutes before the operating hour and then used in optimized economic dispatch for each 5-minute increment in the operating hour. Essentially market operations as designed in New York allow wind to be treated like any other generating resource on the 19 19 Predicting Solar Power Production

20 20 New York grid. One difference is that conventional generation sets an upper limit of power output for the real-time market, but a wind plant s upper generation limit is set by the 15-minute forecast throughout the operating hour. An active area of research is the development of forecast products that are designed specifically to predict significant ramping events. Significant, shortterm changes in weather can cause rapid changes in the output of variable generation resources. On January 26, 2008 ERCOT experienced a significant ramp event when wind generation output at 2,000 MW began a three and a half-hour ramp down to 360 MW. The down ramp was 2 hours sooner and somewhat faster (8 MW/minute vs. 5 MW/minute) than the dayahead forecast had predicted. More accurate advanced notice of these events would allow grid operators time to prepare and plan for these significant changes in output from variable sources of generation. While various ISOs/RTOs are investigating the use of variable generation forecasting for predicting ramp events, there are currently limited systems being used today. ERCOT deployed a first-of-its-kind wind forecasting tool designed to help system operators prepare for large and sudden changes in wind production in 2010. While BAs are most concerned with area forecasts for the entire fleet of variable generation resources operating within their footprint, scheduling authorities that own or operate variable generation resources may require plant specific day-ahead and hour-ahead forecasts to schedule plant output into wholesale power markets. As an example, Southern California Edison has over 1 gigawatt of installed solar PV within its service territory, most of which is scheduled through the wholesale market mechanism. It procures solar forecasts from several vendors, develops an in-house forecast, and uses the centralized solar forecast provided by CAISO to inform its bidding strategy. It has entered into a number of power purchase agreements (PPAs) to meet the State s renewable energy portfolio (RPS) standard. The power from these resources must be scheduled in the day-ahead and realtime markets to recoup some of the costs of the PPA. Southern California Edison s solar forecasts for each solar plant go out 168 hours ahead rolling over every hour, with updates every 10 minutes for hours two and three and three updates a day for the out hours. Southern California Edison uses solar forecasting to develop bidding strategies to maximize the value of the solar-generated power they own. The Sacramento Municipal Utility District (SMUD) has been conducting research on solar forecasting for several years. SMUD experiences peak load of 3,300 MW and minimum load of approximately 800 MW. They currently have 100 MW of utility-scale solar and 50 MW of DG solar within their service territory. They installed a network of 77 irradiance sensors used in in a year-long process to assess the forecasting accuracy of six different forecast providers. They anticipated using these forecasts to help improve the scheduling of regulation reserves. However, they determined that the uncertainty in the forecasts did not allow for improvements in scheduling regulating reserves. SMUD currently subscribe to a solar forecasting service used primarily for day-ahead scheduling to determine the need for other non-variable sources of generation. While SMUD has begun to consider forecasting DG solar, they are currently not doing so as system impacts of DG solar to date have been minimal. Additional uses of variable generation forecasts include transmission outage planning, forecasting load net of DG solar, and the development of bidding strategies used by energy traders. Ahlstrom et al. (2013) suggest that financial traders may play an increasingly important role in improving variable generation forecasts. These market participants have a greater tolerance for price risk relative to other market participants as traders seek to profit from the price spread in different markets. Thus, financial traders have an economic incentive to obtain the most accurate variable generation forecasts to create a competitive advantage in the energy trading business. solar electric power association

SECTION 4 4.1.1 caiso s approach to using solar power predictions California is the leading solar state in the country with a total of over 4 GW of installed solar PV capacity to date, with projections of 12 GW to meet the State s 33% RPS goal by 2020 according to the CAISO. Approximately one half of the installed solar in California is distributed (e.g., not registered resources within the CAISO system). Out of necessity, the CAISO is on the leading edge of solar integration in the U.S. The CAISO began subscribing to a centralized solar forecasting service for large utility-scale plants in mid-2011 provided by AWS Truepower. The cost of the solar forecasting service is spread across all market participants. The solar forecasts provided to CAISO are deterministic, which means they do not consider a probabilistic range of likely solar energy production. The solar power prediction system uses a forecasting approach that employs the direct use of both satellitebased data (to determine the current location and movement of clouds) and recent measurements from the solar generation facility, as well as the output data from NWP models. Separate statistical prediction equations are used for each forecast look-ahead hour. Thus, the statistical training schemes are able to select the variables that have the greatest predictive information for each look-ahead hour. Typically, this approach heavily weights the recent measurements from the generation facility and the satellite-based data will receive greater weight for shorter look-ahead periods (for example, 1-6 hours) and put more weight on the NWP data for longer forecast time horizons (i.e. greater than 6 hours ahead). Figures 11 and 12 present data from the CAISO for Monday August 30, 2013, including the day-ahead (DA) forecast, hour-ahead (AH) forecast, and the actual measured output from solar power plants within northern and southern California respectively. Figures 13 and 14 present the same comparisons for Wednesday, December 18, 2013. The August day was likely a clear day with limited cloud impacts resulting in a close convergence between the forecasts and the actual output. In contrast, the December day was likely characterized by clouds impacting solar plant output and thus greater errors are visible between the forecast values and the actual output from the fleet of solar plants located in both northern and southern California. Solar forecasting is currently used for system operations, unit commitment decisions, and the CAISO s Participating Intermittent Resource Program. CAISO operates a Renewables Desk and is in the process of implementing a Ramp Forecast tool developed at the Pacific Northwest National Laboratory. The renewable desk receives variable generation forecasts. Using the hour-ahead forecasts analysts working at the renewable desk tweak the hour-ahead forecast in real-time based on current weather conditions providing advisories to generation dispatch to go up or down in the 0 2 hour timeframe. To facilitate renewable resources participation in its markets, CAISO developed the participating intermittent resource program (PIRP). The program provides centralized wind and solar power production forecasts to its participants on a daily and hourly basis. In the real-time market, PIRP resources must submit hourly self-schedules to match the projected values developed by the program s forecasting service. Monthly deviations of the actual delivery of energy from the scheduled energy are settled based on locational marginal prices. Renewable resources are also allowed to bid into the energy markets if they choose, but this precludes participation in the PIRP for that bidding period. Figures 15 and 16 illustrate the CAISO day-ahead and real-time market schedules respectively, along with the timing of when the centralized solar forecasts are published for all market participants to access. The centralized day-ahead solar forecast is published at 5:30 am and the day-ahead market closes at 10:00 am for the next operating day, which begins at midnight. The centralized hour-ahead forecast is delivered 105 minutes before the operating hour and the hour-ahead scheduling and bids must be submitted 75 minutes before the operating hour. CAISO is planning to move to a 5-minute real-time economic dispatch of wind and solar modeled after the NYISO in the coming years. Starting on April 1, 2014 variable sources of 21 21 Predicting Solar Power Production

