Green Machine Organic Rankine Cycle Field Test May December 2013

Transcription

1 Green Machine Organic Rankine Cycle Field Test May December 2013 Principal Investigator: Chuen-Sen Lin Project Manager: Daisy Huang Other Participants Gwen Holdmann Vamshi Avadhanula David Light Thomas Johnson Ross Coen David Pelunis-Messier University of Alaska Fairbanks PO Box Fairbanks, AK August 27, 2014

2 DRAFT REPORT: PROJECT TITLE: An Analysis of Organic Rankine Cycle Green Machine System Performance in Alaska Green Machine Organic Rankine Cycle Laboratory and Field Test COVERING PERIOD: November 2011 December 2013 DATE OF REPORT: August 27, 2014 RECIPIENT: PROJECT LEAD: PROJECT MANAGER: OTHER PARTICIPANTS: Alaska Energy Authority 813 West Northern Lights Boulevard Anchorage, AK Chuen-Sen Lin Department of Mechanical Engineering University of Alaska Fairbanks P.O. Box Fairbanks, AK Daisy Huang, Alaska Center for Energy and Power, UAF Department of Mechanical Engineering University of Alaska Fairbanks P.O. Box Fairbanks, AK Gwen Holdmann, Alaska Center for Energy and Power, UAF Vamshi Avadhanula, UAF David Light, Alaska Center for Energy and Power, UAF Thomas Johnson, Alaska Center for Energy and Power, UAF

3 Acknowledgement We would like to thank the following individuals and organizations for their contribution to this project: David Pelunius-Messier, Tanana Chiefs Conference for your collaboration throughout this project and many others; The Alaska Energy Authority, Denali Commission, Alaska Department of Environmental Conservation, and US Environmental Protection Agency for funding various aspects of this project; Alaska Power and Telephone and in particular Ben Beste, Vern Neitzer, and the rest of the crew at the Tok powerhouse for field testing the unit; Devany Plentovich, Program manager for the Alaska Energy Authority for her guidance and eminent patience with this effort particularly the final report; McKinley Services for assisting with the installation; Mike Ruckhaus, UAF Facility Services and Chilkoot Ward and the staff at the UAF Powerplant for allowing us to test the Green Machine at their facility; Maria Richards from the Southern Methodist University Geothermal Lab; Mark Hall with the Heat is Power Association; and David Sjoding of Washington State University for peer reviewing this report. ii

4 Executive Summary This study involved testing and evaluation of a 50 kw, model Block 1 Green Machine (GM), an Organic Rankine cycle (ORC) system capable of generating power from low-quality heat sources, such as heat rejected from a diesel engine. The testing was conducted in two phases. Under Phase I, the GM was installed at the University of Alaska Fairbanks power plant and operated under controlled conditions to determine power output under different heating and cooling rates. It was also operated at full load for 600 hours for reliability testing and 400 subsequent hours for other testing. Under Phase II, the GM was installed at the Alaska Power and Telephone (AP&T) power plant in Tok, Alaska. A major goal of the Phase II study, the ORC field test, was to determine the potential for critical issues to occur during real-life application of the GM in village diesel power plants and to assess performance and economic potential based on the field test data. The GM as installed in the Tok power plant ran for 1,137 hours. Altogether, laboratory and field tests combined, the GM ran for approximately 3,000 operating hours. During that time, the system was considered reliable, and maintenance requirements were considered minimal. The system performed close to its nameplate capacity of 50 kw under controlled laboratory conditions and about half that under field conditions, averaging 22 kw net power output. This lower output was expected because of the operating conditions in the field deployment (379 kw heat supplied to power unit evaporator). After 1,137 hours of operation in Tok, the system developed a refrigerant leak due to an incompatibility between the seal material and the lubricant used in the refrigerant. The manufacturer, ElectraTherm, reports that this problem is known and has been fixed in the current comparable updated version of the GM, which is a Block 5, 65 kw ORC system. Currently, insufficient funds prevents repair of the GM in Tok, and it remains idle in the AP&T powerhouse. As of this writing, AP&T is in the process of negotiating with ElectraTherm to replace the unit with an updated model, the Green Machine While the unit was operational, AP&T saved an estimated 1350 gallons of diesel fuel, or $6740 during the period of time the unit was operational. Assuming a full year of operation and the use of a more efficient cold water pump, total fuel savings could be 12,370 gallons of diesel fuel, which would save the utility $61,850 per year based on current (2012) fuel prices. A more realistic scenario in which the unit would be operated annually from April 1 st though November 15 th would result in a savings of 7,730 gallons of fuel, representing $38,640 annual savings. Under this scenario, the simple payback assuming 0% interest on the capital expense and no subsidy is 10.7 years. We expect this payback would be the same or better for one of ElectraTherm s newer units. ElectraTherm reports that 27 commercial units are in operation worldwide, including two machines with over 17,000 hours run time and an additional six machines with over 10,000 hours run time. Most of these are the newer series 4200, 4400, or However, this study necessarily limits itself to reporting on the performance of the model used in its laboratory and field tests. iii

