基于ARIMA和神经网络的电能质量稳态指标预测

根据有功功率与五项电能质量稳态指标的相关性以及有功功率的数据特点，提出了一种对电能质量稳态指标的预测方法。该方法利用ARIMA时间序列算法对有功功率进行了预测，并根据有功功率与五项电能质量稳态指标的相关性建立神经网络预测模型对五项常规指标进行预测。通过分析预测结果与真实值的误差可得平均误差均在20%以内，该方法可以有效预测出电能质量指标序列的变化趋势，从而对电力系统的稳定性、安全性和经济性起到很好的作用。

Predicting of Power Quality Steady Indicators Based on ARIMA and Neural Network

Based on the active power with five power quality indicators as well as the relevance of active power data characteristics,pro-pose a steady-state power quality indicators forecasting method. This method uses the ARIMA time series algorithm to predict the active power,and in accordance with the relevance of the active power with five steady-state power quality indicators,establish neural network model to predict the five conventional indicators. By analyzing the predicted and actual values of the error can be an average error of less than 20%,so the method can predict the sequence change trends of power quality,and thus playing a very good role for the power system stability,security and economy.