Using Active Customer Participation in Managing Distribution Systems Visvakumar Aravinthan Assistant Professor Wichita State University PSERC Webinar December 11, 2012
Outline Introduction to distribution advancement Limitations with current operation states o Some examples Improving reliability of the systems Active consumer participation How to unify consumer participation with distribution operation 2
Smart Grid Distribution Advancement Introduction 3
Smart Grid What would be new in smart grid 1 o Self-healing from power disturbance events o Enabling consumer active participation o Resilient against physical and cyber attack o Power quality for 21st century needs o Accommodating all generation and storage o New products, services, and markets o Optimizing assets and operating efficiently [1] Department of Energy, Online: http://energy.gov/oe/technology-development/smart-grid 4
Current State Self-healing Consumer participation Physical and cyber attacks Power quality Generation and storage Markets Asset Management Gen. Trans. Dis. Con. 5
Distribution: What Can Be Done? ISO Price Info. Emergency Operation Load expectation State of operation Distributed Resources Market enabled - Flexible grid - Efficiency New Operation - Measurements - Communication - Control paradigms - Components - Data management Emission Mitigation Reliability - Component life - Consumer satisfaction DSM - Consumer acceptance - Price elastic load - Data sharing issues Directives - Distribution pricing - Direct control Self healing - Consumer awareness - DSM to manage load shedding 6
Distribution: Advancement Distribution Advancement 7
Distribution: Advancement Distribution Advancement 8
Distribution: Advancement Distribution Advancement 9
Distribution: Advancement Distribution Advancement 10
Distribution: Advancement Distribution Advancement 11
Consumer Participation DSM Objectives Load Shape Request Data Request 12
Consumer Participation DSM Objectives Load Shape Request Data Request Regulatory Requirements 13
Consumer Participation DSM Objectives Load Shape Request Data Request Consumer Privacy Regulatory Requirements 14
Consumer Participation DSM Objectives Load Shape Request Data Request Consumer Privacy Regulatory Requirements Impact Analysis Utility Cost Benefit Analysis 15
Distribution Operation Examples 16
In a power system Reliability o Lots of data available o Little information extracted Example 1: Moghe et. al. Example 2: Russell et. al. Low level anomalies for 6 days Animal Contact Power restored in 1 hour [1] R. Moghe, M. Mousavi, J. Stoupis, J. McGowan, Field investigation and analysis of incipient faults leading to a catastrophic failure in an underground distribution feeder, in Proc. of Power Systems Conference and Exposition (PSCE), Seattle, Washington, May 2009 [2] D Russell, R. Cheney, T. Anthony, C. Benner, C. Wallis and W. Muston, Reliability Improvement of Distribution Feeders, In proc. 2009 IEEE PES General Meeting, Calgary Canada, July 2009. 17
Reliability Electric Vehicle Charging o Different charging loads on a distribution transformer Power Consumption (kw) 1.5 1 0.5 No Electric Vehicles All Charging at Same Time 1/2 hour delay in charging Random Charging late Night (Controlled) 0 0:00 6:00 12:00 18:00 0:00 o Loss of life of distribution transformers [1] S. Argade, V. Aravinthan, and W. Jewell Probabilistic Modeling of EV Charging and its Impact on Distribution Transformer Loss of Life, in Proc. 1 st IEEE International Electric Vehicle Conference, March 2012 18
Reliability Voltage regulator operations o Distributed generation at feeder/lateral level o IEEE 13 bus system BASE CASE 1.0 LOAD PROFILE TAP CHANGES 0.8 SITE1-VA_WF (V) 0.6 0.4 SITE1-VA_WF (V) 0.2 0.0 0.9 0 500000 1000000 1500000 Time (ms) Electrotek Concepts TOP, The Output Processor o Distributed solar PV at 40% penetration BASE CASE WITH PV PV LOADSHAPE TAP CHANGE 1.0 1.1 PV LOADSHAPE TAPCHANGE 2.0 SITE1-VA_WF (V) 0.8 0.6 0.4 0.2 SITE1-VA_WF (V) LOADSHAPE-VA_WF (V) 1.5 1.0 0.5 TAPCHANGE-VA_WF (V) Increase 0.0 0.9 0 500000 1000000 1500000 Time (ms) Electrotek Concepts TOP, The Output Processor 0.0 0.9 0 500000 1000000 1500000 Time (ms) Electrotek Concepts TOP, The Output Processor [1] V. Ravindran, V. Aravinthan, and W. Jewell Impacts of High Penetration Distributed PV Sources on Voltage Regulation, in Proc. 43 rd Frontiers of Power Conference, Oct. 2012 19
Distributed Generation Impacts of geographically scattered DGs o Voltage rise with 30% PV penetration on IEEE 123 test feeder [1] V. Ravindran, V. Aravinthan, and W. Jewell Impacts of High Penetration Distributed PV Sources on Voltage Regulation, in Proc. 43 rd Frontiers of Power Conference, Oct. 2012 20
Base Case Distributed Generation 30% PV Penetration Note: Red below 1 p.u, Green 1-1.02 p.u, Blue above 1.02 p.u [1] V. Ravindran, V. Aravinthan, and W. Jewell Impacts of High Penetration Distributed PV Sources on Voltage Regulation, in Proc. 43 rd Frontiers of Power Conference, Oct. 2012 21
