Stochastic security constrained unit commitment incorporating demand side reserve

Increasing the social welfare and achieving real competitive markets are the main purposes of consumers participation in electricity markets. Hence, Demand Response Programs (DRPs) have been designed to consider the consumers participation. One of these programs named Emergency Demand Response Program (EDRP) is based on consumers’ responses to high electricity prices and to the incentives that are paid by Independent System Operators (ISOs) in the critical hours. In this paper, a model has been proposed for EDRP participation in the Stochastic Security Constrained Unit Commitment (SCUC) program. Both the spinning reserve and the demand side reserve, which is modeled as EDRP, are taken into account as the operating reserve services. The effects of EDRP and its parameters on the system operation cost have been investigated under power system uncertainties. The two-stage stochastic SCUC model has been utilized for simultaneous clearing of energy and reserve markets. Mixed Integer Linear Programming (MILP) has been used for modeling the proposed method in the GAMS (General Algebraic Modeling System) environment. The proposed model is applied to a 6-bus test system and a modified IEEE Reliability Test System (RTS) to demonstrate its effectiveness.     

Analysis of FACTS devices on Security Constrained Unit Commitment problem

This paper focuses on solving Security Constrained Unit Commitment (SCUC) problem using ABC algorithm incorporating FACTS devices. The objective of the SCUC problem is to obtain the minimum operating cost simultaneously maintaining the security of the system. The SCUC problem is decomposed into Unit Commitment (UC), the master problem and Security-Constrained Economic Dispatch (SCED) as the sub-problem. The existing generation constraints, such as hourly power demand, system reserves, and minimum up/down time limits, ramp up/down limits are included in the SCUC problem formulation. The ability of FACTS devices to control the power flow through designated routes in transmission lines and thereby reducing the overloading of lines are studied. The solution of SCUC problem is also analyzed during a single line outage contingency. The SCUC is carried out incorporating FACTS devices such as SVC, TCSC, STATCOM, SSSC, UPFC and IPFC. The modeling of the FACTS devices within the power system network and finding a suitable location are discussed. The SCUC has been solved and validated on an IEEE 118-bus test system and a practical South Indian 86 bus utility.     

深度学习辅助约束辨识的电力市场快速出清方法

日前电力市场出清需要求解大规模安全约束经济调度问题,尽管实际采用线化处理方法,但需要考虑N－1场景下的大量安全约束,导致其规模庞大,难以快速求解。提出了一种深度学习辅助的日前市场快速出清方法,以满足快速出清计算场合的应用需求。首先,设计基于深度神经网络的安全约束经济调度模型计算框架,将深度学习技术应用于日前电力市场出清计算过程,兼顾安全约束经济调度模型的求解速度和求解精度;其次,提出面向起作用约束辨识的深度学习策略,从特征向量、深度神经网络结果处理2个方面,实现安全约束经济调度起作用约束集的辨识,并将其纳入日前市场出清模型,以简化模型的复杂度;最后,在接入新能源的IEEE 30标准测试系统中验证了所述方法的有效性。利用深度神经网络预辨识安全约束经济调度模型的起作用约束,有利于降低模型复杂度,提高日前市场出清的计算效率。     

深度学习辅助约束辨识的电力市场快速出清方法

日前电力市场出清需要求解大规模安全约束经济调度问题,尽管实际采用线化处理方法,但需要考虑N－1场景下的大量安全约束,导致其规模庞大,难以快速求解。提出了一种深度学习辅助的日前市场快速出清方法,以满足快速出清计算场合的应用需求。首先,设计基于深度神经网络的安全约束经济调度模型计算框架,将深度学习技术应用于日前电力市场出清计算过程,兼顾安全约束经济调度模型的求解速度和求解精度;其次,提出面向起作用约束辨识的深度学习策略,从特征向量、深度神经网络结果处理2个方面,实现安全约束经济调度起作用约束集的辨识,并将其纳入日前市场出清模型,以简化模型的复杂度;最后,在接入新能源的IEEE 30标准测试系统中验证了所述方法的有效性。利用深度神经网络预辨识安全约束经济调度模型的起作用约束,有利于降低模型复杂度,提高日前市场出清的计算效率。     

