Transactive Control: A Novel Technology for Smart Grids Anuradha Annaswamy Active-adaptive Control Laboratory Department of Mechanical Engineering Massachusetts Institute of Technology 1 Depart of Mechanical Engineering Seminar Series, WPI, October 9, 2013 Outline • A Smart Grid – A Paradigm Shift • Transactive Control – Dynamic Market Mechanisms – Integrated Secondary and Primary Control • Case Studies 2 Anuradha Annaswamy, Transactive Control Paradigm Shift: From Current to Smart Grids Increasing supply-demand gap Network Operations Environmental concerns Transmission Utilities Aging infrastructures Industrial Consumers Distribution Demand Response Smart Meters Distribution Consumers Smart Devices Industrial Consumers Utilities Generation Facilities Conventional Grid Distribution Consumers Network Operations Main features: Energy providers • Renewable energy DERs resources Microgrid • Demand Response • Storage Transmission • Advanced Metering 3 Infrastructure Anuradha Annaswamy, Transactive Control Distributed Community Storage Distribution Smart Grid Control • To maintain power balance in the system. • To ensure that operating limits are maintained – Generators limit – Tie-lines limit • To ensure that the system frequency is constant (at 50 Hz or 60Hz). • To achieve the above with renewable energy despite intermittency & uncertainty • To ensure affordable power 4 Anuradha Annaswamy, Transactive Control GRID CONTROL: CURRENT PRACTICE Anuradha Annaswamy, Transactive Control TRANSACTIVE CONTROL: The connecting entity Anuradha Annaswamy, Transactive Control THE OVERALL VISION Vision for Smart Grid Control: 2030 and Beyond: Reference Model and Roadmap (Eds. M. Amin, A.M. Annaswamy, C. DeMarco, and T. Samad), IEEE Standards Publication, November 2013. Anuradha Annaswamy, Transactive Control Distributed Decision and Control • Primary control – Immediate (automatic) action to sudden change of load. – For example, reaction to frequency change. • Secondary control – Restore system frequency, – Restore tie-line capacities to the scheduled value, and, – Make the areas absorb their own load. • Tertiary control – Make sure that the units are scheduled in the most economical way. 8 Anuradha Annaswamy, Transactive Control Transactive control: An Emerging Paradigm* The use of dynamic market mechanism to send an incentive signal and receive a feedback signal within the power system’s node structure • Incentive Signal: Dynamic Pricing • Feedback Signal: Adjustable Demand * Hammerstorm et al., “Standardization of a Hierarchical Transactive Control System” 9 Anuradha Annaswamy, Transactive Control Transactive Control: Example • Pacific Northwest Demonstration Project • 112 Households participating in 2009 • 60,000 households in an ongoing project (2010-2015) • Spans several states Courtesy of Olympic Peninsula Project, IBM TIS: Transactive Incentive Signal TFS: Transactive Feedback Signal 10 Anuradha Annaswamy, Transactive Control Transactive control: Our Definition The use of dynamic market mechanism to send an incentive signal and receive a feedback signal within the power system’s node structure • Incentive Signal: Ex. Dynamic Pricing • Feedback Signal: • Adjustable Demand (Market Level) • (Price Responsive, and Regulation Responsive) • Area Control Error (Secondary Level) • Governor Control (Primary Level) Transactive Control Control architecture that coordinates Market Transactions Active Control at the AGC level with Regulation Demand Response 11 Anuradha Annaswamy, Transactive Control Transactive Control Framework* Incentive Signal Feedback Signal Market Transactions Demand Demand Generation Generation Demand … Generation ~5 mins TRANSACTIVE Area Control Error (ACE) ~10secs ~sec CONTROL AREA-LEVEL Set-points SECONDARY (FREQUENCY) CONTROL UNIT- LEVEL PRIMARY (POWER) CONTROL Set-points * A. Kiani, A.M. Annaswamy, and T. Samad, “A Hierarchical Transactive Control Architecture for Renewables Integration in Smart Grids.” HYCON Workshop, Brussels, 2012. 12 Anuradha Annaswamy, Transactive Control Primary Level - time scale t é PG ù ú zp = ê êë PL úû Steady state é w [k] ù ref ú u[k] = ê ref ê PL [k] ú ë û f : Tie - Line flow D : uncertainty 0 = Ax pss [k] + Bz pss [k] + Fu[k] + D pss 0 = Cx pss [k] + Dz pss [k] + f pss [k] + Df pss xGss Primary Secondary k K Tertiary t k+1 K+1 13 Anuradha Annaswamy, Transactive Control Secondary Level 0 = Ax pss [k] + Bz pss [k] + Fu[k] + D pss 0 = Cx pss [k] + Dz pss [k] + f pss [k] + Df pss ß xs [k +1] = xs [k] + Bsus [k]+ Cs D s [k] xs [k] : wGss us [k]: u[k +1]- u[k] D s [k] : Uncertainty in generation, load, and tie-line flow Goal: xs ® xt xt es a reference signal set by the tertiary level Controller Secondary Level Dynamics es = xs - xt : Area Control Error (ACE) 14 Anuradha Annaswamy, Transactive Control xs Tertiary Level Controller Tertiary xt [K] Gain es [k] Gain Secondary level dynamics Primary Level Dynamics How do we design the Tertiary Level? 