Poster Session and Reception

Tuesday, May 10, 2016 - 4:00pm - 6:00pm
Lind 400
  • Decentralized Optimal Control of Inter-area Oscillations in Power Systems
    Xiaofan Wu (University of Minnesota, Twin Cities)
    To improve upon the limitations of conventional wide-area control strategies, we study the problem of signal selection and optimal design of sparse and block-sparse wide-area controllers. In our design, we preserve rotational symmetry of the power system by allowing only relative angle measurements in the distributed controllers. For the New York-New England test system, we examine performance tradeoffs of different control architectures and show that optimal retuning of fully-decentralized control strategies can effectively guard against local and inter-area oscillations.
  • Virtual Battery Capacity of Deferrable Energy Demand
    Daria Madjidian (Massachusetts Institute of Technology)
    Integration of renewable resources at large scale poses significant challenges for control and operation of power systems. While conventional energy storage, such as batteries and fly-wheels, would be instrumental in solving these problems, current technologies prohibitively expensive for large-scale deployment. At the same time, it is widely acknowledged that a substantial portion of the electricity demand is flexible and could be used to absorb supply side fluctuations. In this research, we investigate the ability of a collection of flexible energy consumers to act as a virtual battery, that is, a resource that can store and release energy in a controllable and reliable fashion up to predetermined limits on volume and charge/discharge rates. Our contribution is twofold. First, we provide upper and lower bounds on the battery capacity that can be attained, and show that there is a fundamental trade-off between volume, charge rate and discharge rate. Second, we introduce a class of priority based scheduling policies that allow the collection of loads to emulate the batteries on the lower bound.
  • Qatar Water Distribution Systems: Transition to Smarter Networks
    Mohsen Aghashahi (Texas A & M University)
    Qatar drinking water supply and distribution infrastructure has experienced a remarkable expansion during the past decade. Transition of this critical infrastructure sector to smarter systems is necessary to ensure reliable, energy-efficient network. This research advances existing mobile and stationary water sensor networks technology to enable realtime sensing and communication. Optimal pervasive control methods are developed to facilitate system automation and fault prevention informed by sensor readings and simulation projections. The outputs of this project will be a versatile, smart cyberphysical water system and a pilot pipeline facility that can be used for broad purposes and applications.
  • Big Data Analytics for Demand Response and Distribution Intelligence
    Amin Rasekh (Sensus)
    A smart grid is an integrated set of products, solutions and systems that enable utilities to remotely and continuously monitor and diagnose problems, prioritize and manage maintenance issues and use data to optimize all aspects of the network. With Advanced Metering Infrastructure, SCADA and other systems, electric utilities are receiving thousands of times the data compared to just a few years ago. Big Data analytics efficiently stores and validates information the utilities collect from smart meters, SCADA systems, customer billing software and news services. It also enables demand Response applications that help utilities effectively monitor usage and control peak demand.
  • New Hybrid Optimization Algorithm to Improve Reliability and Reduce Losses in Micro-Grid
    Miguel Velez-Reyes (University of Texas at El Paso)
    This poster presents an efficient planning of photovoltaic (PV) arrays coordinated with a battery energy storage system (BESS) in a micro-grid (MG) by using a new hybrid optimization technique based on improved harmony search (IHS) and memetic algorithm (MA). At first, optimal location for PV panels is determined by solving a multi-objective function with an overall goal of reliability improvement, and reduction of losses and active power demand purchased from upstream grid. Then, it is assumed that BESS will be connected to the same buses where the PV arrays are installed. BESS are used to minimize power mismatch between generation and consumption, and voltage regulation. The capacity of each BESS installed in certain buses has been determined based on active power demand and forecasted PV output power on each (certain) bus. To prove the effectiveness of the proposed hybrid HIS+MA algorithm, results are compared with those obtained from genetic algorithm (GA). Simulations are carried out for three different cases, i.e., sunny day, cloudy day, and rainy day. IEEE 34-bus standard test system is used in the case study.
  • Connecting Distributed Optimization and Distributed Control in Power Grids
    Xuan Zhang (Harvard University)
    Due to increasing uncertainties resulting from renewable energy penetration and variability in both supply and demand, the control and economic optimization for power networks will need to run on faster time-scales. Moreover, distributed and decentralized control architectures are necessary as power systems are distributed large-scale networks with a lot of complexity, which makes centralized control expensive to implement. This poster proposes a reverse- and forward-engineering framework for designing real-time distributed control and economic optimization algorithms in power grids, which consists of two steps. Firstly, we reverse-engineer given system dynamics as an optimization algorithm to solve a certain optimization problem. Secondly, we use proper forward-engineering approaches to systematically design distributed closed-loop control. As a result, the system automatically tracks the optimal solution of a predefined optimization problem and the control scheme can be implemented in a distributed and closed-loop manner. To illustrate the effectiveness, we apply this framework to design primary load frequency control, economic Automatic Generation Control, as well as economically optimal frequency control in power systems. Numerical investigations demonstrate good performances of the proposed control schemes.
