基于CEEMD-IBA-LSSVM的微电网短期负荷预测研究与应用

Research and Application of Microgrid Short-term Load Power Prediction Based on CEEMD-IBA-LSSVM

为了有效获取微电网负荷中的隐藏信息和潜藏特征,进一步提升微电网短期负荷预测的精准度与效率,提出一种基于基于互补集成经验模态分解(CEEMD,complementary ensemble empirical mode decomposition)和改进蝙蝠算法(IBA,improvement bat algorithm)优化最小二乘支持向量机(LSSVM,least squares support vector machine)的微电网短期负荷预测模型,先利用CEEMD对负荷序列进行有效分解,减轻局部信息相互影响;再引入引入反向学习、动态自适应惯性权重与拉格朗日插值法等方法改进蝙蝠的全局搜索与局部寻优能力,克服标准蝙蝠算法易早熟、易陷入局部最优值的问题,并利用IBA对LSSVM参数进行优化;最后通过算例验证CEEMD-IBA-LSSVM短期负荷模型效果,结果表明所构建模型的预测准确率约为98.21%,与其他预测模型相比具有较高的运行效率与预测精度。

In order to effectively obtain the hidden information and characteristics in the microgrid load and further improve the accuracy and efficiency of the forecasting in the microgrid short-term load, on the basis of complementary integrated empirical mode decomposition(CEEMD) and improved bat algorithm(IBA) and optimized least square support vector machine(LSSVM), a short-term load forecasting model of the microgrid is proposed. Firstly, CEEMD is used to effectively decompose the load sequence and reduce the mutual influence of local information;Then, the methods such as back learning, dynamic adaptive inertia weight and Lagrange interpolation are introduced to improve the global search and local optimization ability of bats, overcome the problem that standard bat algorithm is prone to premature and fall into local optimal value, and optimize LSSVM parameters by IBA;Finally, an example is given to verify the effect of CEEMD-IBA-LSSVM short-term load model. The results show that the prediction accuracy of the model is about 98.21%, which has higher operation efficiency and prediction accuracy than other prediction models.