5 Table of Contents Acknowledgement...ii Executive Summary... iii List of Figures... vii List of Tables... ix Abstract... 1 Introduction Motivation Background Phase I: Laboratory Testing Introduction Laboratory Test Preparation Selection of a Low-temperature Heat Engine for Testing Required Elements for Test Plan and Test Site System Modeling, Simulation, and Test Parameters Selection Heat Source Heat Sink ORC System Modeling Methodology Selections of Component Parameters and Operation Parameters Simulation Case Study Test Component and Measurement Device Selections Using Simulation Data: Test Plan, Design and Selection of Components, and Procurement Test Plan and Design/Selection of Components Procurement Installation and Instrumentation Commissioning Experimental Setup and Test Schedule Experimental Setup Testing Schedule Parameters Measured and Data Reduction Parameters Measured iv

6 2.8.2 Data Reduction Reductions in Emissions and CO Economic Analysis Reliability Testing and Results Preparation Green Machine Setup Parameters Hot Water Loop Setup Parameters Cold Water Loop Setup Parameters Operation Procedure Checklist Reliability Test Results Green Machine Shutdown during the Reliability Test Performance Testing and Results Performance Test Operation Procedure Results Data Analysis, Performance Curves and Example Based on Developed Performance Curves Analysis Results from Reliability Test Performance from Reliability Test Analysis Results from the Performance Test Discussion Based on Performance Test Results Example Based on the Performance Curves Further Discussion of Adopting the GM for a Rural Genset Discussion (Economics, Emissions, Findings) Findings of General Information Related to the GM ORC Findings in GM Performance and Comments on Applications Match between the GM ORC and a Diesel Generator Set Economic (Payback Period) and Performance Estimates Proposed General Policy in GM Application Conclusions from Laboratory Testing Accomplished Tasks Related to Project Objectives Preparation for Phase II Field Testing v

7 3 Phase II: Field Testing Project Preparation and Test Plan Installation at the Tok Power Plant Heat Source Cooling Source Instrumentation and Measurement/Evaluation Equipment Test Schedule and Test Plan Procurement of Required Equipment and Supplies Transport and Installation of the Green Machine Operation and Maintenance Requirements Data Collection Results from Field Testing Field Data Performance and Economic Analysis Discussion Issues Encountered Lessons Learned Summary References Appendix A. Appendix B. Heating and Cooling System Design Green Machine Startup Parameters, 27 Aug Appendix C. Materials Used in Green Machine Installation Appendix IA Survey of Low Temperature Heat Engine Companies (2008) Appendix IB Survey of Low Temperature Heat Engine Companies (2010) Appendix IIA Preliminary Line Diagram of the Testing System Appendix IIB Preliminary Components Selected for the Testing System Appendix IIIA Methodology Proposed for Stage 2 and Stage 3 Modeling Appendix-IIIB Expressions for Single Phase and Two-Phase Heat Transfer Coefficient of Fluids in Plate Heat Exchangers Appendix IVA Available Floor Space for Test System Installation Appendix VI Estimated additional kw that could be generated using an ORC unit for all Alaskan communities vi

8 List of Figures Fig. 1. Schematic of Organic Rankine Cycle system... 9 Fig. 2. Efficiency of ORC system with varying screw expander inlet quality for heat source temperature of 200 F (93 C) Fig. 3. Efficiency of ORC system with varying screw expander inlet quality for heat source temperature of 250 F ( C) Fig. 4. Design line diagram of the testing system Fig. 5. GM Setup parameters screen Fig. 6. GM Startup Parameters screen Fig. 7. GM Options parameters screen Fig. 8. GM Machine Defaults parameters screen Fig. 9. GM PLC I-O parameters screen Fig. 10. GM Veris Setup parameters screen Fig. 11. Green machine HMI screen-shot during reliability test operation Fig. 12. Hot water and cold water supply temperatures to ORC power unit during reliability test Fig. 13. Net power generated, power consumed by power unit pump and hot water pump during reliability test Fig. 14. Heat input to power unit evaporator vs. hot water flow rates at different hot water supply temperatures and cold water flow rates Fig. 15. Heat rejected to cold water in power unit condenser vs. hot water flow rates at different hot water supply temperatures and cold water flow rates Fig. 16. System operating power output vs. hot water flow rates at different hot water supply temperatures and cold water flow rates Fig. 17. Heat input vs. hot water supply temperature Fig. 18. Heat rejected vs. hot water supply temperature Fig. 19. System operating power output vs. hot water supply temperature Fig. 20. System operating efficiency vs. hot water supply temperature Fig. 21. Payback period vs. hot water supply temperature Fig. 22. CO 2 reductions vs. hot water supply temperature Fig. 23A. Pay period at 0% interest rate on capital for different Green Machine ORC power outputs, fuel prices, and capital costs Fig. 23B. Pay period at 10% interest rate on capital for different Green Machine ORC power outputs, fuel prices, and capital costs Fig. 24. Schematic line diagram of heating and cooling to the Green Machine vii

