Yes / The classical generation scheduling problem defines on/off decisions (commitment) and dispatch level of all available generators in a power system for each scheduling period. In recent years researchers have focused on developing new approaches to solve nonclassical generation scheduling problems in the newly deregulated and decentralized electricity market place. In this paper a GA-based approach has been developed for a system operator to schedule generation in a market akin to that operating in England and Wales. A generation scheduling problem has been formulated and solved using available trading information at the time of dispatch. The solution is updated after information is obtained in a rolling fashion. The approach is tested for two IEEE network-based problems, and achieves comparable results with a branch and bound technique in reasonable CPU time.

Yes / The solution of generation scheduling (GS) problems involves the determination of the unit commitment (UC) and economic dispatch (ED) for each generator in a power system at each time interval in the scheduling period. The solution procedure requires the simultaneous consideration of these two decisions. In recent years researchers have focused much attention on new solution techniques to GS. This paper proposes the application of a variety of genetic algorithm (GA) based approaches and investigates how these techniques may be improved in order to more quickly obtain the optimum or near optimum solution for the GS problem. The results obtained show that the GA-based hybrid approach offers an effective alternative for solving realistic GS problems within a realistic timeframe.

Doctor of Philosophy / Department of Electrical and Computer Engineering / Sanjoy Das / Anil Pahwa / The task of short-term optimal thermal generation scheduling can be cast in the form of a multi-objective optimization problem. The goal is to determine an optimal operating strategy to operate power plants, in such a way that certain objective functions related to economic and environmental issues, as well as transmission losses are minimized, under typical system and operating constraints. Due to the problem’s inherent complexity, and the large number of associated constraints, standard multi-objective optimization algorithms fail to yield optimal solutions. In this dissertation, a novel, two-phase multi-objective evolutionary approach is proposed to address the short-term optimal thermal generation scheduling problem. The objective functions, which are based on operation cost, emission and transmission losses, are minimized simultaneously. During the first phase of this approach, hourly optimal dispatches for each period are obtained separately, by minimizing the operation cost, emission and transmission losses simultaneously. The constraints applied to this phase are the power balance, spinning reserve and power generation limits. Three well known multi-objective evolutionary algorithms, NSGA-II, SPEA-2 and AMOSA, are modified, and several new features are added. This hourly schedule phase also includes a repair scheme that is used to meet the constraint requirements of power generation limits for each unit as well as balancing load with generation. The new approach leads to a set of highly optimal solutions with guaranteed feasibility. This phase is applied separately to each hour long period. In the second phase, the minimum up/down time and ramp up/down rate constraints are considered, and another improved version of the three multi-objective evolutionary algorithms, are used again to obtain a set of Pareto-optimal schedules for the integral interval of time (24 hours). During this phase, the hourly optimal schedules that are obtained from the first phase are used as inputs. A bi-objective version of the problem, as well as a three-objective version that includes transmission losses as an objective, are studied. Simulation results on four test systems indicate that even though NSGA-II achieved the best performance for the two-objective model, the improved AMOSA, with new features of crossover, mutation and diversity preservation, outperformed NSGA-II and SPEA-2 for the three-objective model. It is also shown that the proposed approach is effective in addressing the multi-objective generation dispatch problem, obtaining a set of optimal solutions that account for trade-offs between multiple objectives. This feature allows much greater flexibility in decision-making. Since all the solutions are non-dominated, the choice of a final 24-hour schedule depends on the plant operator’s preference and practical operating conditions. The proposed two-phase evolutionary approach also provides a general frame work for some other multi-objective problems relating to power generation as well as in other real world applications.