  • Swing Contract Market Design for Flexible Service Provision in Electric Power Systems
    Leigh Tesfatsion (Iowa State University)
    The need for flexible service provision in power systems has dramatically increased due to the increased penetration of variable energy resources, as has the need to ensure fair access and compensation for this provision. A swing contract (SC) permits flexible service provision because services can be offered as ranges of values rather than as point values. This study develops a new SC Market Design and optimal market-clearing formulation for electric power systems that permits SCs to be offered by any dispatchable resource. Three distinctive characteristics of the SC Market Design are as follows:

    1) Permits separate market-based compensation for service availability via an SC’s availability (offer) price and for actual real-time service performance via the performance payment method included among the contractual terms of the SC;

    2) Permits bundling of multiple services ( e.g., power, ramp rate, duration, ...) into a single SC;

    3) Each service in an SC can be offered with flexibility (swing) in its implementation range.
  • On the Laplacian of a Signed Graph with Application to Microgrids​
    Li Qiu (Hong Kong University of Science and Technology)
    We investigate the positive semidefiniteness of the Laplacian of a signed graph with negative weights. It is noted that an undirected signed graph defines a unique resistive electrical network, wherein the negative weights correspond to negative resistances. As such, the positive semideniteness of the Laplacians is equivalent to the passivity of the associated resistive networks. By utilizing the n-port circuit theory, we obtain several equivalent conditions for the Laplacian of signed graphs to be positive semidefinite and have a simple zero eigenvalue. The result is used to analyze the small-disturbance angle stability of microgrids.
  • Scalable H-infinity Control with Applications in Heat Transfer
    Carolina Lidström (Lund University)
    We address H-infinity state feedback and give a simple form for an optimal control law applicable to linear time-invariant systems with symmetric and Hurwitz state matrix. More specifically, the control law can be expressed in the matrices of the system's state space representation, given separate cost on state and control input. A corresponding result is given for infinite-dimensional systems. The derived optimal control law is transparent, easy to synthesize and scalable. It is applied to control of the heat equation and control of the temperature in buildings.
  • Smart Power Systems of the Future: Foundations for Understanding and Improving Operational Reliability
    Neil Cammardella (University of Florida)
    The power grid in the U.S. and many regions of the world is undergoing changes because of new technologies and government mandates. It is believed that smart meters and a smarter grid will lead to more efficient use of our infrastructure. In addition, increased renewable energy integration will provide power at low cost. This optimism may be justified, but only if experts in control theory play a leading role.
  • Quantifying Market Incentives for Concentrated Solar Power Generators
    Alexander Dowling (University of Wisconsin, Madison)
    Most techno-economic studies of concentrated solar power (CSP) systems only consider levelized cost of electricity (LCOE) to quantitatively compare economics and performance. This is problematic as LCOE neglects the time varying value of electricity and additional revenue opportunities from providing ancillary services. As consequence, LCOE-centric analyses undervalue the flexibility provided by thermal energy storage and misrepresent the true economics of owning and operating CSP systems. For example, a CSP plant providing 10 MW of regulation capacity for all hours of 2015 in the California energy market would have received $500,000 in capacity payments alone. Similarly, shifting 10 MWe of generation from the average real-time market price (30 $/MWh) to the 1% most extreme prices (97 to 1,621 $/MWh) yields additional revenues of $400,000/yr. Thus selection of the appropriate economic and performance metrics is paramount when evaluating CSP systems and comparing against other energy technologies.

    We present a unified framework that considers market rules, start-up/shutdown restrictions and detailed system physics to analyze multiple CSP performance metrics, especially profits and water usage. The electricity market mathematical model is sufficiently detailed to evaluate revenues from selling/providing both energy and ancillary services, including regulation and (non)-spinning reserve capacity, in the California day-ahead and real-time markets. The CSP physics models include mass and energy balances and nonlinear equipment performance correlations. They are sufficiently complex to resolve temperature profiles in the thermal storage tanks and study the impact of storage dynamics on system performance. Revenue estimation and operational policy determination are formulated as mixed integer nonlinear programs (optimization problems). We present a decomposition strategy that uses physics inspired relaxations to efficiently find approximate solutions that are feasible for all of constraints and locally optimal with respect to a fixed start-up/shut-down schedule. The framework is implemented in Julia/JuMP (an open-source programming language/modeling environment) and leverages large-scale and efficient numerical optimization solvers such as IPOPT and Gurobi.