Figure 11: Forecast Solar vs. Actual Output, CA NP15 Figure 12: Forecast Solar vs. Actual Output, CA SP15 MW 450 400 350 300 250 200 150 100 50 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 HOUR MW 900 800 700 600 500 400 300 200 100 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 HOUR DA FORECAST HA FORECAST ACTUAL DA FORECAST HA FORECAST ACTUAL Figure 13: Forecast Solar vs. Actual Output, CA NP15 Figure 14: Forecast Solar vs. Actual Output, CA SP15 700 1200 MW 600 500 400 300 200 100 MW 1000 800 600 400 200 22 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 DA FORECAST HA FORECAST ACTUAL HOUR 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 DA FORECAST HA FORECAST ACTUAL HOUR 22 generation will be able to bid at 15-minute intervals with the implementation of Federal Energy Regulatory Commission (FERC) Order No. 764, which requires ISOs/RTOs to offer intra-hourly transmission scheduling. With financial support from the California Energy Commission, the CAISO is working with Clean Power Research on forecasting DG solar production using its SolarAnywhere FleetView service. Behind the meter solar resource forecasts are not currently being used for operational purposes by the CAISO; in the future DG solar forecasts can provide value for planning and to more effectively schedule non-solar generation units through more accurate load forecasting net of DG solar generation. The CAISO is using irradiance forecasts to project into the future the impact on system load as additional sources of variable generation are added to meet California s goal of 33% renewable generation by 2020 goal. Figure 17 presents a projection of future load net of variable generation, indicating the potential of having surplus generation on the network in shoulder months when demand is relatively low and variable generation production is high. The so-called duck solar electric power association

SECTION 4 Figure 15: CAISO Day-Ahead Market Timeline Figure 16: CAISO Hour-Ahead Market Timeline 12:00 AM DAY-AHEAD MARKET CLOSES 10:00 AM DAY-AHEAD MARKET RESULTS PUBLISHED 1:00 PM 12:00 PM OPERATING DAY BEGINS 12:00 AM OPERATING DAY BEGINS 12:00 AM HOUR- AHEAD MARKET CLOSES OH-75 MIN. OPERATING HOUR (OH) 5:30 AM DAY-AHEAD SOLAR FORECAST PUBLISHED 1:00 PM REAL TIME MARKET FOR NEXT OPERATING DAY OPENS 1:00 PM REAL TIME MARKET FOR NEXT OPERATING DAY OPENS OH-105 MIN. HOUR- AHEAD SOLAR FORECAST PUBLISHED Figure 17: CAISO Net Load 2012-2020 MW (IN THOUSANDS) NET LOAD MARCH 31 28 26 24 22 20 B 18 16 14 A 12 10 12AM 3AM 6AM 9AM 12PM 3PM 6PM 9PM 12AM HOUR A B OVERGENERATION RISK RAMP NEED ~13,000 MW IN THREE HOURS 2012 (ACTUAL) 2013 (ACTUAL) 2014 2015 2016 2017 2018 2019 2020 (Source: http://www.caiso.com/documents/ FlexibleResourcesHelpRenewables_FastFacts.pdf) 23 chart illustrates how over time solar and other sources of variable generation will change the shape of CAISO s system load, causing much increased ramping needs and the potential of having excess generation. This analysis is helping the CAISO to develop strategies that will increase the flexibility of the system to respond to these future conditions and the potential negative impacts on system reliability. The operational strategies and market reforms initiated in the past decade to leverage wind forecasts to more fully integrate wind resources into existing markets and grid operations are directly applicable to large, utilityscale solar plants. Unlike wind, which is generally a central utility-scale generation resource, the majority of the solar deployed today is at the distribution level. Next we consider the uses of forecasts of solar plant power predictions at the distribution level. 23 Predicting Solar Power Production

4.2 distribution-level applications of solar energy predictions 24 24 Solar energy is forging a unique path at the distribution level with forecasting research and integration. In most regions, the levels of DG solar penetration are such that the overall impacts on system operations and reliability are minimal. The variability of DG solar in most regions is well within what grid operators are accustomed to managing given the typical hourly variability associated with demand. However, there are some limited areas where DG solar penetration has reached a level that has created reliability concerns, and in some cases certain feeders have been closed to additional DG solar deployment. Utility companies in states with high solar penetration, including California, Hawaii, Arizona, and New Jersey, are working to address the unique challenges of forecasting DG solar. However, there are currently no fully operational systems that utilize DG solar forecasts for distribution planning and system operations. This is an active area of active research and development that should yield new approaches and applications in the coming years. Forecasting DG solar faces significant challenges due primarily to a lack of DG system performance data. Solar energy systems deployed within the distribution system are mostly invisible to grid operators and utility companies, meaning that in most cases they do not have access to real-time or historic energy production data from these systems. This fact, in addition to the limited amount of information on the technical specifications of DG solar installations, creates challenges for forecast providers. Furthermore, grid operators typically do not even have nameplate capacity and locational data associated with PV installations that are not participating in their markets. Through net metering and incentive applications, data on modules, inverters, and perhaps array angles and orientations are often collected for hundreds if not thousands of DG systems. The challenge for forecast providers is getting access to these technical specifications to integrate into their solar forecasting systems. As mentioned above, Clean Power Research is pioneering efforts to utilize DG solar specifications in California to produce DG solar forecasts. Most states require solar installers to report system characteristics as part of the net metering and/or interconnection application process. These reporting requirements should be standardized and consolidated into accessible databases for use by the forecasting community. Without performance data, forecasts accuracy can never be completely validated. DG solar forecast validations through statistical sampling of DG solar systems could prove to enhance the quality of DG solar forecasts in the future. Historic PV system performance data is often used to train forecasting models to reduce errors over time, although historical irradiance data can be used for this purpose as well. Additionally, real-time telemetry (both PV power production and meteorological data) is typically needed for shorterterm forecasting (intra-day or intra-hour), which becomes increasingly critical at higher penetration levels. Thus, DG solar represents a new and complex challenge for forecast providers and for end-users seeking to acquire the most accurate DG solar forecasts available. Several organizations are addressing these particular challenges, which will likely yield new solutions in the coming years. The demand-side nature of DG solar shows up through shifting load patterns. While utilities have decades of experience with load forecasting, DG solar results in a new element of variability and uncertainty to system load not captured by existing load forecasting methods. This is a new paradigm for utilities and much work is needed to integrate solar forecasting into existing load forecasting practices. The consistent patterns of load seen over time are changing due to DG solar and thus load forecasting techniques will need to be modified in response, including greater emphasis on intra-day and intra-hour load forecasting. Utilities and grid operators need to distinguish between the load being supplied through the grid relative to that being met by DG solar to provide situational awareness. This is critical to proactively prepare for the net load impacts of DG solar variability resulting from changing weather patterns and emerging cloud formations.. Concern is growing regarding the IEEE 1547 model interconnection standard. This standard requires solar electric power association

