Virtual Energy Storage through Distributed Control of Flexible Loads CaFFEET 2015 Innovative Solutions to Integrate Renewable Energy Ana Buˇ si´ c Inria and ENS – Paris, France Thanks to my colleagues, Prabir Barooah and Sean Meyn, and to our sponsors: French National Research Agency, National Science Foundation, and Google
27 March 8th 2014: Impact of wind Load and Net-load 25 and solar on net-load at CAISO 23 GW 21 19 $/MWh 17 1200 Price spike due to high net-load ramping 15 Toal Load Net-load: Toal Load, less Wind and Solar 1000 need when solar production ramped out Wind and Solar 4 GW 800 2 Negative prices due to high 0 600 Peak ramp Peak 24 hrs mid-day solar production Toal Wind Toal Solar 400 200 0 Ramp limitations cause price-spikes -200 Peak ramp Peak 24 hrs Challenges
Challenges Some of the Challenges 1 Ducks MISO, CAISO, and others: seek markets for ramping products 2 / 15
Challenges Some of the Challenges 1 Ducks 2 Ramps 4 G W ( t ) = Wind generation in BPA, Jan 2015 3 GW 2 Ramps 1 0 Jan 01 Jan 02 Jan 03 Jan 04 Jan 05 Jan 06 2 / 15
Challenges Some of the Challenges 1 Ducks 2 Ramps 3 Regulation October 20-25 October 27 - November 1 8 8 Load Generation and Laod GW 6 6 Generation Hydro GW Thermal Wind 4 4 2 2 0 0 1 1 Regulation GW Error Signal in Feedback Loop 0.8 0.8 Regulation GW 0 0 -0.8 -0.8 -1 -1 Sun Mon Tue Wed Thur Fri Sun Mon Tue Wed Thur Fri 2 / 15
Challenges Some of the Challenges 1 Ducks 2 Ramps 3 Regulation One potential solution: Large-scale storage with fast charging/discharging rates 2 / 15
Challenges Some of the Challenges 1 Ducks 2 Ramps 3 Regulation One potential solution: Large-scale storage with fast charging/discharging rates Let’s consider some alternatives 2 / 15
4 G r ( t ) 3 G r = G 1 + G 2 + G 3 GW 2 G 1 Traditional generation G 2 DD: Chillers & Pool Pumps 1 G DD: HVAC Fans 3 0 Jan 01 Jan 02 Jan 03 Jan 04 Jan 05 Jan 06 Virtual Energy Storage
Virtual Energy Storage Control Architecture Frequency Decomposition Gas Turbine BP Coal Batteries BP Control Flywheels Power Grid Σ BP LOAD C H Voltage − BP Frequency Phase BP C A Actuator feedback loop Today: PJM decomposes regulation signal based on bandwidth, R = RegA + RegD Proposal: Each class of DR (and other) resources will have its own bandwidth of service, based on QoS constraints and costs. 3 / 15
Virtual Energy Storage Frequency Decomposition Taming the Duck 27 March 8th 2014: Impact of wind Load and Net-load 25 and solar on net-load at CAISO 23 GW 21 19 $/MWh 17 1200 Price spike due to high net-load ramping 15 Toal Load Net-load: Toal Load, less Wind and Solar 1000 need when solar production ramped out Wind and Solar 4 GW 800 2 Negative prices due to high 0 600 Peak ramp Peak 24 hrs mid-day solar production Toal Wind Toal Solar 400 200 0 Ramp limitations cause price-spikes -200 Peak ramp Peak 24 hrs ISOs need help: ... ramp capability shortages could result in a single, five-minute dispatch interval or multiple consecutive dispatch intervals during which the price of energy can increase significantly due to scarcity pricing, even if the event does not present a significant reliability risk http://tinyurl.com/FERC-ER14-2156-000 4 / 15
Virtual Energy Storage Frequency Decomposition Taming the Duck One Day at CAISO 2020 r a m p 25 Net Load Curve e d 20 s e a r 15 c I n GW 10 ISO/RTOs are seeking ramping products 5 to address engineering challenges, and to avoid scarcity prices 0 Do we need ramping products? -5 12am 3am 6am 9am 12pm 3pm 6pm 9pm 12am 5 / 15
Virtual Energy Storage Frequency Decomposition Taming the Duck One Day at CAISO 2020 25 Net Load Curve 20 k c u D e h 15 t g n GW i m a T 10 5 This doesn’t look at all scary! We need resources, but anyone here 0 knows how to track this tame duck -5 12am 3am 6am 9am 12pm 3pm 6pm 9pm 12am 5 / 15
Virtual Energy Storage Frequency Decomposition Taming the Duck One Day at CAISO 2020 25 Net Load Curve Low pass 20 15 GW 10 The duck is a sum of a smooth energy signal, and two zero-energy services 5 Mid pass 0 High pass -5 12am 3am 6am 9am 12pm 3pm 6pm 9pm 12am 5 / 15
Virtual Energy Storage Frequency Decomposition Regulation 4 G W ( t ) = Wind generation in BPA, Jan 2015 3 GW 2 Ramps 1 0 Jan 01 Jan 02 Jan 03 Jan 04 Jan 05 Jan 06 6 / 15
Virtual Energy Storage Frequency Decomposition Regulation 4 W ( t ) = Wind generation in BPA, Jan 2015 G Goal: G W ( t ) + G r ( t ) ≡ 4 GW 3 GW 2 Ramps 1 0 Jan 01 Jan 02 Jan 03 Jan 04 Jan 05 Jan 06 6 / 15
