FEDERAL UTILITY PARTNERSHIP WORKING GROUP SEMINAR April 22-23, 2015 Nashville, TN A Look at Secondary Use Energy Storage Michael Starke, PHD Oak Ridge National Laboratory Hosted by:
Project Overview
• Supporting the industry investigation into vehicle battery secondary-use through testing, demonstration, and modeling. – Potentially a cost competitive energy storage technology – Validate reliability and safety – working with industry to troubleshoot and test systems under operational conditions – Examining regulatory environment – investigating hurdles that are institutional – Industry acceptance – build confidence in this technology.
2 Secondary Use of EV Batteries
• Potentially significant electric vehicle market. – Projections from different studies show significant growth. – March 2014, Tesla announces news on the building of a Gigafactory with projections of 500,000 vehicle production capability by 2020. – June 2014, Tesla is releasing all patents to encourage electric car production • What can we do with the on- board battery technologies?
Repackage/Reuse: Could provide a low-cost grid storage solution (if design of repackaged system does not require significant modifications and added expense.)
Already Available in USA • Over 150,000 plug-in electric vehicles (PEVs) currently in USA (study by UCLA Luskin Center for Innovation – December 2013) – ~ 55% of PEVs are PHEV and 45% are BEV – Near 70% of these vehicles are Nissan Leaf, Chevy Volt, or Tesla
Nissan Leaf Chevy Volt Tesla Nearing 40,000 Exceeding 50,000 Vehicles Nearing 20,000 Vehicles 16.5 kWh per pack Vehicles 24 kWh per pack ~825MWh 85 kWh per pack ~960MWh ~1700MWh
• Leads to a an estimated 3.485GWh of existing battery storage. • Estimates on capacity of the batteries. Detailed analysis will need to consider operational constraints, BMS level limits, and other aspects.
Demonstration Sites: Repurposing of Batteries • Utilizes BMW mini-E batteries and BMS/Princeton Power Systems interface hardware • 108 kW/180kWh with DC coupling to PV
• Utilizes General Motors Volt batteries and BMS/ABB interface hardware. • 25kW/50kWh system connected to ORNL test-bed, PV smoothing and shifting. Current Activities
SYSTEMS INTEGRATION An effective partnership that merges equipment, technical know-how, and infrastructure: • Energy Storage – Used EV Batteries HARDWARE SOFTWARE • Energy Management System • Electric Grid
ORNL is testing and demonstrating the technology as a third party. The Technology
GM Chevy Volt ABB Enclosure Battery
Re-Packaged
Automotive Application Grid Application(25kW/50kWhr) – Capacity for 10 Years in – Expected capacity for 10 Years Automotive Application of Operation – Power 111kW – 5 Volt Battery Packs – Liquid Cooled / Heated – 5 kW per Volt Battery – Air Cooled/Heated The Working System
Zone 1: The system has a single- Islanding Contactor phase connection with the grid, PV PV Array Array, AC breakers, islanding
240V contactor, and voltage sensing. Grid 1ϕ Load Zone 2: Inverter measures and senses inputs to control charging HVAC and discharging needs (4 Zone 1 Unit Interlocks Zone 4 quadrant) Voltage Sense AC Input Batt. Door Inv. Door Estop Door/Estop Sw. Switch Switch Zone 3: Batteries connected on DC Inverter Intput Zone 5 link and controlled by BMS. BMS DC Output CAN Comm. Zone 2 uses voltage, current, and CAN Comm. Door/Estop Sw. Battery Stack Intput temperature information to relay Temperature Sensors Battery Voltage/Current control information to inverter. Sensors Management Contactor Control System Cell Zone 4: Safety interlocks to Modem prevent unsafe access Zone 3 Zone 5: Thermal management with fans, heaters, and HVAC. Multi-tiered layers of security are present in the system to ensure a safe operation System Benefits: CES Local benefits: Grid benefits: Real and Reactive Power Support Firming and shifting Renewables and Load leveling • demonstrate that load factor and power factor can / T&D Deferral be maintained. • battery can charge/discharge depending on control and load behavior. Service reliability • during outage, CES unit can still supply load for a Ancillary Services period of time. • regulation/spinning ISO request (ancillary Phase balancing services)
• if three units are installed (each on separate CES phases) additional energy can be used to balance Command and Control phases.
Renewables n
Transformer o i
CES t a t s b u Bi-directional S smart meter EV/PHEV
Junction Box CES Unit DC/AC Repurposed CES AC/DC Battery Disconnect Converter Pack switch Similar benefits can be realized by distributed energy storage for commercial applications Testing Setup at ORNL • ORNL objective for testing: Provide real world examination systems integration and applications with the flexibility to capture many different case scenarios.