SECTION 4 distributed generation systems to trip off during voltage or frequency disturbance events that occur due to line faults or other factors. In the case of distributed solar, a frequency disturbance event that causes large numbers of systems to drop off, it will appear to system operators that system load suddenly increased and lead to further voltage reductions. These types of events could cascade and cause a negative impact on bulk system reliability. Forecasts and nowcasts of DG solar can provide situational awareness of how much solar power that may be vulnerable to such widespread tripping, and would allow for a rapid response to such disturbances. A longer-term solution to this technical challenge is required, including allowing DG units to ride through low voltage events. With relatively simple adjustments, modern grid-tied inverters could provide this functionality. Forecasting of DG solar can have value as an input into existing distribution management and planning tools. Integrating DG solar simulations in load flow models can help identify potential problem areas within the distribution network. Eventually, forecasts of DG solar output can be integrated into distribution and outage management tools, again to efficiently manage distribution assets and plan for future investments. An emerging use of variable generation forecasting is for the optimization of smart grid systems. Papavasiliou and Oren (2008) explore the possibility of directly coupling deferrable loads with wind generators in order to mitigate the variability and randomness of wind power generation. End use loads could participate in contractual agreements to defer their demand for power by a fixed amount of time and wind generators optimally allocate available wind power with the objective of minimizing the cost of unscheduled and variable supply. Letendre and Perotti (2012) explore the economic value of linking renewable energy production with electric vehicle charging, providing a premium green charging service. In both of these potential applications, variable generation forecasts would be a key input to developing optimal strategies for managing demand response resources and vehicle charging with the predicted output from wind and solar generators. Perez et al. (2013) investigate the cost associated with mitigating the variability of solar PV using buffer storage. They find that accurate forecasts of solar plant power output can effectively lower the mitigation cost by nearly 40%. The use of variable generation forecasts to optimize the use of smart grid systems designed to mitigate integration costs will likely be an active area of research and development in the coming decade. Smart grid solutions could potentially be integrated with smart inverter technology to allow grid operators to curtail DG solar when necessary or actively manage loads to address reliability concerns associated with the variable nature of solar production. 25 4.2.1 Case Study: Hawaiian Electric Company The Hawaiian Electric Company (HECO) and its subsidiaries serve 95% of the state s 1.2 million residents on the islands of Oahu, Maui, Hawaii Island, Lanai and Molokai. Each of the five islands served have small, (relative to large synchronous interconnected systems in the continental U.S.) independent power grids, which creates a unique challenge with integrating variable sources of generation. The relative high cost of electricity and the State s aggressive 40% RPS goal by 2030 have resulted in explosive growth in solar deployment within HECO s service territory in recent years. HECO has the highest penetration of DG solar than any other utility company in the U.S. The most populated island of Oahu has a peak load of 1,200 MW has an estimated 212 MW of DG solar and the Hawaiian Island with a peak load of 180 MW has an estimated 37 MW of DG solar representing a penetration rate of 17% and 20% respectively. HECO is the first major U.S. power company experiencing significant impacts from variable generation sources including DG solar. Figure 18 identifies the challenges facing HECO in managing the rapid growth of variable sources of generation within its service territory. DG solar has already had a significant impact on the daytime loads served by HECO. The drop off in solar due to the setting sun 25 Predicting Solar Power Production

Figure 18: Distribution Impacts of DG Solar HECO C MW A B A B EXCESS ENERGY CURTAILED SCADA, FEEDER CAPS OR NEW MONITORING TO ACCOUNT FOR DG CONTRIBUTIONS MIN-MAX OUTPUT OF DISPATCHABLE AND BASE GENERATING UNITS OUTPUT OF INTERMITTENT UNITS TYPICAL HAWAII LOAD PROFILE EVENING PEAKING 12 AM 12 PM 12 AM CURTAILMENT SUPPLY EXCEEDS DEMAND AT MINIMUM LOADS MASKED LOADS DG (PV) CHANGING SYSTEM DAYTIME LOAD SHAPE RELIABILITY SUPPLY DOES NOT DEMAND AT MINIMUM LOADS C ENERGY SHORTFALL MUST BE MET BY FAST START UNITS (Sources: http://www.calsolarresearch.org/images/stories/documents/sol1_funded_ proj_docs/smud/smud-heco-cpuc_tsk4_20apr2012.pdf) Figure 19: Impact on System Load for Hawaiian Island Figure 20: Prototype of Variable Generation Display for System Operators 180 170 SYSTEM LOAD (MW) 160 150 140 130 120 110 100 A B 90 0:00 1:26 2:52 4:19 5:45 7:12 8:38 10:04 11:31 12:57 14:24 15:50 17:16 18:43 20:09 21:37 23:02 0:28 1:55 26 A B 2011 DIFFERENCE 2008 DIFFERENCE SATURDAY 6/18/2011 SUNDAY 6/19/2011 SUNDAY 6/15/2008 SATURDAY 6/21/2008 (Source: http://www.calsolarresearch.org/images/stories/documents/sol1_ funded_proj_docs/smud/smud-heco-cpuc_tsk4_20apr2012.pdf) 26 in combination with the increase in system load, are resulting in significant ramping requirements in the early evening hours. Figure 19 illustrates how daytime system load on weekend days has shifted between 2008 and 2011, clearly illustrating the impact that DG solar is having on the load profile for one of the Hawaiian Islands. While the evening peak load remains constant between 2008 and 2011, the afternoon load has dropped significantly over this time period. (Sources: http://www.ct-si.org/events/apce2013/program/pdf/john%20zack.pdf) solar electric power association