15 Anuradha Annaswamy, Transactive Control xs [k] Electricity Market • Centralized mechanism that facilitates trading of energy between buyers and sellers. • The market operator conducts an auction market and schedules generators based on bids received. • Determines a market clearing price (Locational Marginal Price (LMP)) and provides commitments and schedules based on security-constrained unit commitments • Day-ahead (DA) Markets • Real-time Markets (RTM) Wholesale Market 16 Anuradha Annaswamy, Transactive Control Wholesale Market: A Dynamic System Operating Day 22:00 18:00 Operating Day-1 Real Time Energy Market clears 00:10 Day ahead reliability unit commitment 00:00 16:00 Clear Day-ahead Market using Unit commitment and Economic Dispatch Real Time Energy Market opens DA Energy Market results published 16:00 12:00 DA Energy Market offer and bid period closes Revising bids 17 ISO finalizes operating plan for the next day Power Delivery Time Market Mechanisms - LMP Node 1 Node 2 Node n Demand Demand Demand … Generation Generation Generation LMPi, Schedules Bids (MW-h, $) ISO Nodes in New England, USA 18 Anuradha Annaswamy, Transactive Control Top Layer: A Dynamic Market Mechanism Suggested Price Suggested Price Congestion Rent Generation (with Renewable Energy Resources) Suggested MW Consumption Consumers Substations Suggested MW Generation Phase Angle Real-time Market 1. Equilibrium under constant flux. 2. GenCos and ConCos adjust their power level using a recursive process. 19 3. Price is a Public Signal that guides all entities to adjust efficiently. Anuradha Annaswamy, Transactive Control Modeling of Generating Company • 𝜌 : market price ~ Locational Marginal Price at market equilibrium. • The cost function of each generators unit is That is, if a generator observes a market price above the marginal cost , will expand production until the marginal cost of production equals the price. 20 Anuradha Annaswamy, Transactive Control Modeling of Consumers Company • : marginal benefit of Pd • Consumer utility function: j i.e. Demand with a marginal benefit above the marginal price will lead to an expansion in consumption until equilibrium is attained. 21 Anuradha Annaswamy, Transactive Control Pricing Strategy • Energy imbalance Ek at time k • The pricing policy should depend on the degree of energy imbalance 22 Anuradha Annaswamy, Transactive Control A Dynamic Market Model • The market participants need not have global market information. • Convergence of the dynamic system to the equilibrium condition implies that the market reaches the condition of Nash equilibrium. Distributed Gaming 23 Anuradha Annaswamy, Transactive Control Dynamic Market Mechanism (contd.) xt[K] Xt xt • Quantifies effect of volatility and stability • Can help reduce reserve costs with wind uncertainty 24 Anuradha Annaswamy, Transactive Control Interconnections Controller Tertiary xt [K] Gain es [k] Gain Secondary level dynamics Primary Level Dynamics 25 Anuradha Annaswamy, Transactive Control xs [k] Transactive Control: Lower Levels The overall model, including the primary, secondary, and tertiary level dynamics at multiple time-scales 26 Anuradha Annaswamy, Transactive Control Transactive Control: Stability* * A. Kiani and A.M. Annaswamy, “A Hierarchical Transactive Control Architecture for Renewables Integration in Smart Grids,“ CDC 2012, Maui, Hawaii. 27 Anuradha Annaswamy, Transactive Control Transactive control architecture The use of dynamic market mechanism to send an incentive signal and receive a feedback signal within the power system’s node structure • Incentive Signal: Ex. Dynamic Pricing • Feedback Signal: • Adjustable Demand (Market Level) • (Price Responsive, and Regulation Responsive) • Area Control Error (Secondary Level) • Governor Control (Primary Level) Transactive Control Control architecture that coordinates Market Transactions Active Control at the AGC level with Regulation Demand Response 28 Anuradha Annaswamy, Transactive Control Simulation Results • 4-bus network with two generator units at node 1 and wind at bus 2 (Pg1: Base-load; Pg2: Reserve) • L1, L2: DR-Compatible demand Parameters with following values: Pg2 Pg1 G G Area 2 L1 cg1 = 0.25; cg2 = 0.55; generator cost coefficients bg1 = 40.2; bg2 = 60; generator cost coefficients kg1 = 0.3; kg2 = 0.8; generator time constants cd1 = cd2 = 0.4; consumer utility coefficients bd1 = bd2 = 70; consumer cost coefficients kd1 = kd2 = 0.3; demand time constants k = 0:7; LMP time constant (market time constant) W Area 1 L2 29 Anuradha Annaswamy, Transactive Control Market Stability & Volatility Volatility: With increased demand-elasticity (kd) Stability: With increased latency (kr) 30 Anuradha Annaswamy, Transactive Control Simulation Results: Market Stability & Volatility Volatility: With increased demand-elasticity (kd) Stability: With increased latency (kr) 31 Anuradha Annaswamy, Transactive Control Simulation Results Wind Properties: : Actual Wind Power : Mean value of the projected wind. Current Market Practice : ARMA model of the actual wind power. With Transactive 5.3 Control Stability Analysis of Transactiv G G 1 2 Area 2 Area 1 4 3 L2 PD 2 PD 1 L1 Figure 5.3: 4-bus system example for Transactive control 32 Anuradha ItAnnaswamy, Transactive Control follows that the equilibrium of (5.59)-(5.60) is asymptotically stable if an Simulation Results: Effect of Wind Uncertainty Less reserve is required. Hierarchical coordination Anuradha Annaswamy, Transactive Control Simulation Results: IEEE 30 bus Case Reserve Anuradha Annaswamy, Transactive Control Summary • Transactive Control – Dynamic Market Mechanisms – Integrated Secondary and Primary Control • Case Studies 35 Anuradha Annaswamy, Transactive Control A 2013 Publication! 36 Anuradha Annaswamy, Transactive Control Thank You! 37 Anuradha Annaswamy, Transactive Control