  • Packetized Energy Management: Asynchronous Coordination of Thermostatically Controlled Loads
    Mads Almassalkhi (University of Vermont)
    Because of their internal energy storage, electri- cally powered, thermostatically controlled loads (TCLs) have the potential to be dynamically managed to match their cumulative load to the available supply. However, in order to facilitate consumer acceptance of this type of load management, TCLs need to be managed in a way that avoids degrading perceived quality of service (QoS), autonomy, and privacy. This paper presents a real-time, adaptable approach to managing TCLs that both meets the requirements of the grid and does not require explicit knowledge of specific TCL’s state. The method leverages a packetized, probabilistic approach to energy delivery that draws inspiration from digital communications. We test the method using a set of 300 simulated water heaters and find that the method can closely track a time-varying signal without noticably degrading quality of service.
  • Duration-deadline Jointly Differentiated Energy Services
    Wei Chen (Royal Institute of Technology (KTH))
    The supply/demand balance becomes more challenging in a smart grid due to the more uncertainty and intermittency brought in by the increasing integration of renewables. It comes to researchers’ attention that the conventional scheme of supply following demand is neither economically efficient nor environmental beneficial. To resolve this issue, an alternative paradigm is to exploit the flexibility from the demand side to compensate the uncertainty in the supply side. Following this direction, we study the so-called duration-deadline jointly differentiated energy services. Specifically, the energy services are differentiated by both the duration and deadline requirements. We study the adequacy problem of a given supply profile which amounts to solving a constrained binary matrix completion problem. We also propose a market implementation of such energy services and show the existence of an efficient competitive equilibrium. Several extensions have been discussed as well. Of particular interest is to examine the case when peer-to-peer charging is allowed among the consumers.
  • Optimization of Smart Solar Inverters in a Transactive Control Setting
    R. Blake Rector (Portland State University)
    e develop an objective function for determining an optimal control scheme for a smart solar inverter with onsite battery storage in a transactive control setting. The control scheme seeks to maximize the revenue earned by the inverter for providing transactive services to the grid including real power sales, reactive power support, and spinning reserve capacity. The objective function uses predicted market prices and solar irradiance as inputs. The method is tested using historical market data from the Midcontinent Independent System Operator (MISO), and the results are compared to a business-as-usual control scheme. The method is flexible, fast, and can also be applied under other rate or compensatory structures.
  • Real-Time Price Optimization for Load Frequency Control in Electric Power Systems with Renewable Energy
    Toshiyuki Ohtsuka (Kyoto University)
    We formulate real-time pricing in electric power systems with renewable energy as a nonlinear model predictive control (NMPC) problem and demonstrate that a small change in its performance index yields drastic changes in distributions of renewable energy and electricity prices over a power network. We consider electric power systems consisting of consumers, suppliers, wind farms, generators, and an independent system operator (ISO) and assume that electricity generated by wind farms can be predicted over a certain time interval in the future. Consumers and suppliers are modeled by demand curves and supply curves, respectively, derived from their utility functions and cost functions. The ISO can manipulate electricity prices in each area at each time point to stabilize load frequencies of generators, to maximize total benefits of consumers and suppliers, and to balance demands and supplies. To achieve these objectives, the current electricity prices are determined by real-time optimization to minimize a performance index over a finite future, which is formulated as NMPC. We examine two types of performance indices, and simulation results show that a small difference in the performance indices results in contrasting responses in distributions of renewable energy and electricity prices over the power network. Simulation results also show that numerical optimization for NMPC is executed sufficiently fast for real-time implementation.
  • Robust Security-Constrained Unit Commitment with Dynamic Ratings
    Wei Wang (University of Pittsburgh)
    Unit commitment (UC) is one of the most challenging problems in power market and the uncertainties arose in supply and demand sides make it even harder. Recently, robust optimization (RO) gained a lot attention for solving UC problem since it can guarantee the demand will be satisfied under worst case. However, there are some practical problems, such as gas generator efficiency related to inlet air temperature, in which the uncertainty appears in the body of constraints. To deal with this, a two-stage robust optimization model for UC problem taking both the uncertainties in the inlet air temperature and the forecasted demand into consideration is formulated. Under certain assumption, the model is solved by tailored column and constraint generation algorithm. Numerical experiments are also demonstrated and compared with other solution strategies.
