基于Elman神经网络的温室环境因子预测方法

针对目前温室环境系统中,环境监测数据只能反映当前环境状况,无法预测温室环境变化趋势,导致温室环境控制效果差的问题,提出一种基于Elman神经网络的温室环境因子预测方法。以采集的温室内温度、湿度以及二氧化碳浓度的历史数据作为预测模型的输入,建立Elman神经网络预测模型,进而实现精确的温室环境因子变化预测。结果表明,Elman模型优于BP和RBF模型,温度、湿度和二氧化碳浓度预测结果的均方误差分别为0.003 9、0.005 9和0.028 3,决定系数分别为0.991 5、0.967 8和0.973 9。该模型预测结果较理想,可以为温室环境调控提供一定的决策支持。

Prediction method of greenhouse environmental factors based on Elman neural network

In the current greenhouse environment system, the environmental monitoring data can only reflect the current environmental conditions and cannot predict the changing trend of the greenhouse environment, resulting in poor greenhouse environment control effects. This paper proposed a greenhouse environment factor prediction method based on Elman neural network. With the collected historical data of temperature, humidity, and carbon dioxide concentration in the greenhouse as the input of the prediction model, an Elman neural network prediction model was established to realize accurate prediction of changes in greenhouse environmental factors. The results showed that the Elman model outperformed the BP and RBF models. The mean square errors of the temperature, humidity, and carbon dioxide concentration prediction results were 0.003 9, 0.005 9, and 0.028 3, respectively, and the coefficients of determination were 0.991 5, 0.967 8, and 0.973 9, respectively. The prediction results of this model were ideal, and it could provide certain decision support for greenhouse environment regulation.