技术中心

【英文篇名】 Air conditioning load prediction using neural networks based on wavelet transform 【作者中文名】 曹双华; 曹家枞; 李涛; 沈晓青; 【作者英文名】 By Cao Shuanghua~★; Cao Jiacong; Li Tao and Shen Xiaoqing★Donghua University; Shanghai; China; 【作者单位】 东华大学; 北京捷通机房工程设备有限公司; 南通四建集团有限公司; 【文献出处】 暖通空调, Hv 神经网络; 空调负荷预测; 递归BP网络; 【英文关键词】 wavelet transform; neural network; air conditioning load prediction; recurrent BP network; 【摘要】 基于小波变换的思想建立了递归BP网络模型来预测空调负荷,改进了网络权值、阈值的修改算法,引入了折扣系数法以提高近期预测精度,结合一实例进行了空调逐时冷负荷预测,结果表明该方法预测精度高,适用于空调负荷预测。 【英文摘要】 Based on the idea of wavelet transform, establishes the recurrent BP network model to predict air conditioning load. Adopts the discount coefficient method in the modification algorithm of the network weights and thresholds to improve the closest predictive precision. Predicts the hourly air conditioning load with a real example. The results indicate that the method has a high precision and is suitable to predict air conditioning load.