9 Fig. 25. Well Number Fig. 26. GM as installed in the Tok power plant Fig. 27. GM cold- and hot-water piping Fig. 28. Data acquisition system for Green Machine field test in Tok Fig /40 propylene glycol/water supply and return temperatures to Green Machine Fig. 30. Green Machine operating power output, Green Machine pump power, and coldwater pump power consumption Fig. 31. Comparison of Green Machine UAF lab test results and Tok field results for net power output for same heat source supply temperatures Fig. 32. Payback period at 0% interest rate on capital for different Green Machine ORC power outputs, fuel prices, and capital costs (for full year of operation) Fig. 33. Payback period at 10% interest rate on capital for different Green Machine ORC power outputs, fuel prices, and capital costs (for full year of operation) Fig. 34. Green machine net power output and net efficiency versus hot-water supply temperature viii

10 List of Tables Table 1. Thermodynamic properties and environmental date of R-245fa... 8 Table 2. Operation and performance parameters for data acquisition Table 3. Selected components for heating and cooling loops Table 4. Various hot water and cold water flow rates at which GM will be tested Table 5. Tier 4 interim EPA emissions standards for non-road diesel engines Table 6. Total component cost incurred in building the experimental system Table 7. GM Setup parameters table with range and default values Table 8. GM Startup Parameters table with range and default values Table 9. GM Options parameters table with range and default values Table 10. Checklist for GM, hot water and cold water loops Table 11. Reliability test results at three different times of the test Table 12A. Various hot water and cold water flow rates at which GM was tested Table 12B. Actual Input Conditions for GM Performance Testing Table 13A. Performance results for HW Temp = 155 F; HW flow rate = 120 gpm to 300 gpm; CW Temp 50 F and CW flow rate = 120 gpm, 160 gpm, and 200 gpm Table 13B. Induced performance results from measured readings of Table 13A Table 14A. Performance results for HW Temp = 175 F; HW flow rate = 120 gpm to 300 gpm; CW Temp 50 F and CW flow rate = 120 gpm, 160 gpm, and 200 gpm Table 14B. Induced performance results from measured readings of Table 14A Table 15A. Performance results for HW Temp = 195 F; HW flow rate = 120 gpm to 300 gpm; CW Temp 50 F and CW flow rate = 120 gpm, 160 gpm, and 200 gpm Table 15B. Induced performance results from measured readings of Table 15A Table 16A. Performance results for HW Temp = 215 F; HW flow rate = 120 gpm to 300 gpm; CW Temp 50 F and CW flow rate = 120 gpm, 160 gpm, and 200 gpm Table 16B. Induced performance results from measured readings of Table 16A Table 17A. Performance results for HW Temp = 225 F; HW flow rate = 120 gpm to 300 gpm; CW Temp 50 F and CW flow rate = 120 gpm, 160 gpm, and 200 gpm Table 17B. Induced performance results from measured readings of Table 17A Table 18A. Performance results for HW Temp = 155 F to 220 F; HW flow rate = 120 gpm to 300 gpm; CW Temp 68 F and varying cold water flow rate Table 18B Induced performance results from measured readings of Table 18A Table 19. Reliability test results Table 21. Diesel engine specifications ix

11 Table 22. Estimated ORC performance for operating on waste heat recovery from diesel engine Table 23. Instrumentation equipment and components for data collection from the GM 50 kw field test at Tok, Alaska Table 24. Parameters measured and instrumentation used during Phase II testing Table 25. Reduced form of the recorded Tok field test data and generated Green Machine ORC performance data during the field test period Table 26. ORC performance during continuous operation period from 10/02/2013, 11:00 A.M., to 11/19/2013, 7:00 A.M. ( hours) Table 27. Estimated ORC performance in Tok, Alaska, for a full year of operation (8,760 hours) x

12 Abstract This report describes the testing and evaluation of an ElectraTherm 50 kw, model Block 1, Organic Rankine cycle Green Machine as completed by the Alaska Center for Energy and Power at the University of Alaska Fairbanks. The Green Machine was tested in two phases. Under Phase I (Laboratory Testing), the Green Machine was installed at the University of Alaska Fairbanks (UAF) power plant and run under controlled conditions to determine power output at different heating and cooling rates. It was also run under full load for reliability testing for a total test time of over 1,000 hours. Following the UAF tests, the unit was deemed suitable for deployment in a village power plant, and under Phase II (Field Testing) the community of Tok was selected to host the demonstration. The unit was installed at the Alaska Power and Telephone power plant in late summer of After its commissioning on October 2, 2013, the Green Machine operated for 1,138 hours before a seal on the screw expander failed, which caused refrigerant to leak. Prior to the failure of the seal, the Green Machine performed as expected. This report describes the study s objectives, the testing plan, the design and fabrication of the testing system, data collection, the Green Machine abbreviated field test performance and an economic analysis based on that performance, the lessons learned, and recommended guidelines for Organic Rankine cycle system selection and application for rural Alaska village diesel generators.