SCUC／SCED问题分析

安全约束机组组合（SCUC）与安全约束经济调度（SCED）同时考虑电力系统运行的安全性和经济性,在国外受到广泛关注．基于现代混合整数规划（MIP）算法的SCUC近几年得到实际应用和快速发展,文中对SCUC／SCED研究的最新进展进行了综述。概述了SCUC问题的优化模型和算法的研究状况,分类介绍了基于MIP算法的SCUC研究进展．并分析和评述了主要方法、原理及其优缺点,进一步提出我国开展SCUC／SCED研究急待解决的问题。     

基于改进约束序优化方法的带安全约束的不确定性机组组合问题研究

为快速有效地求解考虑间歇性可再生能源接入的SCUC问题,在现有约束序优化的基础上,提出了一种适用于不确定性SCUC问题求解的改进约束序优化算法。该方法针对序优化的粗糙模型和精确模型分别融入了离散变量识别策略和无效安全约束削减策略。相比于传统求解方法,提出的改进约束序优化算法在充分发挥传统序优化计算效率方面优势的同时,进一步增强了算法的紧凑性,降低了计算冗余度,有效提升了算法的求解效率。基于IEEE-118节点标准算例的仿真验证了所提算法的正确性和有效性。     

运行备用市场定价机制对机组报价行为的影响

运行备用是保障电力系统高可靠供电的必要手段。能量市场和运行备用市场之间存在着内在联系。如果备用市场的定价机制能够引起能量市场出清价格的变化，对系统的影响将很大。文中首先按照不同的备用市场定价原则，即市场出清价格方式和按报价结算方式，获得了机组基于成本报价时的利润分布，并分析了分布的合理性及其对机组报价行为的影响。然后利用博弈模型，考察了备用市场不同的定价方法对能量市场机组的报价行为的影响，认为在运行备用市场使用按报价结算方式定价比统一的出清电价方式能更好地引导机组在能量市场报成本价格。仿真计算结果说明了理论分析的意义。     

含风电和储能的电力系统安全约束机组组合问题研究

针对含风电的电力系统安全约束机组组合问题,利用储能的快速双向调节能力改善风电出力的波动性和不确定性,提高系统运行的经济性。机组组合模型中考虑风电、负荷预测误差对机组组合的影响,并引入基于可信理论的旋转备用机会约束条件,可避免造成备用的过度配置。另外,通过考虑合理弃风提高消纳风电能力和降低系统运行成本,分析了储能参数以及接入系统方式对机组组合模型中系统运行成本和风电弃风量的影响。最后,采用新英格兰10机系统验证了所提模型的有效性。     

Relative merits of load following reserves & energy storage market integration towards power system imbalances

Traditionally, power system balancing operations consist of three consecutive control techniques, namely security-constrained unit commitment (SCUC), security-constrained economic dispatch (SCED), and automatic generation control (AGC). Each of these have their corresponding type of operating reserves. Similarly, energy storage resources (ESRs) may be integrated as energy, load following, or regulation resources. A review of the existing literature shows that most ESR integration studies are focused on a single control function. In contrast, recent work on renewable energy integration has employed the concept of enterprise control where the multiple layers of balancing operations have been integrated into a single model. This paper now uses such an enterprise control model to demonstrate the relative merits of load following reserves and energy storage integrated into the resource scheduling and balancing action layers. The results show that load following reserves and energy storage resources mitigate imbalances in fundamentally different ways. The latter becomes an increasingly effective balancing resource for high net-load variability and small day-ahead market time step.     

基于多智能体强化学习的电力现货市场定价机制研究(二):结合理论与仿真的定价机制决策框架

该文研究了电力现货市场定价机制的选择和设计问题。首先,初步分析系统边际电价、节点边际电价和分区边际电价等3种定价机制的适用性,提出理论分析和市场仿真相结合的电力现货市场定价机制决策框架。理论分析方面,梳理了电力现货市场定价机制的理论和国际实践,给出3种定价机制的出清模型和电价计算方式,基于安全约束经济调度模型建立兼容性再调度模型;从价格信号作用、对中长期市场交易开展的影响、市场监管和配套机制需求等多个维度,对3种定价机制的特点进行对比分析,并探讨定价机制与市场模式的配合关系和我国定价机制设计中需要考虑的关键问题。市场仿真方面,基于3种定价机制下的发电商报价双层优化模型,对IEEE 39系统进行仿真计算。最后,综合分析仿真结果和定价机制特点,为算例系统选择合适的定价机制,验证了所提电力现货市场定价机制决策框架的有效性。2篇论文共同解决了定价机制设计与发电商报价策略的嵌套难题。     