SECTION 4 With funding from several sources HECO is working on a variety of issues related to high-penetration of PV. This includes the deployment of a targeted sensor network and the development and evaluation of a prototype forecasting system called SWIFT (Solar and Wind Integration Forecast Tool). The primary purpose of SWIFT is to equip system operators with timely and accurate foreknowledge of hour-ahead and dayahead wind and solar-based generation so that they can readily accommodate the increasing penetration of this type of generation while economically balancing other resources and maintaining high levels of grid reliability. SWIFT incorporates state-of-the-art physics-based and statistical forecasting techniques with innovative adaptations for Hawaiian weather regimes while developing, through feedback from utility operators, an ongoing analysis of system performance. Solar forecasting is performed sub-hourly for utility-scale and DG solar down to individual substations. This is one of only a handful of projects that are specifically working to develop tools that integrate variable generation forecasts into a distribution system energy management system (EMS). Figure 20 illustrates a prototype display of wind and solar forecasts for system operators developed by AWS Truepower. One of the key challenges is to create a visualization of the variable generation forecasts in a format that provides situational awareness and decision support for operations. The goals is to design the display and format in a way that communicates a wide range of forecast information, including most probable power production, variability and ramp rates, uncertainty and geographic patterns. Furthermore, system operators will require training to understand the information being presented and how best to respond to the information. System operators need to understand that forecasts provide value even if not 100% accurate, that forecast performance varies with the weather regime, and how to track weather patterns that cause large changes in variable generation output. 4.2.2 Case Study: Tucson Electric Power Tucson Electric Power (TEP) is the second-largest investor-owned utility in Arizona providing service to over 405,000 customers in the Tucson metropolitan area. With more than 2,260 MW of net generating capacity (primarily coal-fired), TEP supplies most of the power it distributes, and it also sells energy wholesale to utilities and power marketers in the western U.S. In addition, TEP serves as the designated balancing authority covering its service territory that spans 1,155 square miles. With a minimum winter load of 1,000 and an estimated 200 MW of solar PV, equally split between utility-scale and DG solar, during certain months solar variability is beginning to impact system operations. TEP is working with researchers at the University of Arizona to develop a solar forecasting system that will eventually be integrated into the company s energy management system. They currently have a prototype solar forecasting system that provides system operators with a web interface (see Figure 21) covering both wind and solar resources within the TEP service territory. The system integrates both utility-scale solar Figure 21: Prototype Web Interface for Variable Generation TEP Service Territory (Source: https://www.tep.com/doc/planning/2014-irp-real-time-pv-power- Forecasting(Holmgren_UofA).pdf) plants forecast with DG solar forecasts updated on a 1-minute frequency out for 3 days ahead. The separate web interface is currently being used by TEP system 27 27 Predicting Solar Power Production

operators in an advisory capacity, allowing adjustments to hour-ahead generation schedules and reserve capacity requirements given anticipated variability in wind and solar generating resources. TEP plans to eventually integrate the variable generation forecasting system into their energy management system, although the timing of this is uncertain. The novel solar forecasting approach developed by researchers at the University of Arizona uses 80 residential rooftop PV systems distributed over a 50 km x 50 km area as irradiance sensors. This approach to solar PV forecasting relies on power measurements of PV systems as a measure of irradiance, which eliminates the need for a dedicated network of irradiance sensors and automatically accounts for the effect of temperature and angle of incidence on power forecasts (Lonij et al., 2013). Measurements from the 80 DG solar PV systems are recorded at 15-min intervals, with each measurement representing the average AC power over the previous 15 minutes. Each of the PV systems used to forecast solar output for all PV systems within the TEP service territory use an inverter by SMA Solar Technology with a data communications card installed to record data, which is transmitted over the internet using an SMA Sunny WebBox. A clear-sky expectation for the output of each system is obtained; deviations from the clear-sky expected output are used to infer cloud impacts of system performance. The output from the sample is scaled-up to represent the output of all utility-scale and DG solar within TEP s service territory. University of Arizona researchers conclude that their approach using 15-min interval measurements from a network of distributed PV systems outperform the persistence model for forecast horizons ranging from 30 minutes up to 90 minutes. This is an improvement over forecasts based on satellite images, which may not available at 15 minute intervals for these short-term forecasts horizons (Lonij et al., 2013). 28 28 solar electric power association

SECTION 5 5. Solar Forecast Provider Survey Results Solar forecasting is typically one of many services a company provides. Many companies offering solar forecasting also provide wind forecasting services and a host of additional services supporting renewable energy project planning, development, and operation, while others provide a variety of weather forecastrelated services. For example, AWS Truepower based in Albany, NY also provides regional resource maps based on mesoscale models for site screening and early-stage feasibility studies for wind projects. Global Weather Corporation also offers web-based data services to weather-sensitive industries such as energy, transportation, agriculture and digital media. Seventeen companies (see Table 5) were identified as providing solar forecasting services and invited to complete an online survey describing the services they currently offer. Thirteen companies completed the survey. Table 5: Solar Forecasting Service Providers Invited to Complete Online Survey Company Name Country Completed Survey Web Address Clean Power Research U.S. http://www.cleanpower.com/ 3 Tier U.S. http://www.3tier.com/en/ AWS TruePower U.S. http://www.awstruepower.com/ JHtech--Solar Data Warehouse U.S. http://www.solardatawarehouse.com/ Windlogic, Inc. U.S. http://www.windlogics.com/ Global Weather Corporation U.S. http://www.globalweathercorp.com/# Green Power Labs Canada http://www.greenpowerlabs.com/ Iberdrola Renewables, LLC Spain http://iberdrolarenewables.us/ 29 irsolav Spain http://irsolav.com/ Meteologica SA Spain http://www.meteologica.com/meteologica/ NNERGIX Energy Management, S.L. Spain http://www.nnergix.com/ Meteocontrol Energy and Weather Services Germany http://www.energymeteo.com/en/ WEPROG Germany http://www.weprog.com/ Reuniwatt France http://reuniwatt.com/en/ ENFOR A/S Denmark http://www.enfor.dk/ Datameteo Italy http://www.datameteo.com/meteo/ DNV-GL (Garrad Hassan, Inc.) Norway http://www.dnvgl.com/ 29 29 Predicting Solar Power Production

Table 6: Forecast Providers Active in Each Area Transmission Grid Operations Distribution Grid Operations Research & Development Solar Energy Scheduling Solar Energy Market Bidding Centralized PV Plant Output Distributed PV Plant Output New Plant Siting Analyses Regional Transmission Organizations Utility Companies Solar Plant Owners/ Operators Universities/ Laboratories Other (please describe below) 8 3 5 7 4 7 5 2 5 6 5 8 8 6 8 3 1 1 2 5 4 5 2 3 1 0 8 0 0 3 3 0 0 0 0 1 2 1 2 0 Activity Levels: Low Moderate High 30 30 30 Figure 22: Spatial Resolution Capability of Solar Forecast Services NUMBER OF RESPONDENTS 9 8 7 6 5 4 3 2 1 0 <1 KM 1 KM 5 KM 8 KM 10 KM SPATIAL RESOLUTION Twelve of the survey respondents indicated that they currently offer a solar forecast service, with one planning to do so within the year. Of the respondents surveyed, ten currently offer wind forecasting services and three do not. Seven of the ten have been providing wind forecasting services for five years or more. In contrast, one of the respondents has been providing solar forecasts for over four years, five for three to four years, three for two to three years, two for one to two years, and one for one year or less. This suggests that solar forecasting services is less mature than wind forecasting services, which is likely due to the fact that wind represents a much larger global power resource today than solar. Survey respondents were asked to indicate the user groups and applications for the solar forecasting services they provide; Table 5 indicates the results. Based on these results, grid operators and utilities are the major end users for solar forecasting services for grid operations and scheduling applications. Solar forecasting services are also being used by these organizations, and by university research scientists, for research and development purposes. Finally, solar plant owners/operators are using solar forecasting services for scheduling and energy market bidding purposes. All survey respondents currently offering solar forecasting services provide day-ahead and hour-ahead forecasts, and all but one offers intra-hour forecasts. Survey respondents were asked to describe the time intervals for their forecasts and how often they are updated. Day-ahead forecasts are often provided in hourly time intervals for the next seven days, and are solar electric power association