Distribution Automation Location on the Feeder and the Frequency o 5 houses connected to a single transformer 5 houses connected to a transformer Hot Summer day in Kansas 1 minute average 22
Distribution Automation Location on the Feeder and the Frequency o 5 houses connected to a single transformer 5 houses connected to a transformer Hot Summer day in Kansas 5 minute average 23
Distribution Automation Location on the Feeder and the Frequency o 5 houses connected to a single transformer o Is the missed information useful 24
Future Needs Distribution Reliability System Requirements Dynamic Pricing Consumer Participation How to connect distribution necessities with active consumer participation o Utility Improve distribution system operation with better observability Connection between DG to load o Consumer Looks for maximum satisfaction Would not like to share the information 25
Smart Grid Distribution Operation Reliability Based Operations 26
Condition Assessment To improve distribution reliability requires a tool to determine condition of components o Lack of communication limits assessments Observing failure modes improve assessment o Identify criteria that are observable General Winding Condition Oil Condition Physical Condition Criterion Age of the Transformer Experience with Transformer Noise Level Loading Condition Core & Winding Losses Winding Turns Ratio Condition of Winding Condition of Solid Insulation Partial Discharge (PD) Test Gas in Oil Water in Oil Acid in Oil Oil Power Factor Condition of Tank Condition of Cooling System Condition of Tap Changer Condition of Bushing [1] V. Aravinthan, W. Jewell, and W. Jewell Identifying worst performing components in a distribution system using Weibull distribution, in Proc. 11 th International Conference on Probabilistic Methods Applied to Power Systems, June. 2010 27
Condition Assessment Develop a failure rate function for each criterion using o Historic data if available o Else, standards or guidelines if available o Else, hypothetical functions (experience) Historic Data (Transformer) o Example: Age of the component [1] V. Aravinthan, W. Jewell, and W. Jewell Identifying worst performing components in a distribution system using Weibull distribution, in Proc. 11 th International Conference on Probabilistic Methods Applied to Power Systems, June. 2010 28
Condition Assessment Develop a failure rate function for each criterion using o Historic data if available o Else, standards or guidelines if available o Else, hypothetical functions (experience) Historic Data (Transformer) o Example: Gas in the oil Standards Eg: IEEE std. C57.104-2008 Status TDCG (ppk) Remarks 1 < 0.72 Normal aging of oil 2 0.72 1.92 excess oil aging 3 1.92 4.63 Excessive oil aging 4 > 4.63 Very poor oil condition Define R(t) for 2 status or Define R(t) for 1 status and 1 parameter Find the unknown parameters [1] V. Aravinthan, W. Jewell, and W. Jewell Identifying worst performing components in a distribution system using Weibull distribution, in Proc. 11 th International Conference on Probabilistic Methods Applied to Power Systems, June. 2010 29
Condition Assessment Develop a failure rate function for each criterion using o Historic data if available o Else, standards or guidelines if available o Else, hypothetical functions (experience) Historic Data (Transformer) o Example: Location of the transformer No enough information F Total no of transformers failed s Total no of similar transformers handled S F Total no of similar transformers failed S U Total no of similar transformers with unknown cause [1] V. Aravinthan, W. Jewell, and W. Jewell Identifying worst performing components in a distribution system using Weibull distribution, in Proc. 11 th International Conference on Probabilistic Methods Applied to Power Systems, June. 2010 30
Condition Assessment Problem: Not all criteria have equal influence on component failure!!! Solution: Use weighted reliability function o Weighted Reliability Function Once the weighted reliability functions are known o Series parallel topology for component Quantitative: Component Condition Score Qualitative: Component Condition Report: Example: Distribution Transformer [1] V. Aravinthan, W. Jewell, and W. Jewell Identifying worst performing components in a distribution system using Weibull distribution, in Proc. 11 th International Conference on Probabilistic Methods Applied to Power Systems, June. 2010 31
Condition Assessment Defective 90 80 % Normal 100-90 % Fair Mild Satisfactory Stable Serious Critical Extremely Critical Faulty 20 10 % Failed 10 0 % Age: 18 yrs TDCG: 1.8 ppk S F =40, S U =10, F=90 & s=60 [1] V. Aravinthan, W. Jewell, and W. Jewell Identifying worst performing components in a distribution system using Weibull distribution, in Proc. 11 th International Conference on Probabilistic Methods Applied to Power Systems, June. 2010 32