Evaluation of costs for procuring reserve capacity in a deregulated power system using multi‐agent model

In this paper, we assume 2 models for securing reserve capacity. One is “Commitment‐based Security Model” and the other is “Reserve Market‐based Security Model.” In Commitment‐based security model, ISO commits procurement of reserve energy to a particular generation company. Meanwhile, in Reserve market‐based security model, ISO procures reserve energy through reserve market. The main object of this research is to investigate which model will be preferable for the viewpoint of consumer's cost. To compare these models, two things are considered in this paper. One is bidding behavior of agents which bids to energy market and reserve market. To consider this, Q‐Learning of multi‐agent model is used. Also, the Unit Commitment (UC) is considered to calculate generation cost. This is to calculate the cost for securing reserve power more precisely. © 2009 Wiley Periodicals, Inc. Electr Eng Jpn, 167(1): 18– 25, 2009; Published online in Wiley InterScience ( www.interscience.wiley.com ). DOI 10.1002/eej.20705     

Market-Based Price Differentials in Zonal and LMP Market Designs

The California ISO is redesigning procedures for scheduling, dispatch, and congestion management, which are critical to reliable, nondiscriminatory transmission service. The redesign implements security constrained unit commitment and locational marginal pricing to reflect actual costs of delivering energy, using an accurate network model to price both congestion and losses. CAISO simulations of the market redesign are a unique opportunity to compare a) estimated LMPs if the redesign were in place during recent historical periods with b) actual conditions, as logged in operating records. The study uses market schedules and bids in an alternating current optimal power flow simulation. Resulting LMPs are similar within current congestion zones, but price differences occur during high loads, reflecting actual changes in system conditions. The frequency and magnitude of LMP differences are consistent with the current market, suggesting that the increased market transparency will produce stable, predictable prices. Case studies demonstrate that optimization using SCUC and SCED, coupled with the state estimator, allows more appropriate responses to system conditions, improved utilization of transmission capacity, reduction in congestion costs, and enhanced system reliability.     

电力市场中弹性运行备用研究

备用市场组织方式关系到备用市场稳定运行.传统顺序出清方式中,备用出清的先后顺序对出清结果影响重大,市场的公平性不容易保障.该文在弹性备用框架下,提出各类备用联合出清的新型交易形式.引入可靠性指标LOLP（电力不足概率）,将备用容量报价和电量报价结合起来作为综合报价指标,并分析了按报价付费模式和统一出清价模式在备用市场中的应用.在此基础上,结合可靠性评估理论,建立了同时出清10分钟备用和30分钟备用的弹性备用综合模型.文章采用启发式算法求解,简便实用.标准测试系统RTS96的仿真计算结果表明该模型的有效性     

全周期变时段“三公”调度发电计划优化模型

以年度合同进度为主要控制目标的"三公"调度是我国目前最为普遍的调度模式,常用的安全约束机组组合(security constrained unit commitment,SCUC)/安全约束经济调度(security constrained economic dispatch,SCED)模型不能满足"三公"调度计算周期要求,为此建立了考虑电网安全约束的全周期发电计划优化模型,并依据不同时段发电计划的需求差异,进一步建立了全周期变时段发电计划优化模型,该模型可根据时段长度自动选择各种约束。算例结果表明全周期变时段发电计划优化模型能够适应"三公"调度需要,计算性能能够满足大规模电网发电计划精细化管理需要。     

两种第三方主体参与的现货市场出清模式比较

随着高比例新能源接入电力系统，常规机组开机减少，系统备用容量不足，需要充分挖掘第三方主体的备用能力。本文提出并比较了第三方主体参与备用辅助服务的2种现货市场出清模式：电能量和机组备用联合出清+第三方主体备用单独出清模式、电能量和两侧备用联合出清模式。论文首先结合市场发展阶段、备用资源紧缺程度等方面分析了2种出清模式的优缺点；然后，给出了2种市场模式下的出清模型，并分析了第三方主体参与对市场出清结果的影响及其费用和收益情况。最后，采用改进的IEEE30节点系统进行了仿真，结果表明电能量和两侧备用联合出清模式可以更加充分调用第三方主体的备用，降低系统发电和备用总成本。     