SECTION 5 updated hourly or two to four times per day. Survey respondents indicated that hour-ahead forecasts can be updated hourly, every 15 minutes or every 5 minutes depending on the specifications of the client. There was significant variability in the time intervals for intrahour solar forecast from 1 minute up to 30 minute time intervals were reported, with 5, 10, and 15 minute intra-hour intervals being common. Forecasts can be delivered to clients minutes after the computer runs have been completed. There were many comments suggesting that forecast providers have the capacity to deliver a tailored solar forecast to meet a client s particular needs. Respondents were asked to describe the models used to develop their solar forecasts for each time horizon. A hybrid approach to developing solar forecasts seems to be the norm based on survey responses. Each company develops a unique approach using physical models including NWP and satellite cloud images coupled with post-processing and other statistical techniques to improve forecast accuracy. The approach of a few respondents seems to lean more toward a purely statistical approach. Survey respondents were asked to indicate the spatial resolution capability for their solar forecasting services; Figure 22 presents the results. Three respondents have the capability to produce a forecast for a spatial resolution less than one kilometer, while most can provide a forecast based on one, five and ten kilometer resolutions. Comments from those selecting the other category suggest that the spatial resolution can vary depending on the NWP model used and the geographic location. Some parts of the globe do not have the level of spatial resolution these models offer in the U.S. and Europe for example. Ten of the thirteen survey respondents indicated that they do not publish error metrics for their solar forecasting services. One of the three respondents who does publish error metrics reported a root square mean error (RMSE) of 10% or less in central Europe and another indicated that it reports a mean absolute error (MAE) during calm months of 16% for one hour-ahead and 25% for three hours ahead. On the most variable months the MAE is 27% for one hour ahead and 44% for three hours ahead. As discussed above, it is difficult to interpret these results without information on the basis from which these metrics were calculated. One survey respondent stated that forecast metrics are too site and data-type specific to publish blanket numbers, but they will provide regional numbers upon request. Ten of the thirteen survey respondents indicated that they price their forecasting service as a recurring subscription (e.g., monthly or annually). Five indicated that the pricing can be based on a fixed price for defined services, while six respondents indicated that they engage in custom negotiated pricing based on service needs. Survey respondents were not asked to indicate what they charge for their services as it would be difficult to compare pricing without understanding the specifications of the solar forecasts being provided. Survey respondents were asked to indicate what they require from their clients to produce a solar power forecast. Table 6 contains a tabulation of the survey results for this question. Some of the other information requirements indicated by survey respondents include historical irradiance measures and solar plant output and array angles and orientations. It should be noted here that area forecasts of distributed solar plants will require this information for hundreds if not thousands of roof top and other distributed solar energy systems. Most state net metering and incentive programs require that this information be provided, but it may not be in a central database for use by solar forecasters as inputs to their forecasting systems. Table 7: Solar Forecasting Providers Information Needs Required Information Number of Responses Solar Plant(s) Location(s) 13 Solar Plant(s) Name-Plate Capacity Plant Type (fixed, single axis or duel axis tracker) Module Specifications 4 Inverter Specifications 4 Shading Analysis 3 Other 5 12 12 31 31 31 Predicting Solar Power Production

Survey respondents were asked to identify challenges working with utility and grid operators and how these challenges could be addressed. No strong trend emerged from the survey respondents. Several respondents indicated challenges getting good quality data from their clients, including data on the solar energy plants being modeled, historic production from solar energy plants, and the location of distributed solar systems within their service territories. Challenges in terms of getting the needed data to flow seamlessly into the forecasting models were also mentioned by several survey respondents. Several comments indicated that effective communication in terms of needs and the value of solar forecasting is a challenge. The solutions to these challenges mentioned include creating industry standards, providing an open forum for end users to interact with forecast providers to discuss needs, better information exchange tools and investments in the necessary IT systems. Finally, survey respondents were asked to describe the solar forecasting system improvements they plan to implement in the next two years. Virtually all respondents indicated ongoing research and development efforts to improve the quality of their forecasting services. Several specific initiatives mentioned by respondents include enhancement of DG forecasting services, creation of tools to allow grid operators to make better use of probabilistic solar power forecasts, and emphasis on improving very short-term nowcasting services. 32 32 32 solar electric power association

SECTION 6 6. Future Developments and Recommendations The field of solar power forecasting is dynamic with a number of significant research projects underway and several initiatives to share research and current practices. Over the past decade, the California Public Utilities Commission s (CPUC s) California Solar Initiative (CSI) provides funding to several organizations to advance the state of the art in solar forecasting through its research and development program. The U.S. Department of Energy s (DOE) SunShot program provides funding for solar forecasting research through its focus area on high penetration PV. The CPUC and the DOE hosted two high penetration PV forums, one in 2011 and the other in 2013. Each forum included presentations by experts in the arena of solar forecasting. In addition, The Utility Variable-Generation Integration Group (UVIG), previously known as the Utility Wind Integration Group (UWIG), is a membership-based organization with the mission to accelerate the development and application of good engineering and operational practices supporting the appropriate integration and reliable operation of variable generation on the electric power system. UVIG has been hosting regular workshops on variable generation forecasting, with the latest forecasting workshop taking place in late February of 2014 in Tucson, AZ. While much progress has been made in recent years advancing the state of the art in solar forecasting and creating greater awareness among various stakeholder groups on solar energy forecasting, the industry is still in the early stages of development. More work is required to engage stakeholders to clarify industry needs, develop standards, and overcome barriers to greater use of solar forecasting including increasing confidence in forecast accuracy among end-users. Robust and highly reliable solar forecasting services are essential to address one of the key barriers to expanding solar deployment solar power variability. 6.1 Current solar forecasting research projects One of the critical factors to improving solar forecasting accuracy is through enhanced treatment of clouds, which is the key variable contributing to solar variability. In addition, better techniques are needed for short-term 0 30 minutes (nowcasting) as this was identified through expert interviews as a key deficiency in current solar forecasting approaches. More accurate short-term solar forecasts will give grid operators the confidence to be able to address short-term ramps in solar facility generation. Finally, improved models and approaches to accurately predict solar generation from distributed, behind the meter solar energy systems is needed. In 2012 the U.S. DOE allocated $9 million in funding over three years for improving the accuracy of solar forecasting through the SunShot Systems Integration efforts. Two projects were selected for funding Predicting Solar Power Production leveraging the expertise of university research centers, government laboratories, and private firms with expertise to help utilities and grid operators better forecast when, where, and how much solar power will be produced at U.S. solar energy plants. According to the U.S. DOE web site, these solar forecasting projects will allow power system operators to integrate more solar energy into the electricity grid, and ensure the economic and reliable delivery of renewable energy to American families and businesses. One of the projects lead by the University Corporation for Atmospheric Research (UCAR) is focused on improving physical models used in solar forecasting particularly the treatment of clouds within these models. The second project lead by IBM is focused on improving statistical forecasting approaches using super computers and deep machine learning techniques. The two projects overlap to some degree as an initial effort is underway 33 33 33

