【作者】 邓佳佳；

【导师】 黄元生；

【作者基本信息】 华北电力大学 ， 技术经济及管理， 2012， 博士

【摘要】 随着风电和光伏发电等间歇式分布式能源在电网中比重的加大,其出力的随机波动性给电网安全运行带来了很大风险,需要从电网优化调度以及网络优化、规划等多方面研究以缓解分布式能源发电对电网安全可靠运行带来的巨大挑战。考虑分布式能源的电力系统优化运营问题是一个大规模、多目标、多约束、非线性的优化问题,其控制变量具有离散性与连续性相混合的特点,因此,该优化运营问题的建模与求解算法就成了十分重要的研究课题。根据上述分析以及负荷预测对于调度运营的重要性,本文对短期负荷预测模型、考虑分布式能源的电力系统优化运营问题展开研究,包括日前闭环经济调度问题,考虑风电接入的网络潮流优化问题和考虑分布式能源的网络规划问题,具体研究内容包括以下几个方面：(1)短期负荷预测是日前经济调度的重要依据,提出了一种基于GARCH误差修正模型的LS-SVM短期负荷预测模型,利用LS-SVM模型对日前负荷进行预测,采用GARCH误差修正模型对预测误差进行修正从而提高预测的精度。(2)建立了一种基于闭环控制的日前经济调度模型,线性化处理考虑机组组合问题的经济调度模型以及各种约束,同时考虑机组的启停成本,电网运行的安全约束,包括支路输电容量约束以及安全约束。模型求解采用基于预防、校正的交替求解方法,对机组组合模型和基于安全约束的经济调度模型进行迭代交替求解。(3)建立了一种考虑风电并网的多目标电力系统供电网络优化运营模型,综合考虑发电成本最小和电压稳定裕度最大两个优化目标。问题求解采用了自适应调整的粒子群算法对模型进行优化求解。(4)建立了一种基于变权预测校正内点法的分布式网络规划模型,对不同类型的可再生分布式发电机组进行优化配置,以网络损耗最小为优化目标,同时考虑了所有分布式发电机组可能的运行方式。为了提高求解效率,模型采用了直角坐标潮流方程和约束条件。对上述问题的建模与求解不仅可以实现电力系统运行安全性、经济性、环保性等要求,还可以降低发电、输电成本,协调电厂与电网、电网与用户之间的冲突,因此,对考虑分布式能源的电力系统优化运营问题进行研究具有重要的理论价值和现实意义。更多还原

【Abstract】 With the increasing of intermittent wind power and photovoltaic power in distributed grid, stochastic volatility of power generation has brought risk to the operation of power grids. Therefore, we need alleviate the great challenges of distributed energy generation on the grid safe and reliable operation by optimizing the scheduling, dispatching, planning and other aspects from the grid. Operation optimization models of power system considering distributed energy is a large-scale, multi-objective, multi-constraint and nonlinear optimization problem, and control variables have the characteristics of mixed discrete and continuous phase. Therefore, optimization models of power system operation and their solution algorithm considering the distributed energy system have become an important research topic.According to analysis mentioned above and the importance of load forecasting to scheduling operations, short-term load forecasting model and optimization models of power system operation are analyzed in this paper, including closed-loop economic dispatch problem, network optimization problems considering wind power access and network planning problem considering distributed energy. Specific research topics are listed in the following:(1) Short-term load forecasting is an important basis for day-ahead economic dispatch, LS-SVM short-term load forecasting model based on GARCH error correction is proposed in this paper, using LS-SVM model to predict load data. GARCH model is applied to forecast error correction to improve the accuracy of prediction.(2) The day-ahead economic dispatch model based on closed-loop control is presented, linearizing a variety of constraints and taking into account the unit cost of start and stop, security constraints of power grid operation, including the power flow constraints, voltage security constraints as well as N-1security constraints. Prevention-correction method is employed for solving the unit commitment model and security-constrained economic dispatch model alternately.(3) A multi-objective power flow optimization model is established considering the wind power and system voltage stability margin simultaneously, including the minimum cost of power generation and maximum voltage stability margin. The problem was solved using the adaptive particle swarm optimization optimization algorithm.(4) Distributed network planning model based on varied-weight predictor-corrector interior point method is put forward to optimize the allocation of different types of renewable distributed generator. Minimum network losses is used as the objective function, taking into account the operation possibility of distributed generator. In order to improve the efficiency of the problem, the model applies power flow equations and constraints in rectangular coordinates.By modeling and solving the problem mentioned above, we can not only realize the basic requirements of power system on security, economics, environmental protection, but also can reduce costs of power generation and transmission, and coordinate the conflict between power plants and power grids. Therefore, research on operation optimization models of power system considering distributed energy has important theoretical value and practical significance.更多还原