不确定性中长期电力需求预测方法探讨

Study on uncertainty mid-long term power demand forecasting method

鉴于经济发展及国家政策制定实施中存在大量不确定因素,引入模糊理论构建了模糊马尔科夫链状预测模型。通过对用电量观测数据进行模糊划分,解决了马氏理论中状态难以精确划分的难题,形成了用电量增长率的模糊状态隶属函数,并利用转移概率修正值计算出待预测时刻变量对各模糊状态的隶属程度,实现了对未来电力需求增长规律的预测分析。结果表明,本模型不仅能够提供预测分布区间,同时还保证了较高的预测准确度,是一种值得深入探讨的中长期电力需求预测方法。

Due to numerous uncertainty factors existing in economic development and state policies, it introduces fuzzy theory and constructs the fuzzy Markov chain prediction model（FMCM）. The difficulty within the accurate delimitation of classic Markov process（CMP） is settled through fuzzy dividing of electricity data observation, then fuzzy membership functions of power demand growth rate are established. The future power demand growth rule forecasting is achieved by using the membership degree of each forecasting time variable to each fuzzy state calculated by transition probability revised values. The results show that the model can not only provide prediction intervals, but also ensure higher prediction accuracy to improve that it is a remarkable mid-long power demand forecasting method.