34 34 34 to characterize solar forecasting error metrics. In February of 2014 in advance of the UVIG meeting, the U.S. Department of Energy hosted a workshop to discuss an initial proposal for solar forecasting error metrics based on collaborative research. The UCAR-lead U.S. DOE-funded project is titled A Public-Private-Academic Partnership to Advance Solar Power Forecasting. The project participants include: National Renewable Energy Laboratory, Brookhaven National Laboratory, Penn State University, Colorado State University, University of Hawaii, University of Washington, University of Central Florida, Long Island Power and Light, Public Service of Colorado, Sacramento Municipal Utility District, Hawaiian Electric System, New York Power Authority, NYISO, Xcel Energy, Telvent DTN, Atmospheric and Environmental Research, Global Weather Corporation, MDA Information Systems, and the National Oceanic and Atmospheric Administration. The National Center for Atmospheric Research (NCAR) scientist Dr. Sue Ellen Haupt is serving as the principal investigator on the project. The NCAR-lead project team will advance methods for solar radiation measurement and cloud observation and tracking techniques, methods to quantify and track aerosols that affect cloud formation and radiative transfer, including the prediction of aerosols, haze, and contrails, short-term prediction of cloud properties based on observations, and nowcasting techniques. According to Dr. Haupt, a major goal of the project is to improve the way clouds are dealt with in various forecasting models. Specifically, researchers will be upgrading WRF (Weather Resource and Forecasting Model) creating a WRF Solar with a new physics package that improves cloud modeling. The goal is to produce a seamless approach to solar forecasting that improves the physical models used for each forecast horizon. The second U.S. DOE funded solar forecasting project is housed at the IBM Thomas J. Watson Research Center and is titled Watt-Sun: A Multi-Scale, Multi-Model, Machine-Learning Solar Forecasting Technology. The project manager is Dr. Hendrik Hamann. Project partners include Argonne National Laboratory, Arizona Research Institute for Solar Energy, National Renewable Energy Laboratory, Northrop Grumman, Northeastern University, Green Mountain Power, Tucson Electric Power, ISO-New England, Juwi Solar, NOAA, and Petra Solar. Based on information published on the U.S. DOE web site, IMB and its partners will integrate big data processing and cloud modeling into a universal platform that combines different prediction models and uses state-of-the-art machine learning technologies to improve the accuracy of predictions. Similar to the recently demonstrated IBM Watson computer system, the proposed Watt-sun technology will leverage deep machine learning and self-adjusting voting algorithms to decide between various forecasting models and expert systems. The CPUC s CSI research and development program provided funding for two current solar forecastingrelated projects as part of its third solicitation. The Integrating PV into Utility Planning and Operation Tools project led by Clean Power Research (CPR) includes validation and integration of PV fleet simulation tools that will enable utilities and ISOs/ RTOs to cost-effectively integrate distributed PV resources into their planning, scheduling and operating strategies. Project partners include CAISO, University of California San Diego (UCSD), Electric Power Research Institute, State University of New York, and SEPA. This project will build on CPR s CSI Grant Solicitation #1 award titled, Advanced Modeling and Verification for High Penetration PV, which was granted in 2010. The CPR-led project will extend the SolarAnywhere enhanced resolution solar resource database to create a high resolution (1 km, one-minute resolution) database, and benchmarking of data accuracy. This source of high resolution data will be used to forecast PV fleet performance. The project will also validate previously developed PV fleet simulation methodologies using measured ground data from fleets of PV systems connected to California s grid. Finally the project will integrate PV fleet simulation methodologies, which are powered by the high resolution solar resource database, into utility software tools for use in activities ranging from distribution planning to balancing area operations. The tools and data streams developed as part of this work will be made available to California utilities, ISOs/RTOs and others to help cost-effectively and reliably integrate distributed PV into the grid. solar electric power association

SECTION 6 The UCSD is leading a second CSI-funded project titled, High-Fidelity Solar Forecasting Demonstration for Grid Integration. Project partners include Green Power Labs, Clean Power Research, Southern California Edison, Sacramento Municipal Utility District, National Renewable Energy Laboratory, and Power Analytics. This project seeks to demonstrate that solar resource forecasting can be the most costeffective strategy for integrating large amounts of PV generation into distribution systems. Initially, historical aggregate PV ramp rates will be analyzed and a baseline of solar forecast performance under the extreme ramps will be created. Additionally, sites for additional measurements to improve solar forecast accuracy will be proposed based on statistical analysis. For resource adequacy applications, the UCSD team will then improve and demonstrate forecasts for the marine layer meteorological conditions that affect a large fraction of solar PV in San Diego Gas and Electric territory on summer mornings. High resolution NWP and statistical models will be developed and applied to improve forecast accuracy. Local applications of solar forecasting using sky imagery will also be demonstrated on five typical feeders with variations in PV penetration, location/meteorology, and voltage regulation equipment. On these feeders, fast demand response potential based on demand and solar forecasting and dynamic loading will be demonstrated. Through increased granularity in modeling distribution feeder voltages and their reaction to fluctuations in solar PV output, geographic diversity is expected to reduce previously observed impacts at high solar PV penetration. There are a number of additional research and development efforts underway to support high penetration of PV, including the work at HECO discussed above and work at the Sacramento Municipal Utility One CSI project includes collaboration between SMUD and HECO conducting a research, demonstration, and deployment project that targets testing and development of hardware and software for high-penetration PV. This effort is intended to address key grid integration and operational barriers that hinder larger-scale PV adoption into mainstream operations and onto the distribution grid. As part of this effort, SMUD and HECO have identified respective case studies at specific locations in California and Hawaii, solar assessment and forecasting needs, and PV grid integration and visualization needs. Current and future solar forecasting research and development projects will likely improve the quality of solar forecasting services as new techniques and innovations are ultimately integrated into commercial provider forecasting services. One of the experts interviewed for this project indicated that wind forecasting accuracies increased considerably over time from these types of research and development programs, with errors falling from initial values of 20% to 12% over a five-year period. Solar forecasting services today, however, can create value for grid operators and utilities at current levels of forecast accuracy, while future improvements will enhance the value that solar forecasting brings to solar grid integration efforts. 35 6.2 Conclusions and Recommendations Solar energy deployment is accelerating in the U.S. and globally. Today, the U.S. has an installed base of solar energy generation systems totaling over 10 GW. It is anticipated that solar deployment will continue to expand exponentially in the coming decade. The U.S. DOE s SunShot Vision Study finds that cumulative installations of approximately 302 GW of PV and 28 GW of CSP by 2030, and 632 GW of PV and 83 GW of CSP by 2050 are possible if the SunShot price targets are realized (U.S. DOE, 2012). Predicting solar energy Predicting Solar Power Production system output will be an essential tool to accommodate these potential future levels of solar penetration. While much progress has been made in the past decade on developing solar forecasting tools, much additional effort will be required to increase forecast accuracy, integrate forecasting into utility planning and operations, and expand market acceptance. The research and interviews conducted for this report suggest several initiatives that should be undertaken to accelerate the adoption of solar forecasting in response 35 35