13 1 Introduction This study involved testing and evaluation of a 50 kw, model Block 1 Green Machine (GM), an organic Rankine cycle (ORC) system, and took place in two phases. Phase I involved laboratory testing under controlled conditions at the University of Alaska Fairbanks (UAF) power plant. The GM was run for 50 hours under different prescribed conditions of heating and cooling and for 1,000 hours at full load. Phase II involved field testing under real-world conditions at the Alaska Power and Telephone (AP&T) power plant in Tok, Alaska. The major goal of the Phase I study was to operate the GM in a controlled environment to determine its reliability and gain first-hand information about its performance. It was intended that the results would be used by the funding agency, the Alaska Energy Authority (AEA), as well as utilities and vendors to ascertain whether this technology has the potential to improve overall plant efficiency by using rejected heat from diesel generators to generate additional electricity by employing ORC technology. Phase I results suggested that in circumstances where excess heat is available and a sufficient difference in temperature can be achieved between the excess heat utilized from the diesel engines and a cooling side for operating the evaporator, the GM may be an economical option worthy of consideration. Based on these results, funding for a field test Phase II was sought and awarded. The primary goal of the Phase II study, the ORC field test, was to evaluate the performance of the GM during operation in a village diesel power plant. Other goals included gathering and analyzing performance data, evaluating operation and maintenance requirements, analyzing the economics of potential power generation given the operating parameters of the local power plant, and determining factors that should be considered when deciding whether to deploy ORC technology in an Alaska diesel power plant. 1.1 Motivation Isolated rural villages in Alaska annually consume about 30 million gallons of diesel fuel to generate 370,000 MWh of electrical energy, produced by individual diesel-fired generator sets. In general, the ratio of electrical power produced to fuel energy consumed is less than 40%. The rest of the fuel energy is lost as heat. While some power plants in Alaska s rural villages use a portion of this energy for other beneficial uses, such as space and water heating, most of this energy is wasted. The goal of adding ORC products to an existing power cycle is to reclaim some of this heat to generate more power, thus increasing the overall fuel efficiency of the power plant. While ORC technology is mature for larger-scale power generation, the products for smaller-capacity generator sets, appropriate for the typical size of Alaska village power plants, are still new to the market or are in the prototype phase. Many villages have been approached by ORC product developers to invest in this new technology, so the Alaska Center for Energy and Power (ACEP), together with its funding partners, the Denali Commission and AEA, established a program to test the viability of ORC products in Alaska. 1

14 ElectraTherm s Green Machine (GM) was identified as one of the devices with the highest potential for success. The GM discussed here, an early Block 1 model, is designed to generate up to 50 kw of power using the ORC, which is a process that can obtain energy from lowervalue (lower-temperature) heat sources than are commonly used for power generation. (As of this writing, the current market-ready version of the GM 65 kw machine is the 4400; the Block 1 50 kw machine described in this report is no longer being sold by ElectraTherm.) 1.2 Background In 2010, ACEP partnered with Tanana Chiefs Conference (TCC), a nonprofit consortium of 42 communities in Interior Alaska, to obtain and test a 50 kw GM. In 2011, the unit was purchased by TCC with grant funding from the Denali Commission. In November, ACEP installed the unit in the coal-fired power plant at the University of Alaska Fairbanks (UAF), using heat from the plant s steam loop tempered with potable water. The GM was operated for 50 hours under different controlled heating and cooling conditions, and for 1,000 hours of reliability testing under full load. TCC facilitated communication between ACEP and the villages under its umbrella to select an Alaska community whose power company would be willing to field test the GM. The community of Tok was selected by TCC, due to its road accessibility and relative proximity to Fairbanks, availability of waste heat at least seasonally, and the competence, expertise, and interest of the staff managing the local powerplant, operated by Alaska Power and Telephone (AP&T). 2

