Search

This paper compares risk-averse optimization methods to address the self-scheduling and market involvement of a virtual power plant (VPP). The decision-making problem of the VPP involves uncertainty in the wind speed and electricity price forecast. We focus on two methods: risk-averse two-stage stochastic programming (SP) and two-stage adaptive robust optimization (ARO). We investigate both methods concerning formulations, uncertainty and risk, decomposition algorithms, and their computational performance. To quantify the risk in SP, we use the conditional value at risk (CVaR) because it can resemble a worst-case measure, which naturally links to ARO. We use two efficient implementations of the decomposition algorithms for SP and ARO; we assess (1) the operational results regarding first-stage decision variables, estimate of expected profit, and estimate of the CVaR of the profit and (2) their performance taking into consideration different sample sizes and risk management parameters. The results show that similar first-stage solutions are obtained depending on the risk parameterizations used in each formulation. Computationally, we identified three cases: (1) SP with a sample of 500 elements is competitive with ARO; (2) SP performance degrades comparing to the first case and ARO fails to converge in four out of five risk parameters; (3) SP fails to converge, whereas ARO converges in three out of five risk parameters. Overall, these performance cases depend on the combined effect of deterministic and uncertain data and risk parameters. Summary of Contribution: The work presented in this manuscript is at the intersection of operations research and computer science, which are intrinsically related with the scope and mission of IJOC. From the operations research perspective, two methodologies for optimization under uncertainty are studied: risk-averse stochastic programming and adaptive robust optimization. These methodologies are illustrated using an energy scheduling problem. The study includes a comparison from the point of view of uncertainty modeling, formulations, decomposition methods, and analysis of solutions. From the computer science perspective, a careful implementation of decomposition methods using parallelization techniques and a sample average approximation methodology was done. A detailed comparison of the computational performance of both methods is performed. Finally, the conclusions allow establishing links between two alternative methodologies in operations research: stochastic programming and robust optimization. [ABSTRACT FROM AUTHOR]

We consider the operation of a power system with high renewable penetration and minimal network bottlenecks and provide a technique to determine the appropriate level of reserve to enforce. Within a daily scheduling framework, we model the production uncertainty of renewable units via scenarios and include security constraints per scenario to ensure an uneventful operation in case of failure of key thermal units. The rarely constraining network is modeled via areas with limited capacity at their interfaces. The considered decision-making framework results in a two-stage stochastic programming problem whose first stage decides on the commitment of thermal units and whose second stage reproduces, per scenario, the operation of the system with and without the failure of key thermal units. The optimal unit scheduling obtained from the proposed model implicitly identifies a generating reserve pattern that we compare with common rules used in industry for reserve identification, such as 20% or 30% of the demand. We use the Illinois 200-bus system to illustrate and computationally characterize the methodology proposed. • An economically efficient yet secure reserve determination procedure. • Renewable uncertainty is modeled via scenarios. • Security regarding unit failures is ensured via security constraints. • The resulting optimization model is tractable and does not need decomposition. • The reserve determined is superior to setting a percentage of the load as reserve. [ABSTRACT FROM AUTHOR]

One important task in power distribution system expansion planning is feeder routing (FR), which is to determine the optimal route from a medium voltage substation to load points and the optimal size of the conductors to be installed. This paper proposes a novel graph-based model for the resilient feeder routing problem using geographical information system (GIS). We show that incorporating GIS data enhances the FR solution optimality. Moreover, by introducing a representing graph, radiality of the planned network is guaranteed. A second-order cone programming (SOCP) model is used to model power flows through feeders. Since the representing graph ensures radiality, the SOCP model is exact. The uncertainty of rooftop solar generations and demand forecasting errors are taken into account, and a stochastic programming-based solution algorithm is developed. The proposed model represents practical aspects such as economic objectives (installation cost, power losses, resiliency), technical constraints (voltage limits, radiality constraint, reliability), and geographical constraints. The efficiency of the algorithm is demonstrated using three case studies: a small test system, a realistic one, and a synthetic large test system. [ABSTRACT FROM AUTHOR]