36 36 36 to the anticipated growth in solar energy system deployment in the coming decade. These initiatives can be grouped into four broad categories: development of standards/guidelines, economic assessment of solar forecasting value, distributed solar working group and workshops, and expanded engagement of policymakers and regulators. Development of Standards/Guidelines The solar forecasting industry lacks a set of standards or industry guidelines, which acts as a barrier to greater use and understanding of the value of solar forecasting to utility planning and operations. Guidelines or standards are needed to define forecast time horizons relevant to utility scheduling and wholesale power market operations. In addition, as discussed above, there are no current standards for reporting solar forecasting error. A general agreement on the relevant error metrics and methods for calculating error is needed to increase market acceptance. In addition, standards and guidelines are needed that describe the necessary data and communication systems that allow data from forecast end-users and forecast providers to flow seamlessly back and forth. Standardization of data requirements and communication protocols would increase end-user confidence and thus accelerate market acceptance. The U.S. DOE has begun the work of developing these standards through the two solar forecasting enhancement projects described above. An inclusive, open exchange between forecast end-users and forecast providers, in addition to other relevant stakeholders, is needed to move this process forward. A concerted effort with a steering committee made up of experts from various stakeholder groups could conceivably produce a workable set of standards and guidelines within 18 months. Economic Assessment of Solar Forecasting Value A literature review finds a lack of information on the economic value that solar forecasting can provide to grid operators, utilities, and other stakeholders. There are a multitude of factors that influence the price of electricity (e.g., availability of system generation, transmission outages, etc.), making it nontrivial to isolate just one at any given time. As a result, forecast error on one day may have negligible impact on energy transactions, but the same magnitude error on another day could contribute to a large increase in costs. While the value of solar forecasting qualitatively seems quite large relative to the costs associated with producing solar forecasts, there are currently no quantitative analyses to support this notion. As discussed above, some early analysis suggests that solar forecasting can reduce the needed investment in buffer energy storage. As discussed in this report, predicting solar energy system output for various forecast horizons provides value to numerous stakeholders. A methodology should be developed to assess these values from various stakeholder perspectives. A peer review of the methodologies should be conducted to gain a consensus on the valuation approach. With broad acceptance, the value of solar forecasting should be estimated from the various stakeholder perspectives and disseminated through conference presentations and published articles. Sensitivity analyses should also be conducted to determine the incremental value that increased forecasting accuracy delivers for each of the various applications. For some applications, increased forecast accuracy may deliver significant value and for other applications the benefits of increased accuracy may be rather limited. A more complete understanding of the incremental value of increased forecast accuracy is needed to efficiently allocate research and development resources. In sum, increased understanding of the economic value of solar power predictions could accelerate market acceptance. Distributed Solar Working Group and Workshops While there are various groups working on different aspects of solar deployment and resource forecasting, there is a need for a focused effort on DG solar. A working group comprised of relevant stakeholders should be convened to identify the critical issues and begin a collaborative process to address the unique challenges of DG solar. These challenges include the lack of visibility of DG solar, which can be addressed in part through the development of forecasting techniques and models. Best practices in compiling available data on DG solar systems locations and technical specifications needs to be addressed. In addition, methods to forecast all DG solar forecasts solar electric power association

SECTION 6 from a representative sample of systems (often referred to as up scaling) needs to be developed and tested. While there are performance data on DG solar from proprietary monitoring systems, this data is not generally available for public access yet the data could prove to be valuable in terms of benchmarking DG solar forecasting models. While UVIG and the U.S. DOE & CPUC workshops on variable generation and high penetration of PV provide an opportunity for researchers and practitioners to collaborate and share information, more focused workshops on DG solar could prove to be valuable. Periodic workshops organized by the DG working group with a clear focus on the unique challenges of DG solar could help to accelerate information sharing and emerging solutions to the unique challenges that DG solar presents to grid operators and distribution utilities. Expanded Engagement of Policymakers and Regulators Policymakers and regulators need to gain a greater understanding of the value that solar forecasting can bring to meeting policy objectives, including renewable energy portfolio standards. To date there are no policies or regulations providing guidelines how, when and where solar forecasting should be considered. Policies and regulations can help to create requirements in terms of reporting system characteristics providing the needed data to produce accurate solar forecasts. While larger systems are often subject to reporting requirements by grid operators, system reporting requirements and net metering standards for DG solar systems is a patchwork of standards and specifications across states and utility service territories. This creates a challenge as many RTOs/ISOs cross state and utility boundaries and thus they do not have access to a uniform set of data on DG solar systems within their service territories. Forecasting distributed solar energy production requires technical specifications for perhaps thousands of individual systems. More uniform system reporting standards and a centralized database of this information would streamline the process of predicting future output from behind the meter solar. This is part of the broader need for visibility of solar generators by grid operators to maintain system reliability discussed in this report. Predicting Solar Power Production Another issue that could potentially be resolved by policymakers and regulators relates to the allocation of the costs associated with procuring a centralized solar forecasting service. CAISO, the only RTO/ISO that subscribes to a centralized solar forecasting service today, allocates the cost of the service to all market participants. This may or may not be the appropriate cost allocation method for other ISOs/RTOs. In addition, it is an outstanding question as to whether distribution companies should allocate the cost of a solar forecasting service into the rate base. These questions need to be settled based on an informed discussion of the value that solar forecasting brings to various stakeholders. Finally, policymakers and regulators can engage in collaboration with industry stakeholders to promote market structures or scheduling guidelines that leverage modern solar forecasting services. While the seven regional RTOs/ISOs are moving to real time dispatching of generation sources, which is a benefit to variable sources of generation, they are each at different stages of implementation. In addition, not all RTOs/ISOs allow variable generation to bid in the day-ahead markets. Allowing day-ahead bidding from variable generation sources based on the best dayahead forecasts is an important step to allow solar and wind resources to effectively compete with traditional sources of generation. Those regions outside these RTO/ISO boundaries have significant opportunity to alter generation scheduling timelines to more closely model market operating timelines in regions with wholesale power markets. 37 37 37

