1 1 Melding Big Data and CIM for Bold Power Systems Insights Dr. Siri Varadan, PE UISOL, An Alstom Company
2 Outline Big Data CIM and Big Data Use Cases Conclusions
3 Big Data Definition Gartner defines Big data using the four V s. Velocity Volume Variety Variability Big Data is a Scalar quantity That is, it may be described completely by just a number
4 Big Data in Power Systems AMI data PMU data Data from IEDs The Data could include: Voltages Currents Power flows Status Time series data Temperatures Pressures Other (DGA etc.) Power system data is Scalar and more it has location, topology, directionand context Data correlations for certain data sets in power systems are governed by physical laws
5 Big Data in Power Systems The Question Scalar data in itself provides insights Nonscalar data in itself provides insights Can the two be combined to gain additional, unique insights using Data Analytics? Analyticsis defined as the discovery and communication of meaningful patterns in data
6 Big Data in Power Systems The Answer Is not 42 It depends On what you are looking for, or the Lens Lensis a specific business function that helps summarize data from a specific perspective On the availability of an appropriate Lens and its granularity
7 Big Data in Power Systems An Example Lens 345 _0 o
8 Big Data in Power Systems Kinds of Lenses Depends on Area of focus (G, T, D etc.) Driven by the business purpose Applicable to Real-time or Non-Real time data Detection of an incipient system separation Calculation of ATC/TTC Asset Health in real-time Failure rates of asset classes Theft detection
9 Big Data and CIM Premise Big Data when combined with non-scalar attributes like topology, location within the context of a power system model (as in CIM) provides unique insights. Questions: What insights? How? Take AMI data and combine it with CIM data for a feeder!
10 Use Cases Distribution Focus Opportunities to Use AMI Data Description Additional Data Utilized * Distribution Loss Analysis Identify trend of loading on feeders, analyzing potential breakdown of theft and line losses (Where, How much) and notify user. Distribution Transformer Monitoring and Health Indexing Complete Feeder Reliability Analysis Compute, trend and notify user of excessive feeder and transformer loadings over time (Which one, How much, How long) Track, trend and predict feeder reliability using asset health indicators for key feeders Distribution SCADA or Pi Historian, GIS/CIM feeder connectivity, OMS or DMS operational data, CIS data, CMMS, Vendor Catalogs
11 Use Cases Distribution Loss Analysis Aggregate AMI data for customers on a per transformer, then feeder basis (CIM model provides topological connectivity, CIS provides customer information) to get a near-real-time load profile at each distribution transformer Run a load flow using the topology model (from CIM) and aggregated loads to get line loss for the same time period The power flowing out at the feeder head (obtained from SCADA) over the same period should match with the sum of the aggregated loads and loss. If not, loss is detected! Energy loss, not Power loss, is of interest!
12 Use Cases Distribution Loss Analysis Heat-Maps showing feeders with suspiciously high losses Depending on granularity of meter data, the process can become nearreal-time
13 Use Cases Distribution Transformer Monitoring and Health Indexing Define metrics for distribution transformer monitoring based on aggregated AMI data Hours in service # of overloads Time spent overloaded Create Health indexes for feeders based on individual metrics that comprise the feeder Create Health indexes for substations based on performance of connected feeders AMI Data could be the basis for Transformer Load Monitoring
14 Use Cases Distribution Transformer Monitoring and Health Indexing Displays based on metrics defined earlier AMI data used to calculate metrics AMI data merged with CIM model data
15 Use Cases Distribution Transformer Monitoring and Health Indexing Data Visualization (using SpotFire )
16 Use Cases Complete Feeder Reliability Analysis Define metrics for Feeder performance based on AMI data Could essentially establish measures like CAIDI and CAIFI on a per customerbasis These measures will be a true reflection of what the customer experienced! AMI Data could be the basis for a new set of customer centric standards!
17 Use Cases Complete Feeder Reliability Analysis Heat-maps to show feeders that have poor performance Performance calculated on the basis of AMI data
18 Use Cases Other Bold Insights Situational Awareness Real-time operations and control Outage Management Crew dispatch Asset Intelligence Real-time health indexing Condition based asset de-rating Lifecycle management Customer satisfaction Workforce management Planning Power systems offer infinite possibilities for Data Analytics!
19 Conclusions Big Data Analytics must have a business purpose. That is, the Lens must be business driven Depending on the business purpose, the appropriate Lens of adequate granularity may be developed The amount of business value is clearly dependent on the Lens You get what you pay for! Big Data analytics gets more meaningful in the context of the CIM There is a lot of visual appeal to data when combined with GIS based maps for power systems A picture is worth a thousand words!
20 Conclusions Big Data, CIM and Other Systems The Vision AMI SCADA Data Aggregation (In Memory) GIS TLM MDM OMS DMS PI Operational Data Sources XML, UDF CIM XML Asset Connectivity Data (contextual) Advanced Visualization (Spotfire) Event Detection & Notification
21 Conclusions Big Data, CIM and Beyond CIM is growing rapidly to include T&D CIM is growing to encompass asset data pertaining to laboratory tests (for DGA, Oil Analysis) for asset health indexing CIM is growing to include other aspects of power systems Work Management Asset Management Maintenance Management Customer support Operations and Network Control CIM will be the skeleton off which all Data will ultimately hang!
22 Thank you! 22