基于迁移模糊系统的短期电力负荷预测建模  

Short-Term Power Load Forecasting Modeling Based on Transfer Fuzzy System

作　　者：李秋琰 

机构地区：[1]东华理工大学理学院,江西 南昌

出　　处：《应用数学进展》2024年第4期1671-1689,共19页Advances in Applied Mathematics

摘　　要：针对电力负荷数据缺失导致预测精度降低的问题,本文提出基于TSK迁移模糊系统(TSK-TFS)结合变分模态分解(VMD)、迁移成分分析(TCA)和改进斑马优化算法(IZOA)的短期电力负荷预测模型(IZOA-VMD-TSK-TFS-TCA)。首先利用VMD将电力负荷数据分解为若干子序列,并利用TCA将与电力负荷相关的因素降维;其次对斑马优化算法进行改进,利用改进后的斑马优化算法(IZOA)对TSK-TFS的参数寻优,并利用减法聚类算法得到聚类个数,把源域中的数据输入TSK模糊系统训练得到前件参数和后件参数并保留,继承参数并利用一部分目标域数据训练得到后件参数;最后根据得到的后件参数并经过计算得到测试集(另一部分目标域数据)若干子序列的预测值,将各个子序列的预测值叠加得到短期电力负荷的预测值。仿真实验结果表明,本文提出的IZOA-VMD-TSK-TFS-TCA短期电力负荷预测模型具有较高的预测精度,经过统计检验也证实了该模型具有较优的预测性能。To address the issue of reduced prediction accuracy caused by missing power load data, this paper proposes a short-term power load prediction model (IZOA-VMD-TSK-TFS-TCA) based on TSK Transfer Fuzzy System (TSK-TFS) combined with Variational Mode Decomposition (VMD), Transfer Component Analysis (TCA) and Improved Zebra Optimization Algorithm (IZOA). Firstly, VMD is used to decompose the power load data into several subsequences and TCA is used to reduce the dimensionality of factors related to power load. Secondly, IZOA is used to optimize the parameters of TSK-TFS and Subtractive Clustering Algorithm is used to obtain the number of clusters. The data in the source domain is input into the TSK fuzzy system to obtain the predecessor parameters and successor parameters and retain them. The parameters are inherited and the successor parameters are obtained using some of the target domain data. Finally, the predicted values of several subsequences of the test set (another portion of the target domain data) are obtained according to the obtained parameters of the subsequent components and calculated, and the predicted values of each subsequence are superimposed to obtain the predicted values of short-term power load. The experimental results show that the IZOA-VMD-TSK-TFS-TCA short-term power load prediction model has higher prediction accuracy. Statistical tests also confirmed that the model had superior predictive performance.

关 键 词：短期电力负荷预测 TSK模糊系统 迁移学习 变分模态分解 迁移成分分析 改进斑马优化算法 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]