Works Cited 38 Ahlstrom, M., D. Bartlett, C. Collier, J. Duchesne, D. Edelson, A. Gesino, M. Keyser, D. Maggio, M. Milligan, C. Mohrlen, J. O Sullivan, J. Sharp, P. Storck, and M. Rodriguez. (2013). Knowledge is power: Efficiently integrating wind energy and wind forecasts. IEEE Power and Energy Magazine, 11(6): 45 52. Bird, L., and D. Lew. (2012). Integrating wind and solar energy in the U.S. bulk power system: Lessons from regional integration studies. NREL/CP-6A50-55830. Botterud, A., J. Wang, C. Monteiro, and V. Mirand. (2009). Wind power forecasting and electricity market operations. In Proceedings of the 32nd Annual Conference of the International Association of Energy Economics. Hoff, T., R. Perez, J. Kleissl, D. Renne, and J. Stein. (2012). Reporting of irradiance modeling relative prediction errors. Progress in Photovoltaics: Research and Applications, 21(7):1,514-1,519. Hoff, T. and R. Perez (2012) Modeling PV fleet output variability. Solar Energy. 86(8): 2177-2189. Hoff, T. and R. Perez. (2011). PV power output variability: Calculation of correlation coefficients using satellite insolation data. A report by Clean Power Research. Hoff, T. and R. Perez (2010) Quantifying PV output power variability. Solar Energy. 84(10): 1782 1793. Kann, S., S. Mehta, M. Shiao, C. Honeyman, N. Litvak, J. Jones, J. Baca, and W. Lent. U.S. solar market insight report Q3 2013 full report, A report by Greentech Media Company and Solar Energy Industry Association. Kleissl J. (2010). Current state of the art in solar forecasting, Final Report California Renewable Energy Forecasting, Resource Data and Mapping Appendix A. California Renewable Energy Collaborative. Kostylev, V. and A. Pavlovski. (2011). Solar power forecasting performance towards industry standards. In 1st International Workshop on the Integration of Solar Power into Power Systems, Aarhus, Denmark. Lew, D., G. Brinkman, E. Ibanez, A. Florita, M. Heaney, B. Hodge, M. Hummon, G. Stark, J. King, S.A. Lefton, G. Jordan, and S. Venkataraman. (2013). The Western Wind and Solar Integration Study Phase 2. NREL/TP-5500-55588. Letendre, S. and R. Perez. (2006). Understanding the benefits of dispersed grid-connected photovoltaics: From avoiding the next major outage to taming wholesale power markets. The Electricity Journal, 19(6): 64 72. Letendre, S. and M. Perotti. (2012). The business case for matching renewable energy production with vehicle charging. In Proceedings of the EVSE26, Los Angeles. solar electric power association 38

Works Cited Lorenz, E., J. Hurka, D. Heinemann, and H. Beyer. (2009). Irradiance forecasting for the power prediction of grid-connected photovoltaic systems. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2(1): 2 10. Mills, A., M. Ahlstrom, M. Brower, A. Ellis, R. George, T. Hoff, B. Kroposki, C. Lenox, N. Miller, J. Stein, and Y. Wan. (2009). Understanding variability and uncertainty of photovoltaics for integration with the electric power system. LBNL-2855E. Lonij, V., A. Brooks, A. Cronin, M. Leuthold, K. Koch. (2013). Intra-hour forecasts of solar power production using measurements from a network of irradiance sensors. Solar Energy, 97: 58-66. Pelland, S., J. Remund, J. Kleissl, T. Oozeki, and K. De Brabandere. (2013). Photovoltaic and Solar Forecasting: State of the Art. IEA PVPS Task 14, Subtask 3.1 Report IEA-PVPS T14-01: 2013. Pelland, S., G. Galanis, and G. Kallos. (2011). Solar and photovoltaic forecasting through post-processing of the Global Environmental Multiscale numerical weather prediction model: Solar and photovoltaic forecasting. Progress in Photovoltaics: Research and Applications, 21(3): 284 296. Papavasiliou, A., and S. Oren. (2008). Coupling wind generators with deferrable loads. In Proceedings of Energy 2030 Conference, Energy 2008 IEEE. Perez, R., E. Lorenz, S. Pelland, M. Beauharnois, G. Van Knowe, K. Hemker, D. Heinemann, J. Remund, S. Mu ller,w. Traunmuller, G. Steinmauer g, D. Pozo, J. Ruiz-Arias,V. Lara-Fanego,.L. Ramirez-Santigosa, M. Gaston-Romero, and L. Pomares. (2013). Comparison of numerical weather prediction solar irradiance forecasts in the U.S., Canada and Europe. Solar Energy, 94: 305 326. Perez, R., S. Kivalov, J. Dise, D. Chalmers, and T. Hoff. (2013). Mitigating short-term PV output intermittency. In Proceedings of EU PVSEC Conference, Paris, France. Perez, R., S. Kivalova, J. Schlemmera, K. Hemker, and T. Hoff. (2011). Parameterization of site-specific short-term irradiance variability. Solar Energy, 85(7): 1343 1353. Perez, R., S. Kivalov, A. Zelenka, J. Schlemmer, and K. Hemker. (2010a). Improving the performance of satellite-to-irradiance models using the satellite s infrared sensors. In Proc. of American Solar Energy Society s Annual Conference, Phoenix, AZ. Perez, R., S. Kivalov, J. Schlemmer, K. Hemker, D. Renné, and T. Hoff. (2010b). Validation of short and medium term operational solar radiation forecasts in the U.S.. Solar Energy, 84(12): 2161 2172. Perez R. and R. Seals. (1997). Comparing satellite remote sensing and ground network measurements for the production of site/time specific irradiance data, Solar Energy. 60 (2): 89 96. Wilcox, Stephen. (2012). National Solar Radiation Database 1991 2010 Update: User s Manual. National Renewable Energy Laboratory: NREL/TP-5500-54824. 39 Predicting Solar Power Production 39

1220 19th Street, NW, Suite 800, Washington, DC 20036-2405 solarelectricpower.org Tel: +1.202.857.0898