The focus of this thesis is the quantification of the value of operation flexibility in systems with large penetration of wind generation. This begins with the quantification of the impact of wind generation (WG) uncertainty on the system's needs for frequency regulation and reserve. This is done by combing the stochastic behaviour of wind generation, demand uncertainty and generation outages. Two different approaches are compared to access the implications of using normal distribution approximations or direct representations of different sources of uncertainty. This is followed by an investigation of the relative impact of different sources of uncertainty on the reserve levels. For large wind penetration, wind becomes the dominant source of uncertainty driving most of the need for reserve. Procuring such large requirements increases the need for flexibility and the overall system operations cost. To mitigate these additional costs and improve system flexibility, the study explores the use of a combination of spinning and standing reserve to meet the increased reserve requirements. This combination minimises the cost of reserve and increases system flexibility. These benefits are more pronounced if a more accurate representation of uncertainty is used. Following this, a detailed analysis of the value of generation flexibility is performed. The analysis is based on the modification of traditional scheduling models to include WG and to take into account the relevant features of system operation flexibility. The value of flexibility is quantified for different conventional generation mix, different response and reserve technology compositions and generation technology flexibility, across a wide range of wind penetration levels. The key drivers for the value of flexibility are shown to be the increased response and reserve requirements (especially reserve requirements), the conventional generation mix and the inherent flexibility of must-run generation. This is driven mostly by the system's need for curtailing wind to maintain the generation/demand balance. To obtain a significant reduction of carbon emissions, however, a combination of must-run generation with a large penetration of wind is required. This results in a high economic and environmental value being placed on must-run generation flexibility. The high economic and environmental value attributed to flexibility is seen as an opportunity to explore alternative sources of flexibility, such as storage and demand side flexibility (DSF). To this end, this work also investigates the role that such enabling technologies can play in enhancing system flexibility, by contributing to standing reserve and load-levelling. To this end a new system operation tool is developed. This tool simulates system operation for forecasted and realised wind generation to optimise reserve scheduling and utilisation. This is required to quantify the value of value of using storage and DSF to provide reserve. This tool is used to quantify the economic and environmental value of these technologies for different conventional generation mix and wind penetration. The studies show that both technologies have economic and environmental benefits and this is more pronounced for low flexible conventional generation mix and higher wind penetration. The value is driven mostly from increasing the system's ability of using WG. This highlights the role that storage and DSF can play in enabling low carbon systems composed by a combination of low flexible conventional generation with a large wind penetration. Finally, the thesis also examines the role of storage and DSM to support network operation, particularly in systems like the UK where the connection of WG capacity is limited by network constraints. The use of these technologies, to increase network flexibility in situations of congestion, is explored through the development and application of a multi-period optimal power flow with storage and DSM included as part of the optimisation constraints. The study concludes that both technologies present benefits and have a complementary role. Its value is maximised under different conditions and depends on the cost of generation and location of demand, across the network.

Flexibility describes the system ability to cope with events that may cause imbalance between electricity supply and demand while maintaining the system reliability in a cost-effective manner. Flexibility has always been present in the power system to cater for unplanned generator outages and demand uncertainty and variability. The arrival of wind generation with its variable and hard to predict nature increases the overall needs for system flexibility. This thesis provides a systematic approach for investigating the role of flexibility in different power system activities including generation scheduling, generation planning and market operation, and furthermore proposes two 'offline' indices for flexibility evaluation. Using the tools and metrics presented in this thesis, it is possible to perform the following tasks: • Conduct generation scheduling simulation to evaluate the impacts of wind on the flexibility requirement of power systems; • Use the unit construction and commitment algorithm to 1) estimate the maximum allowable wind capacity for an existing system; 2) find the optimal investment of new flexible units for accommodating more wind generation; and 3) decide an optimal generation mix for integrating a given wind penetration; • Use the market model to reveal the value and profitability of flexibility and evaluate the corresponding effects of alternative market design; • Use the two proposed flexibility indices to quantitatively assess the flexibility of individual generators and power systems without undertaking complex and time consuming simulations.