We study market equilibria that are achieved by strategic firms that participate in electricity and natural gas markets. Strategic firms submit their offers and bids to both markets with the aim of maximizing profit or utility and we consider firms that can include a combination of electricity and natural gas supply and demand. The strategic actions of these firms are represented by upper-level problems that are optimized subject to shared lower-level problems that represent the clearing of electricity and natural gas markets. This market structure and our modeling approach yields a multiple-leader/two-follower complementarity problem. We develop a modeling approach that can find equilibria with different characteristics, e.g., maximized social welfare, producer profits, or consumer welfare. We demonstrate numerically that producers aim typically to increase market prices while consumers seek to decrease them. [ABSTRACT FROM AUTHOR]

As the power industry considers moving from a deterministic paradigm to a stochastic one to clear the day-ahead market in systems with significant stochastic production, the following question arises: how should a stochastic clearing algorithm be formulated? Our analyses indicate that a three-stage stochastic approach is clearly superior to a two-stage one. To show this, we propose a model that includes three stages: the first one represents the day-ahead market, the second stage, the intraday market, and the third one, the real-time operation. The objective is to clear the day-ahead market, but with a prognosis of the future: the intraday market and the real-time operation. To assess the impact of the intraday market on the day-ahead outcomes, we compare the results obtained from the proposed model with those of a simpler but more common two-stage model, which represents the day-ahead market and the real-time operation. [ABSTRACT FROM AUTHOR]

In highly polluted regions, it might be necessary to operate power plants enforcing appropriate emission limits to help ensuring air quality. If the air quality is low and pollutants are difficult to diffuse, few additional emissions should be allowed to avoid further deteriorating the air quality. Wind power helps reducing pollution levels as it substitutes polluting thermal production. However, wind and air pollution are generally anti‐correlated. Storage can be used to shift wind power production from hours when higher pollution levels are allowed to hours when lower pollution levels are allowed to help meeting emission constraints at critical hours. Storage availability also achieves a comparatively lower production cost for the system as a whole. To represent the stochastic nature of wind power production and weather‐dependent emission limits, the authors propose a two‐stage stochastic programming model to analyse the trade‐off emission versus cost in an electric energy system with storage facilities. The first stage represents the day‐ahead scheduling, while the second one represents the real‐time operation under different scenarios. The authors analyse the impact of emission limits and/or storage on generation scheduling by comparing different models using an illustrative example and a case study based on the IEEE 118‐node system. [ABSTRACT FROM AUTHOR]

This paper analyzes the impacts of flexible demands on day-ahead market outcomes in a system with significant wind power production. We use a two-stage stochastic market-clearing model, where the first stage represents the day-ahead market and the second stage represents the real-time operation. On one hand, flexibility of demands is beneficial to the system as a whole since such flexibility reduces the operation cost, but on the other hand, shifts in demands from peak periods to off-peak periods may influence prices in such a way that demands may not be willing to provide flexibility. Specifically, we investigate the impacts of different degree of demand flexibility on day-ahead prices. A number of scenarios modeling the uncertainty associated with wind production at the operation stage, and nonconvexities due to start-up costs of generators and their minimum power outputs are taken into account. [ABSTRACT FROM AUTHOR]

This paper proposes a pricing scheme for the day-ahead market in power systems with a large percentage of renewable stochastic production. To clear the day-ahead market, instead of a simplistic deterministic model, we use a two-stage stochastic programming model that embodies a prognosis of future operating conditions. Non-convexities due to start-up costs and the on/off status of generators and their minimum power outputs are properly taken into account. Our goal is to obtain uniform day-ahead clearing prices that deviate in the least possible manner from marginal prices and that allow producers to recover their costs without uplifts. The proposed methodology is illustrated using a simple example and a realistic case study. [ABSTRACT FROM AUTHOR]

This paper presents a tri-level reliability-constrained robust power system expansion planning (SEP) framework modeling the uncertainty of electricity demand, wind power generation, and availability of units and lines. In the proposed tri-level SEP model, electricity demand and wind power generation being continuous uncertain variables are modeled by bounded intervals and availability of units and lines being discrete uncertain variables are modeled by probability distributions. Simulation results on the Garver 6-bus, the IEEE 24-bus, and the IEEE 73-bus test systems confirm the effectiveness of the proposed tri-level framework to cope with multi-fold planning uncertainties. [ABSTRACT FROM AUTHOR]

Given the significant amount of installed generation-capacity based on wind power, and also due to current economic downturn, the subsidies and incentives that have been widely used by wind-power producers to recover their investment costs have decreased and are even expected to disappear in the near future. In these conditions, wind-power producers need to develop offering strategies to make their investments profitable counting solely on the market. This paper proposes a multi-stage risk-constrained stochastic complementarity model to derive the optimal offering strategy of a wind-power producer that participates in both the day-ahead and the balancing markets. Uncertainties concerning wind-power productions, market prices, demands' bids, and rivals' offers are efficiently modeled using a set of scenarios. The conditional-value-at-risk metric is used to model the profit risk associated with the offering decisions. The proposed model is recast as a tractable mixed-integer linear programming program solvable using available branch-and-cut algorithms. Results of a case study are reported and discussed to show the effectiveness and applicability of the proposed approach. [ABSTRACT FROM PUBLISHER]

This paper proposes an efficient solution approach based on Benders’ decomposition to solve a network-constrained ac unit commitment problem under uncertainty. The wind power production is the only source of uncertainty considered in this paper, which is modeled through a suitable set of scenarios. The proposed model is formulated as a two-stage stochastic programming problem, whose first-stage refers to the day-ahead market, and whose second-stage represents real-time operation. The proposed Benders’ approach allows decomposing the original problem, which is mixed-integer nonlinear and generally intractable, into a mixed-integer linear master problem and a set of nonlinear, but continuous subproblems, one per scenario. In addition, to temporally decompose the proposed ac unit commitment problem, a heuristic technique is used to relax the inter-temporal ramping constraints of the generating units. Numerical results from a case study based on the IEEE one-area reliability test system (RTS) demonstrate the usefulness of the proposed approach. [ABSTRACT FROM AUTHOR]

Renewable energy sources are here to stay for a number of important reasons, including global warming and the depletion of fossil fuels. We explore in this paper how a thermal-dominated electric energy system can be transformed into a renewable-dominated one. This study relies on a stochastic programming model that allows representing the uncertain parameters plaguing such long-term planning exercise. Being the final year of our analysis 2050, we represent the transition from today to 2050 by allowing investment in both production and transmission facilities, with the target of achieving a renewable-dominated minimum-cost system. The methodology developed is illustrated using a realistic large-scale case study. Finally, policy conclusions are drawn. [ABSTRACT FROM PUBLISHER]

This paper proposes an optimal power flow (OPF) model with flexible AC transmission system (FACTS) devices to minimize wind power spillage. The uncertain wind power production is modeled through a set of scenarios. Once the balancing market is cleared, and the final values of active power productions and consumptions are assigned, the proposed model is used by the system operator to determine optimal reactive power outputs of generating units, voltage magnitude and angles of buses, deployed reserves, and optimal setting of FACTS devices. This system operator tool is formulated as a two-stage stochastic programming model, whose first-stage describes decisions prior to uncertainty realization, and whose second-stage represents the operating conditions involving wind scenarios. Numerical results from a case study based on the IEEE RTS demonstrate the usefulness of the proposed tool. [ABSTRACT FROM PUBLISHER]

This paper proposes an alternative formulation for the problem of dispatching generation units in a predominantly hydro system considering a medium-term horizon. Hydro-power dispatch is expressed as a equilibrium problem with equilibrium constraints (EPEC) and its solutions are compared to the solutions of a classical optimal hydro dispatch problem. A stochastic formulation is used to represent uncertainties in water inflows. Results are presented for realistic test systems with interconnected reservoirs. [ABSTRACT FROM PUBLISHER]

When deciding on wind power investments, three major issues arise: the production variability and uncertainty of wind facilities, the eventual future decline in wind power investment costs, and the significant financial risk involved in such investment decisions. Recognizing the above important issues, this paper proposes a risk-constrained multi-stage stochastic programming model to make optimal investment decisions on wind power facilities along a multi-stage horizon. The proposed model is illustrated using a clarifying example and a case study. [ABSTRACT FROM PUBLISHER]

Investment in wind power facilities involves a high level of uncertainty. To properly model such uncertainty, we consider a large number of scenarios and formulate this investment problem as a mathematical program with equilibrium constraints. The target of this problem is to maximize the profit from wind power investment in a target year, and it is subject to complementarity constraints describing a large number of market clearing conditions. Since the profit as a function of the investment variables has as sufficiently convex envelope, the considered problem can be solved by Benders decomposition. Thus, we propose, describe, and analyze a Benders decomposition algorithm to efficiently tackle the wind power investment problem. [ABSTRACT FROM PUBLISHER]

Within a weekly market horizon, this paper considers a power producer that sells its energy both in the pool and through weekly forward contracts. The paper provides a methodology that allows the producer to derive the self-scheduling of its production units, to select weekly forward contracts, and to obtain the offering strategy for Monday's pool. The proposed technique is based on stochastic programming and allows the producer to maximize its expected profit while controlling the risk of profit variability. A comprehensive case study is used to illustrate the characteristics of the proposed methodology. Appropriate conclusions are finally drawn. [ABSTRACT FROM AUTHOR]

This paper presents a technique to derive the best offering strategy for a wind power producer in an electricity market that includes various trading floors. Uncertainty pertaining to wind availability, market prices at the different trading stages, and balancing energy needs are properly taken into account. Risk on profit variability is suitably controlled at the cost of a small reduction in expected profit. The proposed technique translates into a linear programming problem of moderate size, which is readily solvable using commercially available software. A variety of numerical case studies demonstrate the interest and effectiveness of the proposed technique. Appropriate conclusions are duly drawn. [ABSTRACT FROM AUTHOR]

This paper presents a bilevel programming approach to solve the medium-term decision-making problem faced by a power retailer. A retailer decides its level of involvement in the futures market and in the pool as well as the selling price offered to its potential clients with the goal of maximizing the expected profit at a given risk level. Uncertainty on future pool prices, client demands, and rival-retailer prices is accounted for via stochastic programming. Unlike in previous approaches, client response to retail price and competition among rival retailers are both explicitly considered in the proposed bilevel model. The resulting nonlinear bilevel programming formulation is transformed into an equivalent single-level mixed-integer linear programming problem by replacing the lower-level optimization by its Karush-Kuhn-Tucker optimality conditions and converting a number of nonlinearities to linear equivalents using some well-known integer algebra results. A realistic case study is solved to illustrate the efficient performance of the proposed methodology. [ABSTRACT FROM AUTHOR]

This paper proposes a methodology to determine the required level of spinning and nonspinning reserves in a power system with a high penetration of wind power. The computation of the required reserve levels and their costs is achieved through a stochastic programming market-clearing model spanning a daily time horizon. This model considers the network constraints and takes into account the cost of both the load shedding and the wind spillage. The methodology proposed is illustrated using an example and a realistic case study. Some conclusions are finally drawn. [ABSTRACT FROM AUTHOR]

To make informed decisions in futures markets of electric energy, stochastic programming models are commonly used. Such models treat stochastic processes via a set of scenarios, which are plausible realizations throughout the decision-making horizon of the stochastic processes. The number of scenarios needed to accurately represent the uncertainty involved is generally large, which may render the associated stochastic programming problem intractable. Hence, scenario reduction techniques are needed to trim down the number of scenarios while keeping most of the stochastic information embedded in such scenarios. This paper proposes a novel scenario reduction procedure that advantageously compares with existing ones for electricity-market problems tackled via two-stage stochastic programming. [ABSTRACT FROM AUTHOR]

A generating unit can sell its production in the pool facing the financial risk inherent to volatile pool prices. Alternatively, it can sell its production through the futures market at comparatively stable but smaller average prices. Unlike the pool, forward contracts involve an obligation to sell during a specified time period, and, if the unit fails, it should buy energy from the pool to meet its futures market commitments. Hence, taking into account its forced outage rate, the unit should determine the appropriate mix of pool and futures-market involvement so that its expected profit is maximized for a prespecified risk level on profit variability. Within this context, this paper analyzes the impact of the degree of unavailability of the generating unit on its forward contracting decisions. A detailed case study illustrates the analysis performed. [ABSTRACT FROM AUTHOR]

This paper addresses the optimal involvement in a futures electricity market of a power producer to hedge against the risk of pool price volatility. The considered trading horizon spans one whole year. Recognizing the highly uncertain nature of future pool prices, a stochastic programming framework with recourse is used to model this decision-making problem. The resulting problem is a large scale mixed-integer linear programming problem. Scenario reduction techniques are used to make this problem tractable. Risk is properly modeled using the CVaR methodology. Results from a realistic case study are provided and analyzed. Some conclusions are finally drawn. [ABSTRACT FROM AUTHOR]

To procure the electric energy to be sold to its clients, a retailer has to face two major difficulties. While buying electric energy, it must cope with uncertain pool prices or sign forward contracts at higher average prices. While selling electricity, it should handle the uncertainty of the end-user demand and the fact that customers might choose a different retailer if the selling price is not competitive enough. We propose a risk-constrained stochastic programming framework to decide which forward contracts the retailer should sign and at which price it must sell electricity so that its expected profit is maximized at a given risk level. The developed model takes into account the elastic behavior of end users with respect to the selling price offered by the retailer. A realistic case study illustrates the methodology proposed. [ABSTRACT FROM AUTHOR]

This paper provides a technique based on stochastic programming to optimally solve the electricity procurement problem faced by a large consumer. Supply sources include bilateral contracts, a limited amount of self-production and the pool. Risk aversion is explicitly modeled using the conditional value-at-risk methodology. Results from a realistic case study are provided and analyzed. [ABSTRACT FROM AUTHOR]

This paper considers a profit-maximizing thermal producer that participates in a sequence of spot markets, namely, day-ahead, automatic generation control (AG C), and balancing markets. The producer behaves as a price-taker in both the day-ahead market and the AGC market but as a potential price-maker in the volatile balancing market. The paper provides a stochastic programming methodology to determine the optimal bidding strategies for the day-ahead market. Uncertainty sources include prices for the day-ahead and AGC markets and balancing market linear price variations with the production of the thermal producer. Results from a realistic case study are reported and analyzed. Conclusions are duly drawn. [ABSTRACT FROM AUTHOR]

The first of this two-paper series formulates a stochastic security-constrained multi-period electricity market clearing problem with unit commitment. The stochastic security criterion accounts for a pre-selected set of random generator and line outages with known historical failure rates and involuntary load shedding as optimization variables. Unlike the classical deterministic reserve-constrained unit commitment, here the reserve services are determined by economically penalizing the operation of the market by the expected load not served. The proposed formulation is a stochastic programming problem that optimizes, concurrently with the pre-contingency social welfare, the expected operating costs associated with the deployment of the reserves following the contingencies. This stochastic programming formulation is solved in the second companion paper using mixedinteger linear programming methods. Two cases are presented: a small transmission-constrained three-bus network scheduled over a horizon of four hours and the IEEE Reliability Test System scheduled over 24 h. The impact on the resulting generation and reserve schedules of transmission constraints and generation ramp limits, of demand-side reserve, of the value of load not served, and of the constitution of the pre-selected set of contingencies are assessed. [ABSTRACT FROM AUTHOR]

Abstract: A virtual power plant aggregates various local production/consumption units that act in the market as a single entity. This paper considers a virtual power plant consisting of an intermittent source, a storage facility, and a dispatchable power plant. The virtual power plant sells and purchases electricity in both the day-ahead and the balancing markets seeking to maximize its expected profit. Such model is mathematically rigorous, yet computationally efficient. The offering problem is cast as a two-stage stochastic mixed-integer linear programming model which maximizes the virtual power plant expected profit. The uncertain parameters, including the power output of the intermittent source and the market prices, are modeled via scenarios based upon historical data. The proposed model is applied to a realistic case study and conclusions are drawn. [Copyright &y& Elsevier]

Abstract: As wind power technology matures and reaches break-even cost, wind producers find it increasingly attractive to participate in pool markets instead of being paid feed-in tariffs. The key issue is then how a wind producer should offer in the pool markets to achieve maximum profit while controlling the variability of such profit. This paper compares two families of offering strategies based, respectively, on a naive use of wind production forecasts and on stochastic programming models. These strategies are compared through a comprehensive out-of-sample chronological analysis based on real-world data. A number of relevant conclusions are then duly drawn. [Copyright &y& Elsevier]

In some highly populated areas of the planet, air quality is an issue. Since power plants in those areas significantly contribute to air pollution due to SO 2 and NO x emissions, it is appropriate to operate such power plants enforcing admissible emission limits. Since emissions directly impact the air quality index, which is highly uncertain 24 h in advance, admissible emission limits should be modeled as stochastic parameters.To represent the stochastic nature of such admissible emission limits, we propose a two-stage stochastic programming model for the commitment of power plants. The first stage represents the day-ahead scheduling, and the second stage represents the real-time power system operation under different weather and thus admissible emission limit conditions. We show the effect of stochastic emission limits on power system operations using a simple example and a 118-node case study. [ABSTRACT FROM AUTHOR]