Electric Vehicle Charging Assumed 20% EV Penetration in Busses Zone 3, 4, 5. Type 1 charging assumed, slow charging will contribute to minimum impact on the system Renewable generation / storage is included to at Bus 8 for the 3 rd part 13 Bus IEEE Test Feeder 24 Bus IEEE Reliability Test System 33
Electric Vehicle Charging Two levels of optimization, o Level 1: Schedule day ahead charging (request sent by consumers in advance) Objective: Minimize the system average interruption duration index (SAIDI) (maximize performance) Constraints: Transmission congestion All vehicles requesting charging are charged All vehicles are charged when they are available None of the system components are overloaded 34
Electric Vehicle Charging Two levels of optimization, o Level 2: Find the maximum number of vehicles charged in real time Objective: Maximize the number of vehicles that could be charged Constrains: Acceleration of loss of life of the transformer Maximum cap on the CO 2 emission Optimum number of vehicles from level 1 is charged 35
Electric Vehicle Charging Part 1: No renewable, same level of CO 2 emission as traditional vehicles allowed Zone 5: Moderately loaded feeder section 36
Electric Vehicle Charging Part 2: With renewable 80% of CO 2 emission as traditional vehicles allowed Zone 5: Moderately loaded feeder section 37
Smart Grid Distribution Advancements Consumer Participation 38
Active Consumer Participation Coordinating EV charging o Develop a price model to control the EV charging time Assume that there are number of vehicles that could be charged without degrading the performance at time i Vehicles could schedule charging time one day ahead What if there are more vehicles wanting to be charged o Two level of pricing one for vehicles scheduled other of the additional vehicles Objective is to minimize both the prices 39
Active Consumer Participation Limiting Factors o Consumers prefer to charge at convenience Generally consumer anxiety increases if the charging is delayed o Limit consumers who are not satisfied More charge more anxiety More availability less anxiety 40
Active Consumer Participation Limiting Factors o Price: > + + Reference Price Additional Power Loss due to Large Loads Distribution Overloading o Component Condition Most critical component: Transformer based on IEEE std. C51.97 transformer hotspot temperature should be limited to + + < Ambient Temp. Top oil temp.. rise over ambient Hot spot temp. rise over top oil 41
Active Consumer Participation How consumer anxiety affects additional EVs connected to the grid 42
Active Consumer Participation Distributed generation for improvement in performance o Example: Minimize the feeder power loss with the DG penetration Using exact lumped model o Allow DGs with active power control mode o Reactive power is supplied to minimize power loss o But maximum power factor is limited at generation 43
Active Consumer Participation For the IEEE 13 bus feeder Power Loss (Kw) 100 75 50 25 0 6 11 16 At 0.9pf limit At 0.95pf limit Time (h) For the IEEE 34 bus feeder 30 Power Loss (Kw) 20 10 0 6 11 Time (h) 16 At 0.9pf limit At 0.95pf limit 44
Smart Grid Distribution Operation Connecting Both Together 45
Unification Consumer participation Load Load Load Shifting Time of the Day Flexible Loading Time of the Day EV charging Distribution Transformer Condition (Reliability) based reconfiguration Operating beyond IEEE 1547 (reactive power control) 46
Thank you visvakumar.aravinthan@wichita.edu 47
Support Slides 48
Condition Assessment Lets assume oil is bad and TDCG 4 ppk Normal 100-90 % Fair Mild Defective 90 80 % Satisfactory Stable Serious Critical Extremely Critical Faulty 20 10 % Failed 10 0 % Criterion Weight R(t) Faults seen by the transformer 0.70 0.80 0.860 Geographical location 0.60 0.90 0.940 loading 0.80 0.80 0.840 Age 0.90 0.76 0.784 Noise 0.40 0.90 0.960 Condition of winding 0.90 0.80 0.820 PD test 0.50 0.82 0.910 Core and winding loss 0.80 0.80 0.840 Condition of solid insulation 0.80 0.88 0.904 Tap changer condition 0.60 0.91 0.946 Winding turns ratio 0.70 0.95 0.965 Gas in oil 0.90 0.12 0.108 water in oil 0.90 0.87 0.883 Acid in oil 0.90 0.89 0.901 Oil PF 0.90 0.90 0.910 Tank condition 0.90 0.92 0.928 bushing condition 0.90 0.90 0.910 hot spot temperature 0.70 0.80 0.860 cooling system 0.70 0.80 0.860 [1] V. Aravinthan, W. Jewell, and W. Jewell Identifying worst performing components in a distribution system using Weibull distribution, in Proc. 11 th International Conference on Probabilistic Methods Applied to Power Systems, June. 2010 Experience 0.50 0.03 0.515 49
Electric Vehicle Charging Part 2: No renewable 80% of CO 2 emission as traditional vehicles allowed Zone 5: Moderately loaded feeder section 50
Active Consumer Participation How location would influence the price 51
Active Consumer Participation For the IEEE 13 bus feeder Power (p.u.) 1 0.8 0.6 0.4 0.2 0 Load1 PV 0 4 8 12 16 20 24 Time (h) Current (Ampere) 200 150 100 50 0 Iag1 Iag2 Iag3 6 11 16 Time (h) Current (amperes) 200 150 100 50 0 Iag1 Iag2 Iag3 6 11 16 Time (h) 52