Our work is driven by a class of practical problems of sequential decision making in the context of electric power generation under uncertainties. These problems are usually treated as receding horizon deterministic optimization problems, and/or as scenario-based stochastic programs. Stochastic programming allows to compute a first stage decision that is hedged against the possible futures and -- if a possibility of recourse exists -- this decision can then be particularized to possible future scenarios thanks to the information gathered until the recourse opportunity. Although many decomposition techniques exist, stochastic programming is currently not tractable in the context of day-ahead electric power generation and furthermore does not provide an explicit recourse strategy. The latter observation also makes this approach cumbersome when one wants to evaluate its value on independent scenarios. We propose a supervised learning methodology to learn an explicit recourse strategy for a given generation schedule, from optimal adjustments of the system under simulated perturbed conditions. This methodology may thus be complementary to a stochastic programming based approach. With respect to a receding horizon optimization, it has the advantages of transferring the heavy computation offline, while providing the ability to quickly infer decisions during online exploitation of the generation system. Furthermore the learned strategy can be validated offline on an independent set of scenarios. On a realistic instance of the intra-day electricity generation rescheduling problem, we explain how to generate disturbance scenarios, how to compute adjusted schedules, how to formulate the supervised learning problem to obtain a recourse strategy, how to restore feasibility of the predicted adjustments and how to evaluate the recourse strategy on independent scenarios. We analyze different settings, namely either to predict the detailed adjustment of all the generation units, or to predict more qualitative variables that allow to speed up the adjustment computation procedure by facilitating the ``classical' optimization problem. Our approach is intrinsically scalable to large-scale generation management problems, and may in principle handle all kinds of uncertainties and practical constraints. Our results show the feasibility of the approach and are also promising in terms of economic efficiency of the resulting strategies. The solutions of the optimization problem of generation (re)scheduling must satisfy many constraints. However, a classical learning algorithm that is (by nature) unaware of the constraints the data is subject to may indeed successfully capture the sensitivity of the solution to the model parameters. This has nevertheless raised our attention on one particular aspect of the relation between machine learning algorithms and optimization algorithms. When we apply a supervised learning algorithm to search in a hypothesis space based on data that satisfies a known set of constraints, can we guarantee that the hypothesis that we select will make predictions that satisfy the constraints? Can we at least benefit from our knowledge of the constraints to eliminate some hypotheses while learning and thus hope that the selected hypothesis has a better generalization error? In the second part of this thesis, where we try to answer these questions, we propose a generic extension of tree-based ensemble methods that allows incorporating incomplete data but also prior knowledge about the problem. The framework is based on a convex optimization problem allowing to regularize a tree-based ensemble model by adjusting either (or both) the labels attached to the leaves of an ensemble of regression trees or the outputs of the observations of the training sample. It allows to incorporate weak additional information in the form of partial information about output labels (like in censored data or semi-supervised learning) or -- more generally -- to cope with observations of varying degree of precision, or strong priors in the form of structural knowledge about the sought model. In addition to enhancing the precision by exploiting information that cannot be used by classical supervised learning algorithms, the proposed approach may be used to produce models which naturally comply with feasibility constraints that must be satisfied in many practical decision making problems, especially in contexts where the output space is of high-dimension and/or structured by invariances, symmetries and other kinds of constraints.

abstract: The uncertainty and variability associated with stochastic resources, such as wind and solar, coupled with the stringent reliability requirements and constantly changing system operating conditions (e.g., generator and transmission outages) introduce new challenges to power systems. Contemporary approaches to model reserve requirements within the conventional security-constrained unit commitment (SCUC) models may not be satisfactory with increasing penetration levels of stochastic resources; such conventional models pro-cure reserves in accordance with deterministic criteria whose deliverability, in the event of an uncertain realization, is not guaranteed. Smart, well-designed reserve policies are needed to assist system operators in maintaining reliability at least cost. Contemporary market models do not satisfy the minimum stipulated N-1 mandate for generator contingencies adequately. This research enhances the traditional market practices to handle generator contingencies more appropriately. In addition, this research employs stochastic optimization that leverages statistical information of an ensemble of uncertain scenarios and data analytics-based algorithms to design and develop cohesive reserve policies. The proposed approaches modify the classical SCUC problem to include reserve policies that aim to preemptively anticipate post-contingency congestion patterns and account for resource uncertainty, simultaneously. The hypothesis is to integrate data-mining, reserve requirement determination, and stochastic optimization in a holistic manner without compromising on efficiency, performance, and scalability. The enhanced reserve procurement policies use contingency-based response sets and post-contingency transmission constraints to appropriately predict the influence of recourse actions, i.e., nodal reserve deployment, on critical transmission elements. This research improves the conventional deterministic models, including reserve scheduling decisions, and facilitates the transition to stochastic models by addressing the reserve allocation issue. The performance of the enhanced SCUC model is compared against con-temporary deterministic models and a stochastic unit commitment model. Numerical results are based on the IEEE 118-bus and the 2383-bus Polish test systems. Test results illustrate that the proposed reserve models consistently outperform the benchmark reserve policies by improving the market efficiency and enhancing the reliability of the market solution at reduced costs while maintaining scalability and market transparency. The proposed approaches require fewer ISO discretionary adjustments and can be employed by present-day solvers with minimal disruption to existing market procedures. / Dissertation/Thesis / Doctoral Dissertation Electrical Engineering 2018

