|
|||||
|
|
| 基于改进BP神经网络的建筑用电量预测 | |
| 引用本文： | 许慧,黄世泽,郭其一,屠瑜权.基于改进BP神经网络的建筑用电量预测[J].低压电器,2014(18):54-58. |
| 作者姓名： | 许慧 黄世泽 郭其一 屠瑜权 |
| 作者单位： | 1. 同济大学电子与信息工程学院,上海,200331 2. 同济大学交通运输工程学院,上海,201804 3. 浙江中凯科技股份有限公司,浙江乐清,325604 |
| 摘 要： | 在建立城市层面的建筑用电量预测模型时,对BP神经网络结构及其训练算法进行了研究。针对常规BP网络算法收敛速度慢、易陷入局部最小点的缺点,采用了具有较快收敛速度及稳定性的L-M算法进行了改进,建立了基于改进BP神经网络的建筑预测模型。最后通过上海市某栋公共建筑原始用电能耗统计数据作为样本对该预测模型进行了验证,验证结果表明,基于改进BP神经网络的预测模型适合建筑的用电量预测。 |
| 关 键 词： | 非线性 BP神经网络 建筑用电能耗 L-M算法 预测模型 |
Prediction of Building Electricity Consumption Based on Improved BP Neural Network |
|
| XU Hui,HUANG Shize,GUO Qiyi,TU Yuquan.Prediction of Building Electricity Consumption Based on Improved BP Neural Network[J].Low Voltage Apparatus,2014(18):54-58. | |
| Authors: | XU Hui HUANG Shize GUO Qiyi TU Yuquan |
| Affiliation: | XU Hui, HUANG Shize, GUO Qiyi, TU Yuquan (1. College of Electronics and Information Engineering, Tongji University, Shanghai 200331, China; 2. School of Transportation Engineering, Tongji University, Shanghai 201804, China; 3. Zhejiang JONK Technology Co. ,Ltd. , Yueqing 325604, China) |
| Abstract: | According to the complex nonlinear characteristics of the building energy consumption,the structure of BP neural networks and its training algorithm were studied. In the study,as the traditional BP algorithm has some unavoidable disadvantages such as the slow training speed and being easily plunged into local minimums,an optimized L-M algorithm was applied,which has a quicker training speed and better stability,to set up a predictive model of the building based on improved BP neural networks. With the statistical data of the public building electricity consumption in Shanghai as a sample,the prediction model was testified. The test result shows the predictive model based on improved BP neural networks is suited to the prediction of building electricity consumption. |
| Keywords: | nonlinear BP neural networks building electricity consumption L-M algorithm prediction model |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ | |