大波动电网负荷短期预测的仿真研究

Simulation of Short Forecasting for Fluctuant Load of Power Grid

研究大波动电网负荷短期预测问题。由于电网负荷存在频率振荡和电压摆动,不能保障安全供电,应提前预测。针对传统数据预处理方面的缺陷,导致所需数据量大,预测精确度低。为解决上述问题,提出的数据挖掘融合改进马尔科夫链预测模型。首先对冲击负荷和小水电负荷的特性进行分析,运用改进形状系数对历史数据向量进行筛选,找出共线性的指标向量,并用矩阵实验室仿真计算出预测结果。应用实例证明,采用数据挖掘融合改进马尔科夫链的负荷模型预测更加准确地预测出冲击负荷与小水电发电负荷,为大波动负荷的预测提供了设计依据。

The probelm of short forecasting for the fluctuant load of power grid was studied. Because the tradition- al algorithm has the faults in data preprocessing, a large number of data are needed and the result is imprecise. In or- der to solve this problem, the data mining and improved - markov model were put forward. First, characteristics of the impact load and the out - put of small hydro - power plants were analyzed. Then, vectors which comprised histor- ical index data were screened by comparing the improved shape coefficients, and the linear index was acquired. Fi- nally, the result was obtained by using the the MATLAB. Example shows that the impact load and the out - put of small hydro - power plants are predicted accurately by using data mining and Improved - Markov model, so as to pro- vide the design basis for the fluctuate load.