Keeping a contingency reserve in power systems is necessary to preserve the security of real-time operations. This work studies two different approaches to the optimal allocation of energy and reserves in the day-ahead generation scheduling process. Part I presents a stochastic security-constrained unit commitment model to co-optimize energy and the locational reserves required to respond to a set of uncertain generation contingencies, using a novel state-based formulation. The model is applied in an offer-based electricity market to allocate contingency reserves throughout the power grid, in order to comply with the N-1 security criterion under transmission congestion. The objective is to minimize expected dispatch and reserve costs, together with post contingency corrective redispatch costs, modeling the probability of generation failure and associated post contingency states. The characteristics of the scheduling problem are exploited to formulate a computationally efficient method, consistent with established operational practices. We simulated the distribution of locational contingency reserves on the IEEE RTS96 system and compared the results with the conventional deterministic method. We found that assigning locational spinning reserves can guarantee an N-1 secure dispatch accounting for transmission congestion at a reasonable extra cost. The simulations also showed little value of allocating downward reserves but sizable operating savings from co-optimizing locational nonspinning reserves. Overall, the results indicate the computational tractability of the proposed method. Part II presents a distributed generation scheduling model to optimally allocate energy and spinning reserves among competing generators in a day-ahead market. The model is based on the coordination between individual generators and a market entity. The proposed method uses forecasting, augmented pricing and locational signals to induce efficient commitment of generators based on firm posted prices. It is price-based but does not rely on multiple iterations, minimizes information exchange and simplifies the market clearing process. Simulations of the distributed method performed on a six-bus test system showed that, using an appropriate set of prices, it is possible to emulate the results of a conventional centralized solution, without need of providing make-whole payments to generators. Likewise, they showed that the distributed method can accommodate transactions with different products and complex security constraints.

Through integrated modeling of power system operations and market risks, this thesis addresses a variety of important issues on market signals modeling, generation capacity scheduling, and electricity forward trading. The first part of the thesis addresses a central problem of transmission investment which is to model market signals for transmission adequacy. The proposed system simulation framework, combined with the stochastic price model, provides a powerful tool for capturing the characteristics of market prices dynamics and evaluating transmission investment. We advocate the use of an AC power flow formulations instead since it allocates transmission losses correctly and reveals the economic incentives of voltage requirements. By incorporating reliability constraints in the market dispatch, the resulting market prices yield incentives for market participants to invest in additional transmission capacity. The second part of the thesis presents a co-optimization modeling framework that incorporates market participation and market price uncertainties into the capacity allocation decision-making problem through a stochastic programming formulation. Optimal scenario-dependent generation scheduling strategies are obtained. The third part of the thesis is devoted to analyzing the risk premium present in the electricity day-ahead forward price over the real-time spot price. This study establishes a quantitative model for incorporating transmission congestion into the analysis of electricity day-ahead forward risk premium. Evidences from empirical studies confirm the significant statistical relationship between the day-ahead forward risk premium and the shadow price premiums on transmission flowgates.

[pt] O sistema elétrico de potência brasileiro, denominado Sistema Interli- gado Nacional – SIN, possui como órgão responsável pela operação o Op- erador Nacional do Sistema Elétrico – ONS. A fim de utilizar os recursos energéticos de forma a garantir a qualidade, confiabilidade e segurança no suprimento de energia elétrica ao menor custo total de operação, o oper- ador utiliza uma cadeia de modelos de otimização que subsidia a tomada de decisão no Programa Diário de Operação, implementado diariamente nas salas de controle do ONS e de agentes de geração para operação em tempo real. A etapa de Programação Diária do Operador Nacional do Sistema Elétrico busca estabelecer o despacho centralizado da geração e das reser- vas de potência a fim de atender à demanda prevista de energia elétrica considerando os limites da rede elétrica, das tecnologias de geração e a in- certeza de disponibilidade das unidades geradores e linhas de transmissão. Este trabalho propõe um modelo computacional programado em código aberto para a programação diária implementado na linguagem Julia. O modelo pertence à classe de modelos de unit commitment e considera a cootimização do despacho de geração e definição dos níveis de reservas em cada gerador do SIN para atender a critérios de segurança do tipo N − K . / [en] The Brazilian electric power system, called the National Interconnected System - SIN ( Sistema Interligado Nacional), has as its responsible institu- tion for operation the National Electric System Operator - ONS (Operador Nacional do Sistema Elétrico). In order to manage energy resources to en- sure quality, reliability and security of electricity supply at the lowest total operating cost, the operator uses a chain of optimization models that feeds the Daily Operation Program for decision-making, which is implemented everyday in the ONS and generators control rooms for real-time operation. The Daily Scheduling phase of the National Electric System Operator seeks to establish the centralized dispatch of generation and power reserves in order to meet the expected demand for electricity considering the limits of both the electrical grid and the generation technologies, along with the uncertainty of availability of generator units and transmission lines. This work proposes a computational model programmed in open-source for daily operation programming, implemented in the Julia language. The model be- longs to the unit commitment model class and it considers the generation dispatch cooptimization and reserve levels definition in each SIN generator to meet N-K safety criteria.