基于GA-BP神经网络的土壤-空气换热器换热量预测分析

Heat exchange quantity prediction of earth-air heat exchanger based on GA-BP neural network

文章对位于太原市一个日光温室内的土壤-空气换热器进行夏季工况试验,获得了不同运行工况下换热管内空气的温度和湿度的分布数据。试验结果表明:土壤-空气换热器具有一定的除湿效果;当换热管长度为17.2 m,换热管内空气流速为2 m/s时,土壤-空气换热器潜热换热量占全热换热量的31.37%,且潜热换热量在全热换热量中的占比随着换热管长度的增加而逐渐降低。文章将得到的试验数据分为训练样本和测试样本,同时,分别基于BP神经网络和GA-BP神经网络建立了土壤-空气换热器换热量的预测模型,并对模拟结果进行对比。模拟结果表明:GA算法对BP神经网络具有较好的优化作用;与基于BP神经网络建立的土壤-空气换热器换热量预测模型相比,基于GA-BP神经网络建立的土壤-空气换热器换热量预测模型的预测精度较高,收敛所需的迭代次数也较少。

An experimental study on a earth-air heat exchanger was carried out under summer operating conditions in a solar greenhouse in Taiyuan.The distribution data for temperature and humidity in the heat exchanger tube were obtained under different operating conditions.The experimental results show that the earth-air heat exchanger had a dehumidification effect on the high temperature and high humidity air of the solar greenhouse.When the length of the tube was 17.2 m,the air velocity and the ratio of latent heat were 2 m/s,31.37%respectively,and the proportion of latent heat exchange quantity in total heat exchange quantity decreases gradually with the increase of tube length.Meanwhile,the experimental data were divided into training samples and testing samples,and the prediction model of total heat transfer for earth-air heat exchanger was established by BP neural network and GA-BP neural network.The results show that GA algorithm had a good optimization effect on BP neural network prediction model.The prediction accuracy of GA-BP neural network was higher than BP neural network.However the iteration convergence number of GA-BP neural network was less than the latter.