15 2 Phase I: Laboratory Testing 2.1 Introduction In rural Alaska, approximately 180 villages consume about 370,000 MWh [1] of electrical power annually using isolated diesel generator sets. In part because diesel fuel is imported from long distances often just once a year and then stored in bulk, the cost of fuel, and hence the cost of generated power, is very high. If waste heat from the diesel generators can be captured and used for either space heating or supplemental power generation, the fuel savings are significant. Many applications for low quality heat, such as heat rejected from a diesel engine, have been demonstrated. Examples include general heating (e.g., space heating and city water temperature maintenance), direct thermal-to-electricity conversion, heat-to-power conversion using a heat engine, refrigeration, and desalination. Of these applications, waste heat for heating is considered the most efficient application and is commonly practiced seasonally in Alaska. However, in some cases, waste heat for heating is not practical for reasons that include lack of proximity between the powerhouse and buildings, or prohibitively high construction costs and incompatibility of building heat systems. Waste heat for heating in Alaska village diesel generators has been discussed in detail [2]. Waste heat for power through heat engines is recommended under appropriate circumstances because of its acceptable efficiency (i.e., close to 10%), flexibility in heat utilization, and expected low maintenance (similar to steam engine or refrigeration systems). In addition, unlike heating, power is needed year-round. Power usage in many of Alaska s rural villages is about or below 1 MW. For these generator sets, the power produced using waste heat is expected to be below 100 kw. For waste heat engines belonging to this category, the power-to-cost ratio is expected to be very high if the heat engine is facilitated with a radial turbine (a type of expander) for heat-to-power conversion. Many different thermodynamic cycles and different types of heat-to-power expanders have been used to lower the cost. Examples of thermal cycles include the ORC and the ammonia/water (or Kalina) cycle. Examples of heat-to-power conversion expanders include the screw expander, scrolling expander, and piston expander. All of these examples are still in the prototype and proving stage or the prototype fabrication stage. It is well understood that the performance of a heat engine depends on conditions of the heating source and cooling source, both of which largely rely on the load pattern and waste heat properties (e.g., exhaust, jacket coolant) of the diesel generator set and the cooling source available in the village. Therefore, to estimate the performance and economic impact of any waste-heat engine on an individual diesel generator set, the performance data of the heat engine under various heating and cooling conditions are needed. In general, these data are obtainable by testing the heat engine under controlled heating and cooling conditions. The Phase 1 study had four objectives. The first was to demonstrate that an improvement of the efficiency of the diesel power plant by about 10% (i.e., about 4% of fuel efficiency) is achievable through the use of an Organic Rankine cycle (ORC) system, which uses waste heat contained in diesel engine jacket water and exhaust. The second objective was to evaluate the 3

16 feasibility, operation and maintenance requirements, and payback time of applying a selected ORC system. The third objective was to develop guidelines for ORC system selection, operation, and maintenance, and to evaluate the potential impact of applying waste-heat ORC systems on rural Alaska economy, fuel consumption, and emissions and greenhouse gas reductions. The fourth objective was to compare the performance and economics of two ORC systems: (1) a 50 kw system that uses a screw expander and is an emerging technology, and (2) a Pratt & Whitney (P&W) 250 kw unit that uses a radial turbine and is a comparably well-developed technology. 2.2 Laboratory Test Preparation Preparation for Phase I included the selection of an appropriate low-temperature heat engine, the layout of required elements for the testing plan (critical parameters, etc.), and the selection of a testing site (e.g., utility, heat source, and heat sink) Selection of a Low-temperature Heat Engine for Testing The proposed project began by surveying (in 2008 by Jared Kruzek) accessible manufacturers who are involved in the low-temperature heat engine industry about their industrial applications and potential and willingness to deliver, within a reasonable time, a lowtemperature heat engine with a power capacity between 10 kw and 100 kw. Eighteen manufacturers were contacted, and their general product information was reviewed. The manufacturers selected for further consideration included an ammonia/water system manufacturer and an ORC system manufacturer. The ammonia/water system manufacturer was contacted because of its previous credible working experience with the Alaska power industry, because its ultra-small size of 10 kw could be applied in Alaska s smallest and most high-cost communities, and because it showed its complete design layout and willingness to spend its own funds for fabricating a prototype system. In 2010, pre-shipment testing conducted by the ammonia/water system manufacturer showed that a major component (the heat-to-power conversion unit) of the system had two major drawbacks, which hampered overall system performance and caused repeated delays in delivery. While the manufacturer continued work on improving the product, no timeline for delivery could be established. The drawbacks included (1) migration of lubrication fluid from the power unit into the loop of working fluid, which lowered the heat transfer efficiency significantly after a short period of operation, and (2) much lower than expected heat-topower-conversion efficiency of the power unit. Also in 2010, the ORC manufacturer showed promising test results for its ORC prototype. The After conducting an updated market survey to assure no other promising technologies were overlooked, ACEP decided to move forward with procuring and testing the ORC, which was a 50 kw rated unit manufactured by ElectraTherm under the model name Green Machine, Block 1 model. 4

17 2.2.2 Required Elements for Test Plan and Test Site The test plan included testing the ORC system for reliability and performance. The heat engine required a heat source, which provides driving energy to the heat engine and emulates the heating conditions (i.e., temperatures and flow rates) received from the waste heat generated by the village diesel generators. A heat engine also needs a cooling source, which absorbs the dissipated heat from the heat engine and emulates the variety of cooling conditions possibly provided by the cooling sources existing in the villages (e.g., surface and ground water, radiator, and cooling tower). In addition, heat energy transmitting devices (e.g., heat exchanger, pipe, pump, and valve) are needed to transmit heat between the working fluid in the ORC system and the media of the heating and cooling sources. Other examples of required system elements include devices for safety and reliable performance. The purpose of the reliability test was to observe the endurance in operation and consistency in performance of the ORC system under the rated operation condition. The purpose of the performance test was to look at performance details of the ORC system and its components (i.e., efficiencies, energy consumptions) under numerous operation conditions of heating and cooling. There were more parameters to be measured and measurement devices to be installed for the performance test than for the reliability test. To guarantee that all measurement components needed for this project were included in the final testing system design, the preliminary design line was diagrammed, showing all the parameters to be measured. The line diagram of the preliminary design and required components are given in Appendix IIA along with information on components (Appendix IIB). Based on the preliminary design and components information, the requirements for space, facilities, and utilities of the test site were then estimated. The selection of the final design concept needed to be conducted based on the existing facility and resources of the testing site, available overall test budget, desired operation and maintenance requirements, and time constraints. 2.3 System Modeling, Simulation, and Test Parameters Selection In order to find the performance of the ORC system and the performance of its individual components, sensors and measured data of physical properties (i.e., temperatures, pressures, and flow rates) of the working fluid pertinent to the components are needed. The collected data are then analyzed to give the performance results. Preliminary designs and selection of appropriate testing components (e.g., sizes and types) are obtained through the process of system modeling and simulation. Modeling and simulation may also help determine operation parameters that are critical to system performance. By applying testing data, the model can be further improved and simulation results may become good enough to be useful in predicting long-term performance and applying the ORC system to any individual diesel generator. The sections that follow describe the model construction process for different stages of modeling and corresponding functions, and a constructed first-stage model for performance simulation and testing components selection. The model includes three components: heat source, heat sink, and the ORC system. The fluid used in heating and cooling loops was water; the working fluid used in the ORC system was R-245fa refrigerant. 5

