Performance Simulation of Energy Storage Technologies for Renewable Energy Integration

Transcription

1 Performance Simulation of Energy Storage Technologies for Renewable Energy Integration Cesar A. Silva Monroy Ph.D. Student Electrical Engineering University of Washington Energy Seminar October 8, 2009

2 Overview Introduction Power System Applications Modeling Pumped Hydro Energy Storage Compressed Air Energy Storage (CAES) Batteries Superconducting Magnetic Energy Storage (SMES) Flywheels Ultracapacitors Conclusions References

3 Introduction Renewable energy resources such as wind and solar are stochastic in nature Current power systems must keep the power balance between generation and demand (+ losses): P demand = P generation Power imbalance between demand and generation is aggravated by stochastic resources Energy storage can change the way we operate power systems Future power system will need to keep energy balance: E demand = E generation Energy Storage has the potential to enable high penetration of renewable energy resources

4 Power System Applications Load leveling Investment deferral Active and reactive power flow control Emergency power supply Focus is wind and solar integration: Generation shaping

5 Generation Shaping Wind energy is random, intermittent, over large scales and short times (10 minutes) Load is slowly varying over 10 minutes Wind variation must be met by change in controllable output Generation kept on line and off market to provide response to wind costs money and emissions

6 Generation Shaping Storage a solution P P t t P P t t P t Storage

7 Generation Shaping Benefits Smooth, controllable wind farm output Reduces wind farm transfer requirement Issues Adds to wind farm costs, and thus cost of wind power Regulation currently estimated to add 10% to cost of wind not enough to pay for storage

8 Modeling Generic model Employed for optimization of power system operation Time frame: minutes years No transient behavior Capture minute to minute variations State variables

9 Modeling Parameters Energy Capacity Power input and output capacities Efficiencies: Charge, Discharge, Selfdischarge Life cycling characteristic Minimum charge Other parameters particular to each technology (Resistance, Mass, etc.)

10 Modeling Input variables: Power input Power output Time step Output variables: State of charge Emissions (NOx, SOx, CO 2 ) Number of cycles

11 Ideal Energy Storage Template for developing specific models 100% efficient Infinite charge/discharge capabilities High energy density (energy/volume ratio) Infinite life time Zero emissions

12 Ideal Energy Storage Charge: T s : Time step E = E 0 +P in T s E: energy stored after T s E 0 : energy stored before T s P in : Power input Discharge: State of charge: 1 SOC 0 E = E 0 -P out T s SOC = E/E max

13 Ideal Energy Storage Number of cycles N c : N c = N 0 +PT s /2E max 1 cycle = 1 charge and 1 discharge Efficiency Charge: Discharge E = E 0 +P in T s η c E = E 0 -P out T s /η d

14 Pumped Hydro Energy Storage Hydraulic potential energy E = mgh Charging: Pump water to a higher level reservoir Discharging: Use stored water to run turbines connected to electric generators 1. Transmission 2. Transformer 3. Motor-generator 4. Lower reservoir 5. Tail race 6. Pump-turbine 7. Penstock 8. Upper reservoir 9. Local loads Diagram of pumped hydroelectric energy storage [1]

15 Pumped Hydro Energy Storage Capacity: given by volume Response times are from 1 to 10 min to go from full load to full generation Pumping efficiency is modeled as charge efficiency Generating efficiency is modeled as discharge efficiency Water evaporation is modeled as the selfdischarge rate (very low) No cycling effects No emissions

16 CAES Concept Stores energy in the form of a compressed gas: E = PV ln(p in /P out ) Charging: Air is compressed in natural or artificial underground caverns Discharging: Compressed air is released to in the combustion process of a natural gas turbine (diabatic storage) CAES reduces overall fuel consumption CAES concept plant (Norton mine) [2]

17 CAES Characteristics Capacity: limited by size and conditions of storage cavern (up to thousands of MWh) High power output ramp rate (30% of maximum load per minute) Compression process is complex to model About 0.75 MWh of energy are needed to store enough air for 1 MWh of energy released: Lossless charge process Discharge process: E = E 0 -P out T s η d No cycling effects There are emissions associated with generation

18 Batteries Chemical potential energy Discharge: electrons flow from anode to cathode, anode material is oxidized, cathode material is reduced Charge: Current flow is reversed, anode material is reduced, cathode material is oxidized [3]