ORNL Distribution 13.8kV, 3 Phase, 60Hz 480V System 50kVA inverter Distribution 480V/2.4kV 750kVA, 13.8kV/2.4kV Panel
2.4kV System 480V, 3 Phase, 60Hz Circuit #2 50kW PV 2.4kV, 3 Phase, 60Hz 37.5kVA, 120V-240V/ 750kVA, 2.4kV/480V 480V
Disconnect Switch 480V, 3 Phase, 60Hz 240V, Split Phase, 60Hz Disconnect Community Energy Switch Storage (240V, Split Phase, 25kW) 15kW PV Programmable Load Bank (240V, 1 Hardware Split Phase, 2 Communications 24kW) 3 Controls
Hard/Soft: Communication and Control Communications and ROOFTOP PV DECC
Facility Control
CT/VT and Measure & Validate
ORNL/ Distribution • Communications and GRID control done through
CES Serial, Modbus over Serial, and TCP/IP
CT/VT
• All integrated through Matlab/Labview
• Load Bank utilized for LOAD CT/VT Communications Emulation. Cable, RS232/RS485 over Modbus M&V
Communications Cable, RS232 CONTROL/COMPUTING Hardware: Equipment Inside DECC
480V/240V(split-phase) Transformer
Emergency Disconnect
Programmable Load Bank Islanding Contactor/Relay Hardware: Equipment Outside DECC
Emergency Disconnects
PV Array
Battery Inverter Enclosure Interface Manual Control Set of pre-programmed controls
CES Alarms State display Controls and Programs • Auto-runs at 12:00AM • Controls depend on selected settings. Cloud Cover Temperature Data Processing % Cloud Cover Temperature (C)
Data
solar irradiance
Historian Load Bank Temp Solar Irradiance/ Residential Model Storage PV output Consumption Temperature (C) Data kW PV
Control Mode, P, Q kW load
SOC Estimate
Load Factor Control Points
PV Forecast Emergency Main Control GA Optimization Monitoring Load Bank text message email Measured Data
Data Acquistion
Shutoff Measurements and Simulation Additions • Load Bank is controlled to Load Bank Interface follow residential load profiles through macros. • Residential profiles are developed through modeling and historical data collection.
Residential Model Consumption
Macro is running
Power Consumed by Bank 1 Residential Modeling
• Residential data has been QINTER
Tindoor sub-metered and collected QHVAC QSOLAR for several years. Used to Cair Rin = 1/UAinsul
develop and validate load 0.95 Rin = 1/UAmass models. Tamb
• Markov Chains are used to Sleeping Tmass
0 5 .0 2 1 .0 5 0 drive residential loads such Cmass Home 5 0 3 .0 .0 5 as washer/dryer/water 0 0.125 Model/HVAC heaters… Grooming Laundry
0 .8 0.15 8 5 0. Activity Simulation Markov Chains PV Forecasting for Optimization
Collect Cloud Forecast
Cloud Forecast
Neural Networks Weather Historical Data (Irradiance Underground Forecast) Ambient Temp Solar Irradiance
Solar Panel/ Module Temp Solar Thermal Model Model
PV Curves
Maximum Value
MMPT (Power) Testing Procedure (Systems Tests) • Objectives: Start Requested Power: 5 – Obtain standard metrics (round-trip kW
Charge Battery at efficiency/ensure within bounds of Requested Power Level to 70% SOC standards) Rest Battery for 30 – Demonstrate application examples Minutes Discharge Battery at Increase Requested Requested Power Power Level • Standard Metrics: Level for 30 minutes
Rest Battery for 30 – Round-trip efficiency Minutes
– Harmonics, etc. No Time Since Last Power Level Change > 24 Hours?
• Applications Yes
No – Load factor, At Maximum Power Level? – Power factor, Yes – Renewable Integration, Stop – Islanding
Multiple Value Streams: Stacking Benefits (Load Factor/Power Factor, Renewable Integration)
Grid (nearly flat)
Power (W) Power (Var)
Time (hr) Time (hr)
SOC (%) Histogram power factor SOC target to return to 50%
Target Set to 0.97
Time (hr) Time (hr) TE: PV Smoothing/Capacity Firming
Power (kW) Objectives: Integrate PV by removing oscillations and error in forecast.
Benefits: 1) Removing oscillations in PV output can impact local voltage. 2) In some cases these oscillations lead to 03/22/2014 14 significant tap changes Total PV Power Power (kW) CES Power in transformers. 10 Net Power Predicted PV Power Smoothing this 6 behavior with storage can extend transformer 2 Power (kW) life. -2
-6 8 10 12 14 16 18 20 Time (Hr) TE: Islanding Mode Battery-supplying Objectives: entire load Utilize storage for emergency backup power
Benefits: 1) Provides power during an outage 2) Can be used to Grid power support contingency type events as well to reduce load consumption.
Initial Economic Approach
Battery Optimal Model Battery Dispatch (Mixed Integer) Grid Linear Services Optimization
Cost/Savings Data Initial Economic Results
• Arizona Public Service Company residential rate structures • Year-long simulated load for 3 homes • Dispatch the battery to minimize the homeowners’ cost • Utilized efficiencies of real system, 10year/3000 cycle battery Initial Economic Results Future Tasks
<500VDC, G1, G2, G3, G4, G5, G6 • Modeling and economics 25kW Preq, Qreq, Mode
Estop1, Inverter Control Va, Vb, Vc BMS SOC1, G11 S11 System kWh1, Ia, Ib, Ic kW1 L1 I1 C1 assessment for DES. Estop1, SOC1, kWh1, kW1 G11, G12 V1, I1 V1 G12 S12 Battery DC/DC Converter 1 • Development of
BMS G21 <500VDC, S21 DC/AC Inverter/Grid Connection
25kW L2 G1 I2 C2 refurbished secondary S1 S2 S3 G22 S22 Battery V2 G1 G2 G3 3ϕ 480VAC La Ia Va 600VDC Lb 100kW Ib Vb Lc use ES. Ic Vc
BMS G31 <500VDC, S31 S4 S5 S6 G5 25kW L3 G4 G6 I3 C3
G32 S32 Battery V3
BMS G41 <500VDC, S41
25kW L4 I4 C4
G42 S42 Battery V4
BMS <500VDC, G51 S51 25kW L5 I5 C5
G52 S52 Battery V5
26