18 2.3.1 Heat Source The physical heat source loop for the new test site was expected to include a hot water source from a steam/water heat exchanger, a VFD pump, and pipes and fittings. Other components for measurement and control are included. Heating fluid enters the heat source heat exchanger of the ORC system through pipes, and transfers heat to the working fluid, the R-245fa refrigerant. The loop controls the temperature and flow rate of the existing hot water from the steam/water heat exchanger. All important information from the steam and heating water loop (e.g., fluid temperatures, pressures, flow rates, VFD rpm, pump power consumption) for each operating condition was collected for system and components performance analysis. The model constructed corresponding to the heat source included all important operation parameters (i.e., temperature and flow rate control and pump power evaluation) of all the function features. The model could be modified easily to cope with different types of heat source and components Heat Sink At the selected test site, the physical heat sink loop included a cooling fluid source from a fire hydrant and its manual flow rate control valve, pipes, and fittings, temperature and flow rate measurements, and control devices. The cooling fluid entered the condenser of the ORC system through the pipes. The loop had limited controllability in temperature and flow rate of the cooling fluid entering the condenser. Information on fluid properties along the pipes, as well as power and water consumption corresponding to conditions of each operation, were collected. The model constructed corresponding to the cooling source included all important operation parameters (i.e., temperature and flow rate control and pump power evaluation) of all the function features. In addition, the model could be modified easily to cope with different types of cooling sources and components ORC System A general ORC unit includes at least a pump, an evaporator, a heat-to-power converter, and a condenser. Other components needed in modeling depend on the versatility of the physical construction of the ORC system. For example, one known property of the ORC system with a screw expander is its ability to allow mixed vapor/liquid working fluid in the heat-to-power conversion unit (in this case the screw expander), so that the system can add a component to control the working fluid flow rate and/or quality of fluid entering the expander to optimize system performance. The physical ORC system used for this project was an integrated unit, which may not be practical for conducting accurate measurements of working fluid properties for performance analysis of individual components without modifying the system (modifying the system may result in losing the warranty). Also, detailed engineering information on individual components may not be available due to concerns about intellectual property. The system does allow the addition of more sensors to access approximate working fluid properties related to performance of many of the components. This feature is helpful in getting better analysis results of the components. 6

19 2.3.4 Modeling Methodology Detailed engineering data for individual components may not be needed for testing ORC system reliability and physical performance related to system power generation under different heat source and sink conditions. Detailed engineering data for components needed for ORC optimal net power generation also may not be needed. In other words, to achieve the objectives listed in the Introduction, detailed engineering data of the ORC system components may not be critical. However, if the purpose of a test is to check that the operation condition (input to the ORC) is optimal and to offer comments on system design, detailed engineering data of components are required. To model the system with reasonable accuracy, the modeling plan was divided into two to three stages, depending on how feasible and desirable it was to know the details of the performance parameters of the ORC system and its components. The first stage involved modeling the ORC system using simple values to represent system performance parameters based on specifications of the ORC system and its components. The purpose of this stage was to qualitatively understand the effects of the operating conditions of the heat source and heat sink on ORC system performance and to find approximate ranges of operation of the parameters of system components. The results would be used for test planning and selections of test system components and measurement devices. The second stage of modeling was to fit the system and component parameters using limited measurement data obtained from limited experimental cases (i.e., heating and cooling conditions). If system simulation results obtained using fitted (approximate) values of system parameters can qualitatively match experimental results (although without appropriate accuracy), extra experimental cases could yield a more complex and detailed model as the third stage. The purpose of the second and third stages of modeling was to develop a simulation model that can predict system performance at any system operation condition. Since modeling was at the component level, the performance prediction was acceptable to the operation conditions, which were moderately outside of the operating conditions covered in the test. The methodology of the first stage is described in detail here. The methodologies of the second and third stages are given in Appendix IIIA for reference purpose only. In the first stage, the selected ORC system was able to optimize the system net output for all heating and cooling conditions, limiting the total output to a maximum of 50 kw. There are other features related to conventional constraints for performance regulation and system protection, such as constraints in maximum temperature and maximum pressure. In a physical prototype, these constraints are imposed by physical mechanisms, such as mechanical and electrical devices. To simplify the mathematical model, some of these functions can be emulated using a conditioning computation statement or neglected because they have much less influence than other functions on the gross performance of the ORC system. The simplified model (Figure 1) includes the evaporator, a screw expander, a condenser, a VFD pump, and controlled heating and cooling sources. The model also allows the quality of the working fluid entering the expander to be adjusted by varying the flow rate of the working fluid for optimal ORC system performance. Simulation results are useful for test planning and test system component or sensor selections. The expander was modeled by a single efficiency at this stage 7

