基于BP神经网络的压缩空气用能预测模型研究

Research on Compressed Air Prediction Model Based on BP Neural Network

文中根据BP神经网络对复杂问题所具有的预测能力,选取某企业压缩空气用能数据作为预测评判指标,建立了用于压缩空气用能预测的BP神经网络模型,并应用于实际问题预测。然而,最初预测结果在几个时刻点出现异常,在此文中针对记录数据可能出现问题的各种原因,再次利用局部预测对原始数据做出修正,最后预测结果已基本符合。实例表明:用BP神经网络模型进行压缩空气用能预测是可行和有效的。

In this paper, based on the BP neural network has the ability to predict about complex issues, the data of some enterprise com- pressed air is chosen as the indexes of criterion, a new model using BP neural network is proposed to predict the application of some en- terprise compressed air. However,in the first time the predicted results of some points are abnormal, according to the recorded data may be various reasons for the problems, use local forecast to make corrections on the original data, finally basically conform to the forecast re- sults. The results show that it is feasible and valid to use BP neural network for compressed air prediction.