  • Energy Storage Dispatch Using Adaptive Control Scheme Considering Wind-PV in Smart Distribution Network
    Miguel Velez-Reyes (University of Texas at El Paso)
    This paper presents an adaptive control scheme for optimal dispatch of energy storage systems (ESS) to follow the pattern of intermittent power output of renewable energy sources (RES) in electric power distribution networks with the goal to minimize costs and reduce the need to compensate the variability and uncertainty of RES, mainly wind and solar. The proposed control scheme utilizes day-ahead forecasts of wind and photovoltaic (PV) power output obtained from hybrid intelligent methods. Once the forecasts are obtained, unit commitment (UC) is executed using forecasted data of load as well as wind and PV power in order to schedule optimal generation. The operational decisions are then fed into the economic dispatch (ED) problem, which has the control scheme embedded. As the actual power output of wind-PV deviates from the desired value due to forecast error, the adaptive control scheme developed in this paper assists the ESS to compensate the difference by charging or discharging. In order to evaluate the effectiveness of the proposed model, this paper considers a case study of a sixteen-bus test system and the results indicate that the ESS can reduce the deviation of the wind-PV power output between 1-5%.
  • Graphical Models for Power Systems Analysis
    Krishnamurthy Dvijotham (California Institute of Technology)
    Several problems arising in the design, analysis and efficient operation of power systems are naturally posed as graph-structured optimization problems. Due to the nonlinear nature of the physical equations describing the power grid, these problems are often nonconvex and NP-hard. However, practical instances of several graph-structured optimization problems have been solved successfully in the graphical models literature by exploiting graph structure and using message-passing or belief propagation techniques. In this work, we show that a similar approach can be successfully applied to power systems, leading to theoretically and practically efficient algorithms. I will discuss two applications in detail: a) Solving mixed-integer optimal power flow problems on distribution networks, and b) Detecting and mitigating market manipulation by aggregators of renewable generation in a distribution-level market. I will also discuss possible extensions of these approaches to other power system/infrastructure network problems.

    Based on joint work with Misha Chertkov, Sidhant Misra, Marc Vuffray, Pascal Van Hentenryck, Niangjun Chen, Navid Azizan Ruhi and Adam Wierman.
  • Polynomial Optimisation in Power Systems
    Jakub Marecek (IBM Research Division)
    Assuming the alternating-current models of power flows, a wide variety of optimisation problems over steady states of power systems can be cast as polynomial optimization problems. This makes it possible to solve large instances in practice and to guarantee asymptotic convergence in theory. In particular:
    -- First (, IEEE T. Power Systems 31:1), we formulate the alternating-current optimal power flow (ACOPF) problem as a degree-two polynomial program and study two approaches to solving it via convexifications. In the first approach, we tighten the first-order relaxation of the nonconvex quadratic program by adding valid inequalities. In the second approach, we exploit the structure of the polynomial program by using a sparse variant of Lasserre's hierarchy. This allows us to solve instances of up to 39 buses to global optimality and to provide strong bounds for the Polish network within an hour.
    -- Second (, submitted), we present a re-formulation of the ACOPF with coordinate-wise constraints, which makes it possible to derive novel semidefinite programming (SDP) relaxations. For those, we develop a first-order method based on coordinate descent, which is often as fast as Matpower's interior point method, the popular heuristic, within the same accuracy. Contrast this with readily available second-order methods for solving the SDP relaxations for real-world large-scale power systems, which often fail to perform even a single iteration within reasonable run-times.
    -- Finally (, submitted), we study means of switching from first-order methods for solving the convex relaxation to Newton method on the original non-convex problem. This allows for convergence under the same conditions as in solvers for the convex relaxation, but with a rate of convergence better than those possible with first-order methods. Again, we illustrate our approach on the ACOPF.
  • Distributed Optimization Decomposition for Joint Economic Dispatch and Frequency Regulation
    Enrique Mallada (Johns Hopkins University)Adam Wierman (California Institute of Technology)
    Economic dispatch and frequency regulation are typically viewed as fundamentally different problems in power systems and, hence, are typically studied separately. In this work, we frame and study a joint problem that co-optimizes both slow timescale economic dispatch resources and fast timescale frequency regulation resources. We show how the joint problem can be decomposed without loss of optimality into slow and fast timescale sub-problems that have appealing interpretations as the economic dispatch and frequency regulation problems respectively. We solve the fast timescale sub-problem using a distributed frequency control algorithm that preserves the stability of the network during transients. We solve the slow timescale sub-problem using an efficient market mechanism that coordinates with the fast timescale sub-problem. We investigate the performance of the decomposition on the IEEE 24-bus reliability test system.