20 and will be modified as more information becomes available in publications and through experimental data. The pump was modeled with varying efficiency based on its operation condition, and the evaporator and condenser were modeled by their respective flow and heat transfer properties and heat transfer areas. The evaporator has the capability to model liquid, liquid/vapor mixture, and vapor flows. Currently, the condenser is modeled as a single section unit. If heat loss to the atmosphere is found significant, heat loss will also be included in the model Selections of Component Parameters and Operation Parameters Ranges of values of operation parameters used for simulation were based on specifications of the components (e.g., hot water flow rate and temperature limits, pressure and temperature limits of the ORC system), properties of the fluids (i.e., heating, cooling, and ORC working fluid), performance of sub-components (e.g., heat exchanger performance versus flow rate), and properties of heat and cooling sources. Known limits, which mostly are based on the recommendation of the manufacturer, included maximum pressure of the ORC system (150 psi), estimated heat source temperature (235 F), controlled heat source capacity (2.4 MMBtu), flow rates of pumps (250 gpm for heating, 375 gpm for cooling), and cooling sink capacity (3.0 MMBtu recommended by an ORC engineer). The values for ORC component parameters adopted from publications (limited data available) and conventional application practice [3] for system simulation included expander efficiency (e.g., 0.78), pump efficiency (e.g., 0.70), heat transfer coefficient of evaporator (e.g., 1500 W/m 2 - K or 265 Btu/ft 2 - F), evaporator area (e.g., 100 ft 2 ), heat transfer coefficient of condenser (e.g., 1400 W/m 2 - K or 247 Btu/ft 2 - F), and condenser area (e.g., 200 ft 2 ). Other limitations included maximum heat source temperature (225 F) and minimum cooling source temperature (50 F). Some of the values were adjusted based on the match between the simulation results and the published and experimental data. The heat exchanger simulation model was based on standard practice [4, 5]. Since the working fluid (R-245fa) properties would affect the ORC system performance and the temperature and pressure limits used for testing, some of the physical properties were obtained (see Table 1). Detailed properties of R-245fa can be found in NIST documents. The results of the first stage simulation included system performance (e.g., net efficiencies, expander power) of the ORC system, and net efficiencies of the test system as functions of operation parameters of the heat source and heat sink. The results include the effects of sizes of heat exchangers and efficiencies of expander and pumps on system performance. The first stage results were used to help design the test plan for the ORC system. Table 1. Thermodynamic properties and environmental date of R-245fa Safety Vaporization Heat (1atm.) Non Kj/Kg Flammable (355.5 Btu/lb) * Ozone depletion potential ** Greenhouse warming potential Boiling T.(1atm.) 14.6 C (58.3 F) Critical Point 154 C (309.2 F) 36.4 bar (527.9 psi) Saturation Slope ODP* GWP** 100 year Isentropic Besides simulating results in system performance, the process also helped with estimating sizes and capacities of components and measurement devices needed for the testing system. 8

21 The first stage of test system modeling was completed and the effects of flow properties of heat source and heat sink on net efficiencies of the ORC system and the test system (i.e., including parasitical power consumptions through heating loop and cooling loop) were obtained from simulation. In addition, temperature drops of heating flow and cooling flow crossing the ORC system were useful for selections of Btu meters for heating and cooling loops and the steam/hot water heat exchanger. Simulated thermodynamic states of working fluid along the ORC system were useful for selection of pressure gauges for components performance monitoring. A schematic of the ORC system used in this simulation model is shown in Fig. 1. The working fluid used in the simulation model was refrigerant R-245fa. The saturated liquid refrigerant from the condenser was pumped at high pressure to the pre-heater. In the pre-heater the refrigerant was heated to the saturated liquid state; the saturated liquid then went to the evaporator. In the evaporator the saturated liquid was superheated or saturated (may include vapor or vapor/liquid mixture). This high-pressure working fluid was converted to low-pressure liquid or vapor/liquid mixture (to the condenser pressure) using a screw expander, which is connected to the generator to produce power. The low-pressure refrigerant from the screw expander was cooled to the desired state in the condenser, and the liquid portion was pumped back to the pre-heater, and the cycle continued. Fig. 1. Schematic of Organic Rankine Cycle system The ORC system model has three major components: the heat source loop, the heat sink loop (open loop), and the ORC system. In the heat source loop for diesel generator waste heat application, the heating fluid may come from the engine jacket water or from a 50/50 glycol/water mixture exiting the exhaust heat exchanger or both combined. In the heat sink loop, the cooling fluid may be from the cooling tower, radiator, or a large body of water, such as from a nearby river or lake. 9

