改进鲸鱼算法优化的多维度深度极限学习机短期负荷预测

Short-term Load Forecasting Based on Multi-dimensional Deep Extreme Learning Machine Optimized by Improved Whale Algorithm

电力系统短期负荷预测的准确性是影响电力系统运行安全的关键因素。以更精准进行短期负荷预测为目标,提出了一种基于改进鲸鱼算法(Improved Whale Optimization Algorithm,IWOA)优化的多维度深度极限学习机(Deep Extreme Learning Machine,DELM)短期负荷预测方法。首先,针对传统鲸鱼算法(Whale Optimization Algorithm,WOA)初始种群分布不够广泛的问题,引入Tent混沌映射对初始鲸鱼种群初始化;其次针对极限学习机模型(Extreme Learning Machine,ELM)数据深层隐藏的信息学习能力差的问题,采用深度极限学习机作为基础负荷预测模型,并以改进鲸鱼算法对其进行参数寻优;最后考虑到温度、湿度等因素对负荷变化影响较大,建立多维度IWOA-DELM负荷预测模型。仿真结果表明,与其他模型相比,多维度的IWOA-DELM模型预测的准确度更高。

The accuracy of power system short-term load forecasting is a key factor affecting the security of power system operation.In order to accurately carry out short-term load forecasting,a short-term load forecasting method based on multi-dimensional deep extreme learning machine(DELM)optimized by improved whale optimization algorithm(IWOA)was proposed.Firstly,Tent chaotic map was introduced to initialize the initial whale population to solve the problem that the initial population distribution of the traditional whale optimization algorithm(WOA)is not wide enough.Then,in order to solve the problem of poor learning ability of information hidden in deep data of ELM model,the DELM was used as the basic load forecasting model,and the IWOA was used to optimize its parameters.Finally,considering that temperature,humidity and other factors have a greater impact on load changes,a multi-dimensional IWOA-DELM load forecasting model was established.The simulation results show that,compared with other models,the multi-dimensional IWOA-DELM model has higher prediction accuracy.