电力系统论文：基于混合算法—径向基神经网络短期负荷预测.doc

电力系统论文：基于混合算法—径向基神经网络的短期负荷预测 【中文摘要】短期负荷预测(STLF)是电力系统运行调度中的一项重要内容,是能量管理系统的一个重要模块。电力市场的引入,对负荷预测的准确性、实时性、可靠性和智能性提出了更高的要求。神经网络模型是一种比较常用的短期电力负荷预测模型,本文针对RBF神经网络的不足之处进行改进,建立新的预测模型,并应用于短期电力负荷预测。本文的主要研究内容如下：(1)粒子群(PSO)算法是一种基于群智能方法的演化计算技术,该算法操作简单,使用方便,收敛速度快。本文将PSO算法与RBF神经网络相结合,形成粒子群-径向基神经网络(PSO-RBF),建立计及各种影响因素的短期负荷预测模型。运用所建立的PSO-RBF模型和RBF模型进行短期负荷预测,并比较所得结果可知,PSO-RBF模型要优于传统的RBF方法。(2)将人工鱼群算法(AFSA)与RBF神经网络相结合,形成人工鱼群-径向基神经网络(AFSA-RBF),运用此模型进行短期负荷预测,并与RBF模型的结果进行比较,AFSA-RBF模型要优于传统的RBF模型。(3)针对PSO算法易陷入局部极值和收敛精度不高的缺点,将AFSA算法引入到PSO算法中,AFSA算法优点在于具有全局寻优的特点,能跳出局部极佳,结合两者的优点,组成混合优化算法,以四个标准函数进行测试,测试结果表明,混合算法在一定程度上避免陷入局部极小,加快收敛速度且提高了搜索精度。本文将AFSA算法与PSO算法组成的混合算法与RBF神经网络相结合,形成混合算法-径向基神经网络,并将其用于澳大利亚新南威尔士州短期负荷预测仿真研究中,仿真结果表明此方法比起PSO-RBF神经网络,更能够提高预测精度,能克服RBF神经网络和PSO算法的缺点。 【英文摘要】Short-term load forecasting (STLF) in power system is an important part of running and dispatching, and it is an important module of energy management system. The operation of power market requires the accuracy, real-time, reliability and intelligence of load forecasting. The neural networks forecasting model is one of the most normal models used in STLF. On base of improving the shortcomings of Neural Networks Forecasting using the RBF algorithm, this paper aims to introduce a new model in STLF. The main study in this paper is as follows.(1) Particle Swarm Optimization (PSO) is an evolutionary computation technique based on the swarm intelligence, which is originated from artificial life and evolutionary computation. In this paper, a mixed PSO-RBF algorithm is formed, which is the combination of PSO and RBF network. Then, a STLF model involving various influencing factors is built. The STLF of power system is performed using the mixed PSO-RBF algorithm and RBF algorithm. The simulation results indicate that this mixed PSO-RBF algorithm is better than RBF algorithm.(2) In this paper, a mixed artificial fish swarm algorithm (AFSA)-RBF algorithm is formed, which is the combination of AFSA and RBF network. Then, a STLF model inv