Virtual Energy Storage Frequency Decomposition Regulation 4 W ( t ) = Wind generation in BPA, Jan 2015 G Ramps Goal: G W ( t ) + G r ( t ) ≡ 4 GW 3 Ramps Ra GW 2 Ramps Ramps Ramps 1 0 Jan 01 Jan 02 Jan 03 Jan 04 Jan 05 Jan 06 6 / 15
Virtual Energy Storage Frequency Decomposition Regulation 4 G r ( t ) 3 Ramp GW 2 1 Goal: G W ( t ) + G r ( t ) ≡ 4 GW G r ( t ) obtained from generation? 0 Jan 01 Jan 02 Jan 03 Jan 04 Jan 05 Jan 06 6 / 15
Virtual Energy Storage Frequency Decomposition Regulation 4 G r ( t ) 3 GW 2 G 1 G r = G 1 + G 2 + G 3 1 0 Jan 01 Jan 02 Jan 03 Jan 04 Jan 05 Jan 06 6 / 15
Virtual Energy Storage Frequency Decomposition Regulation 4 G r ( t ) 3 GW 2 G 1 G 2 G r = G 1 + G 2 + G 3 1 0 Jan 01 Jan 02 Jan 03 Jan 04 Jan 05 Jan 06 6 / 15
Virtual Energy Storage Frequency Decomposition Regulation 4 G r ( t ) 3 GW 2 G 1 G 2 G r = G 1 + G 2 + G 3 G 3 1 0 Jan 01 Jan 02 Jan 03 Jan 04 Jan 05 Jan 06 6 / 15
Virtual Energy Storage Frequency Decomposition Regulation 4 G r ( t ) 3 GW 2 G 1 G 2 G r = G 1 + G 2 + G 3 Where do we find G 3 1 these resources? 0 Jan 01 Jan 02 Jan 03 Jan 04 Jan 05 Jan 06 6 / 15
Power deviation Local decision Grid signal U i Y i ζ t Local t t Load i Control X i t Local feedback loop Demand Dispatch Design
Demand Dispatch Demand Dispatch G r = G 1 + G 2 + G 3 G 1 G 3 ? G 2 G r 7 / 15
Demand Dispatch Demand Dispatch G r = G 1 + G 2 + G 3 G 1 Traditional generation G 2 G 3 G r 7 / 15
Demand Dispatch Demand Dispatch G r = G 1 + G 2 + G 3 G 1 Traditional generation G 2 Water pumping (e.g. pool pumps) Fans in commercial HVAC G 3 G r Demand Dispatch: Power consumption from loads varies automatically and continuously to provide service to the grid, without impacting QoS to the consumer 7 / 15
Demand Dispatch Demand Dispatch Responsive Regulation and desired QoS – A partial list of the needs of the grid operator, and the consumer High quality AS? (Ancillary Service) Does the deviation in power consumption accurately track the desired deviation target? 8 / 15
Demand Dispatch Demand Dispatch Responsive Regulation and desired QoS – A partial list of the needs of the grid operator, and the consumer High quality AS? (Ancillary Service) Reliable? Will AS be available each day? It may vary with time, but capacity must be predictable. 8 / 15
Demand Dispatch Demand Dispatch Responsive Regulation and desired QoS – A partial list of the needs of the grid operator, and the consumer High quality AS? Reliable? Cost effective? This includes installation cost, communication cost, maintenance, and environmental. 8 / 15
Demand Dispatch Demand Dispatch Responsive Regulation and desired QoS – A partial list of the needs of the grid operator, and the consumer High quality AS? Reliable? Cost effective? Customer QoS constraints satisfied? The pool must be clean, fresh fish stays cold, building climate is subject to strict bounds, farm irrigation is subject to strict constraints, data centers require sufficient power to perform their tasks. 8 / 15
Demand Dispatch Demand Dispatch Responsive Regulation and desired QoS – A partial list of the needs of the grid operator, and the consumer High quality AS? Reliable? Cost effective? Customer QoS constraints satisfied? Virtual energy storage: achieve these goals simultaneously through distributed control 8 / 15
Demand Dispatch General Principles for Design Local feedback loop Two components to local control U i Y i U i ζ t ζ t Local t t t Load i Prefilter Decision Control X i t X i t Each load monitors its state and a regulation signal from the grid. Prefilter and decision rules designed to respect needs of load and grid Randomized policies required for finite-state loads 9 / 15
Demand Dispatch MDP model MDP model The state for a load is modeled as a controlled Markov chain. Controlled transition matrix: P ζ ( x, x ′ ) = P { X t +1 = x ′ | X t = x, ζ t = ζ } Local feedback loop Two components to local control U i Y i U i ζ t ζ t Local t t t Load i Prefilter Decision Control X i t X i t 10 / 15
Demand Dispatch MDP model MDP model The state for a load is modeled as a controlled Markov chain. Controlled transition matrix: P ζ ( x, x ′ ) = P { X t +1 = x ′ | X t = x, ζ t = ζ } Local feedback loop Two components to local control U i Y i U i ζ t ζ t Local t t t Load i Prefilter Decision Control X i t X i t Questions: • How to analyze aggregate of similar loads? • How to design P ζ ? 10 / 15
Recommend
More recommend
