摘要
为了分析火电厂冷却塔修建时的沉降规律,本文以某实际工程的监测数据为例。首先构建时间序列(auto regressive moving average,ARMA)模型对冷却塔沉降进行预测,进而利用初次预测的数据和对应冷却塔修建时标高组成误差逆向(back propagation,BP)神经网络模型,得出修建过程的预测值。由预测值与实测数据对比分析表明,本文提出的组合模型能有效、准确地预测冷却塔的沉降变化。
In order to analyze the settlement law of thermal power plant during the construction of cooling tower,this paper takes the monitoring data of a practical project as an example.Firstly,a time series model(ARMA)is constructed to predict the cooling tower settlement,and then the BP neural network model is formed by using the initial prediction data and the corresponding elevation during the construction of the cooling tower to obtain the predicted value of the construction process.The comparison and analysis between the predicted value and the measured data show that the combined model proposed in this paper can effectively and accurately predict the settlement changes of cooling towers,and has a guiding effect on the settlement laws during the construction of similar cooling towers.
作者
温玉维
陈威
WEN Yu-wei;CHEN Wei(Hunan Electric Power Design Institute Co.,Ltd.of CEEC,Changsha 410007,China)
出处
《电力勘测设计》
2020年第S01期183-187,共5页
Electric Power Survey & Design
关键词
冷却塔
沉降监测
ARMA
BP神经网络
组合预测
cooling tower
subsidence monitoring
ARMA
BP neural network
combination forecast
