基于径向基函数神经网络的温室室内温度预测模型

试验证实径向基函数神经网络(Radial Basias Function Neural Network)在函数逼近能力、训练速度方面都有良好的性能．采用最小正交二乘法为训练算法,基于传统的数学分析,用PRIVA公司温室监控系统采集数据,选用当前时刻室外温度、风速、太阳辐照度、顶窗开度、内帘幕展开度、水温、室内温度、相对湿度,再加上1个时间间隔、2个时间间隔以前的室内温度作为输入向量,获得了满意的温室室内温度一步预测模型(均方差等于0．0073)．该模型为设计温室环境控制器及分析温室性能奠定了良好基础．

Greenhouse Temperature Prediction Model Based on Radial Basias Function Neural Networks

The Radial Basias Function Neural Network has an excellent ability in function approximation and a higher speed in training,which is proved by experiment here.In this essay, a satisfactory temperature prediction model is set up by using means of orthogonal least squares learning algorithm and choosing right input vector,which are consisted of indoor temperatures one interval and two intervals ago,and the environment parameters at current time,including outdoor temperature,wind speed,radiation,opening degree of windows,opening ratio of sun- shade curtain,water temperature,indoor temperature,and indoor relative humidity.The data are collected by an environment control system made by PRIVA.The average square difference between the predicted temperature and the actual temperature is 0.0073.The result lays a solid foundation for designing the controller of greenhouse environment.