一种售电公司经济调度双层优化方法

为解决大型售电公司应对新能源接入和需求侧波动带来的随机性风险问题,文章提出了一种基于随机规划的双层优化模型:上层子问题以出清电价和出清电量为决策变量,模拟了市场电能与备用辅助服务的联合出清过程;下层子问题以售电公司的最优购电成本为目标,模拟了售电公司的经济调度行为。模型的求解过程中,通过机会约束规划将模型转换为确定性模型,再由CPLEX求解器分别求解上下层子模型,并通过上下层的迭代求得均衡解。最后,建立了不计及售电公司决策影响的两阶段模型作为对比模型,通过算例分析,验证了该模型可有效反映售电公司决策行为对出清过程的影响,并阐明了不同置信水平对售电公司决策的影响。     

一种售电公司经济调度双层优化方法

为解决大型售电公司应对新能源接入和需求侧波动带来的随机性风险问题,文章提出了一种基于随机规划的双层优化模型:上层子问题以出清电价和出清电量为决策变量,模拟了市场电能与备用辅助服务的联合出清过程;下层子问题以售电公司的最优购电成本为目标,模拟了售电公司的经济调度行为。模型的求解过程中,通过机会约束规划将模型转换为确定性模型,再由CPLEX求解器分别求解上下层子模型,并通过上下层的迭代求得均衡解。最后,建立了不计及售电公司决策影响的两阶段模型作为对比模型,通过算例分析,验证了该模型可有效反映售电公司决策行为对出清过程的影响,并阐明了不同置信水平对售电公司决策的影响。     

南方（以广东起步）电力备用市场设计

现阶段南方区域仍按“两个细则”管理电力备用,其备用品种、价格机制、结算机制及备用资源的分配方式均有不尽合理之处,亟须建设南方（以广东起步）电力备用市场,完善南方区域电力市场体系。通过对国外电力备用市场机制和南方区域备用管理现状问题的分析,提出了适应南方区域电力市场发展现状的电力备用市场机制,对备用市场出清模型进行了探究,并通过算例对不同模式下的备用市场建模和出清结果进行了分析,结果表明,在报量报价和报量不报价的备用市场模式下,备用出清价格都已包含机会成本,可实现备用中标容量的机会成本补偿,备用与电能量市场联合优化机制可确保参与备用市场的机组收益不低于只参与电能量的情形,对机组积极参与备用市场起到了正向激励作用。     

南方（以广东起步）电力备用市场设计

现阶段南方区域仍按“两个细则”管理电力备用,其备用品种、价格机制、结算机制及备用资源的分配方式均有不尽合理之处,亟须建设南方（以广东起步）电力备用市场,完善南方区域电力市场体系。通过对国外电力备用市场机制和南方区域备用管理现状问题的分析,提出了适应南方区域电力市场发展现状的电力备用市场机制,对备用市场出清模型进行了探究,并通过算例对不同模式下的备用市场建模和出清结果进行了分析,结果表明,在报量报价和报量不报价的备用市场模式下,备用出清价格都已包含机会成本,可实现备用中标容量的机会成本补偿,备用与电能量市场联合优化机制可确保参与备用市场的机组收益不低于只参与电能量的情形,对机组积极参与备用市场起到了正向激励作用。     

Stochastic Multiobjective Market Clearing of Joint Energy and Reserves Auctions Ensuring Power System Security

In this paper, a stochastic multiobjective framework is proposed for day-ahead joint market clearing. The proposed multiobjective framework can concurrently optimize competing objective functions including augmented generation offer cost and security indices (overload index, voltage drop index, and voltage stability margin). Besides, system uncertainties including generating units and branches contingencies and load uncertainty are explicitly considered in the stochastic market clearing scheme. The solution methodology consists of two stages, which firstly, employs roulette wheel mechanism and Monte Carlo simulation (MCS) for random adaptive 24-h scenario generation wherein the stochastic multiobjective market clearing procedure is converted into its respective deterministic equivalents (scenarios). In the second stage, for each deterministic scenario, a multiobjective mathematical programming (MMP) formulation based on the epsiv -constrained approach is implemented for provision of spinning reserve (SR) and nonspinning reserve (NSR) as well as energy. The MMP formulation of the market clearing process is optimized while meeting AC power flow constraints and expected interruption cost (EIC). The IEEE 24-bus Reliability Test System (RTS 24-bus) is used to demonstrate the performance of the proposed method.