19 Batteries Assumptions: Current is distributed evenly through all cells in stack All cells have the same SOC at all times All cells have the same capacity Capacity: given by amount of cells in series and parallel Fast power response, in the range of seconds Power converters efficiency are around 90% Self-discharge

20 Batteries Life cycling depends on type of battery: Lead-acid Sodium-Sulfur Vanadium redox (Reflow) Losses depend on voltage and current Equivalent circuit:

21 Batteries Lead acid: OCV = 2.1 V Internal resistance increases with number of cells in series, decreases with number of cycles Voltage decreases linearly Capacity decreases exponentially with number of cycles Energy available decreases with higher output currents (Peukert number k) C r = I k T s k =

22 Batteries Sodium-Sulfur OCV = 2.08 V Internal resistance increases with number of cells in series, decreases with number of cycles Voltage is constant up to DOD of 60-75% Voltage drops linearly for DOD > 60-75% Capacity decreases linearly with number of cycles Peukert effect

23 Batteries Vanadium redox (Reflow) [4]

24 Batteries Energy capacity is limited by reactant tank volumes Power capabilities are limited by number of cells Auxiliary equipment losses OCV = 1.4 V Output voltage: V = OCV +2RT/F ln(soc/(1-soc)) No Peukert effect No cycling effect

25 SMES Stores energy in the magnetic field formed by a dc current circulating in a superconducting magnetic ring E = 0.5 LI 2 Experimental SMES composition [1]

26 SMES Capacity: given by power conversion or coil ratings Very high power capabilities Losses: Power conversion Refrigeration losses: assumed constant Self-discharge values are high if pumps are kept on

27 Flywheels Rotational kinetic energy: E = 0.5Jω 2 Charge: motor accelerates spinning mass (rotor) Discharge: use inertia of rotating mass to drive generator Power conversion system needed Cross-section of a flywheel [5]

28 Flywheels Capacity: given by maximum rotational speed Very high power charge/discharge capabilities Losses: Power conversion system Bearings friction losses can be calculated as function of friction moment Operation of magnetic bearings or low viscosity fluids cause parasitic losses No cycling effects No emissions

29 Ultracapacitor Electric potential energy: E = 0.5CV 2 Charge/discharge: constant current, voltage or power Uses double layer effect [5]

30 Ultracapacitor Model as a capacitor with a series resistance Energy capacity is increased by adding capacitors in series and parallel Very high power capabilities Additional losses due to power conversion No cycling effects Very low self-discharge

31 Summary Technology Emax Pout Losses Cycling Other PHES Reservoir volume Slow η p, η g, selfdischarge No effects CAES Cavern volume Medium η d No effects Emissions Batteries Cell number High Resistive, PC, SD Lifetime decreases Peukert effect Reflow Cell number High Resistive, PC, SD, parasitic No effects SMES Coil rating High PC, Refrigeration, SD, PC No effects Flywheel Rotational speed High Parasitic, friction, SD, PC No effects UC Capacitor ratings High Resistive, PC No effects

32 Conclusions Simulation of energy storage technologies can be carried out with a set of defined parameters Pump-hydro, CAES and Batteries are largescale storage Future work Include cost models Optimal operation Optimal location Optimal size

33 References 1. A. Ter-Gazarian, Energy Storage for Power Systems, Peter Peregrinus, on.htm 3. D. Linden, T.B. Reddy, Handbook of Batteries, 3rd edition, McGraw-Hill, ES_Summer2001/Miyake.pdf 5. Handbook of Energy Storage for Transmission and Distribution Applications, EPRI - DOE, Washington D.C., 2003

34 QUESTIONS?

35 Load leveling P Daily Load Shape time

36 Load leveling

37 Load leveling Benefits Supply cheap off-peak power to on-peak times Keep base load units on line during off-peak Issues Need high price differential to be economic Round trip efficiency must be high Enables base load - CO2 release may increase Daily load shape sets storage and power requirements Major motivator for existing storage facilities

38 Investment Deferral Idea: Optimal utilization of transmission investment Transfer Only a few hours at maximum load % Above

39 Investment Deferral Storage allows line to operate closer to average power output Transfer Storage % Above

40 Investment Deferral Benefits More capacity (MWh transferred) from same line Can defer transmission construction Transmission losses reduced for same energy transfer Also provides peak shaving benefits Issues How does storage capture value of investment deferral and reduced losses in deregulated market?