22 Modeling and simulation was a continuous process. As more system component information was available from experiments and literature, the model was updated to predict more realistic and accurate performance results for given heat and cooling conditions, which were expected to function for the model at the second and third stages. Detailed system component information, such as for the screw expander and for boiling and condensing heat transfer coefficients of refrigerant in the evaporator and condenser, was not available in the literature and could only be evaluated based on experimental data (if experimental data of the ORC system was accessible). The data obtained from the experimental analysis was used to tune the model so that in the future it can be applied to any waste heat source for economic and feasibility analysis of the ORC system. The intention of the second stage and third stage (Appendix IIIA) modeling was to enable the model to be capable of comparing the performance of the ORC system operated under different diesel engine load and environmental conditions, not to reengineer the design of the ORC system. The difference between the second stage and third stage modeling is in the complication of the modeled component (i.e. the model parameters representing the component). In the first stage simulation, five system parameters were being controlled; they included inlet temperature and flow rates for heat source and heat sink input loops and the state of the refrigerant inlet to the expander. The quality of refrigerant inlet to the expander was controlled to investigate its effect on the power output and efficiency of the ORC system for given heat source and heat sink conditions. The system simulation had been performed for different screw expander inlet refrigerant states for given heat source and heat sink inlet conditions. The heat source and heat sink inlet conditions were flow rate and inlet temperatures of respective fluids (here water was considered for both heat source and sink). In the current preliminary simulation, the following assumptions were made: 1. All the ORC heat exchangers, i.e., evaporator, pre-heater and condenser, are 100% efficient. 2. The quality of refrigerant out of the evaporator in the ORC system is controlled. 3. The quality of liquid out of the pre-heater and condenser are saturated liquid. 4. The isentropic efficiency of the screw expander and pump (within the ORC system) are constant at 78% and 70%, respectively. Assumptions affect system performance characteristics. If simulation results do not match the published or measured performance characteristics of the real system, a test plan must be designed to determine which assumptions need to be changed. The model can be modified easily, once better performance characteristics of components are obtained from experiments. In simulating the performance of the ORC system, explicit formulae for heat transfer coefficients of refrigerant on one side of the heat exchanger and water on other side of the heat exchanger should be known. Generally, the heat transfer coefficient of a fluid is expressed in terms of its thermodynamic and transport properties. The heat transfer coefficient also depends heavily on the geometry of the heat exchanger and material of construction. 10

23 All heat exchangers considered in the present case were plate heat exchangers (PHE). A widely accepted expression for heat transfer coefficient of single-phase fluids in a plate heat exchanger is given by Muley and Manglik [6]. In the ORC system, this expression is used for calculating the heat transfer coefficient of hot water and cold water in the evaporator and the condenser, respectively, and the heat transfer coefficient of hot water and refrigerant in the pre-heater. In all of the above cases of heat transfer coefficient calculation, the fluid thermophysical properties were taken at average fluid temperature. An expression for heat transfer coefficient of evaporating refrigerant liquid-vapor mixture in the evaporator is given by Ayub [7]. The expression for heat transfer coefficient of condensing refrigerant liquid-vapor mixture in the condenser is given by Selvam et al [8]. All the above expressions are presented in the Appendix IIIB Simulation Case Study The constructed ORC system model has been used to simulate an example of an ORC system of 50 kw with the system parameters mentioned above, and heat exchanger parameters and computation method listed in Table IIIB-1 of Appendix IIIB. The efficiency versus expander inlet quality is shown in Fig. 2 and Fig. 3. These figures also show the effect of parasitic power and heat sink supply temperature on system efficiency. The parasitic power is the power needed to pump the heat source and heat sink fluids to and from the ORC system. As the heat sink supply temperature decreases (in this case from 21 C to 5 C), the same amount of heat from the condensing refrigerant in the condenser less the amount of cooling fluid must be removed, which may decrease the parasitic power and increase the efficiency of the system. This result may be one of the advantages of using the ORC system during the winter months. 11

24 8 Heat Source Temperature of 200F Efficiency (%) System Pump Work Only All Pump Work (Tc=21C) All Pump Work (Tc=5C) Expander Inlet Quality (kg/kg) Fig. 2. Efficiency of ORC system with varying screw expander inlet quality for heat source temperature of 200 F (93 C) 12 Heat Source Temperature of 250F 11 Efficiency (%) 10 9 System Pump Work Only All Pump Work (Tc=21C) All Pump Work (Tc=5C) Expander Inlet Quality (kg/kg) Fig. 3. Efficiency of ORC system with varying screw expander inlet quality for heat source temperature of 250 F ( C) 12