Measurement validation peak load reduction
Download as PPTX, PDF1 like1,355 views
The document presents an overview of Duke Energy's Distribution System Demand Reduction (DSDR) initiative, which utilizes advanced measurement and validation (M&V) techniques to reduce peak electricity demand across its wide service area. Key elements include a centralized distribution management system, real-time data analysis, and the integration of various technological tools for optimized network operations. The findings highlight significant energy savings and operational efficiencies achieved through this initiative, underscoring the importance of accurate data collection and continuous refinement of methods.
Measurement validation peak load reduction
- 1. Measurement & Validation of Peak Load Reduction Jeremy Carden, P.E. Lead Volt/VAR Engineer Duke Energy Dragan Popovic, Ph.D. Executive Vice President – Smart Grid IT Schneider Electric 1 February 4, 2015
- 2. Presentation Summary Introduction to Duke’s DSDR system Implementation Details Duke’s M&V Approach Benefits to Duke Productizing the Solution Conclusions
- 3. Duke Energy Progress – Distribution System Demand Reduction (DSDR) Peak demand reduction through VVO Deployed on entire distribution grid Controllable load: 8,400 MW at peak 315 substations 1,160 feeders 1.5 million customers 34,000 square miles of service area 3
- 4. DSDR Implementation Details Centralized Distribution Management System (DMS) Integrated with several business applications (GIS, OMS, CIS, EMS, etc.) Model-based with near real-time measurement input from the grid 7 Million GIS assets 400,000 SCADA points 90,000 SCADA measurements used as part of state estimation process IP-based, two-way communications 800 substation voltage regulators 415 substation capacitor banks Over 10,000 feeder devices 2,900 voltage regulators 2,900 capacitor banks 1,500 medium voltage sensors 3,000 low voltage sensors 850 reclosers 4
- 5. M&V Approach for DSDR Post-activation analysis developed by Duke in 2013 Creates a baseline estimate by applying a polynomial regression to pre-activation load measurements Primarily evaluated at the system level Performed on data captured directly from DMS (before/during/after) Data requirements Frequent sample rate (typically 30 seconds) 3 – 6 hours of pre-activation data (winter vs. summer peak) 5 – 13 hours of total data (winter vs. summer peak) 5
- 6. M&V Approach for DSDR Strengths Relatively simple and practical (in comparison to others) Created from actual load measurements Statistically based Results fall within in recommended confidence intervals Favorable results when compared to other methods Weaknesses Relatively new Not an actual load curve Manual and can be time extensive Unreasonable baselines for particular days 6
- 7. 2014 Polar Vortex Example 7
- 9. Results/Benefits Validation of 316 MW of peak demand reduction Avoided peak energy production cost estimates DSDR business case deferred construction of two peaking CT generation units Validation of 178 MW of peak spinning reserves (non-optimized voltage reduction) Avoided energy production cost estimates Validation of benefit from emergency voltage reduction activations Example is 2014 Polar Vortex that helped DEP avoid shedding firm load Development of hourly forecast models Aids in planning and economical dispatch order of resources 9
- 10. Conclusions Continual application and refinement is a necessity Compare and contrast against other methods Duke has utilized 3 different methods Provides validation of results Data, Data, Data Need the tools and resources to collect, extract, and retain Not only for benefit but also for system performance (identifying issues) Visibility into the state of the grid during activation Availability of resources (circuits and devices) Communications status Event/command logs 10
- 11. Productization – ADMS ●Software solution that integrates SCADA/DMS/OMS/DSM into one product – IT/OT convergence ● Managing one network model for all operations based on the GIS ● One user interface familiar to all users within utility ● One data base and one security ●Functionality ● Network monitoring and management ● Improved management of alarming, tagging and history ● Validation of switching operations ● Basic calculations (load flow, state estimation) and network optimization applications ● Closed loop execution (VVO, FLISR) ● Distribution training simulator (DTS) ● Distribution network planning and design tool
- 12. Productization – ADMS Realtime Bus (DNP3, ICCP) Utility Management Systems DMSSCADA OMS Control Center Switching and Logging EMS Simulation Modeling DEM Asset Mngmt Mobility Feeder Automation Substation Automation Transmission Enterprise Bus (IEC CIM) ERP Energy Market GIS Network Management AMI Weather MDM Behind theMeter UMS Common Platform
- 13. Productization – VVO Closed loop ●Configuration ●Profiles – presets of VVO configuration managed by grid engineers ●Control & Monitoring ●24/7 operation control ●Automatic verification of command execution ●Overview of devices availability and commanding ●Modes: Loss Optimization, DSDR, Emergency, Storm ●Manual execution on request ● Off-line and real-time analysis of VVO decisions
- 14. Productization – VVO Closed Loop DMS Distribution Management System 1. Field data collecting 2. Real-time analysis 3. Decision making 4. Commands executing Regulator Capacitor SUB DSCADA Tap Lines Phase Additions Sensors AMI
- 15. Productization – VVO & DSDR ●VVO optimizes network state in all load conditions ● Normal load ●Reduces losses, provides VAR support ●Energy savings vs. energy charging ● Heavy load ●Peak load shaving ●Avoids high cost of peak spinning reserves ●Avoids load shedding in emergency situations ●Achieved benefits ● Loss Optimization – continuous VAR support for transmission needs ● DSDR – 250 MW demand reduction benefit – summer 2014 ● Emergency – level 2 implements 250 MW reduction during cold wave – January 2014
- 16. Productization – VVO & DSDR WP Losses WP Benefits Achieved
- 17. Productization – Network model data challenge ●Quality of calculated state depends on accuracy of GIS data ●Good quality of calculated state needed for VVO multi-objective decisions (LO, DSDR) ●To include more VVO single objective approaches based on basic network model & topology, SCADA measurements, smart meters ● Implementation of VVO in phases, on different parts of network ● Gaining benefit in earlier phases prior to data improvement ●To support automatic detection of parts of network ready for multi-objective VVO with full benefit
- 18. Productization – Conclusions ● ADMS – comprehensive, real-time solution for network management and design ● Productization of ADMS with VVO Closed Loop – DSDR ● VVO – continuous process that optimizes power grid 24/7 ● DSDR project – measurable and verified benefits of implementation of VVO ● Loss Optimization/VAR support ● DSDR – peak shaving ● Emergency ● New ideas for software solution from DSDR project ● Automatic support for different quality of network model data ● New ways of calculating and reporting benefit 18
- 19. 19 Jeremy D. Carden, P.E. jeremy.carden@duke-energy.com Dragan Popovic, Ph.D. dragan.popovic@schneider-electric-dms.com