真空管太阳能热水器热性能逐时循环预测模型  

The hourly cycling forecasting model of thermal performance for an evacuated solar water heater

作　　者：李泽东[1] 高文峰[1] 刘滔[1] 林文贤[1] 赵佳音[2] 

机构地区：[1]云南师范大学太阳能研究所,教育部可再生能源材料先进技术与制备重点实验室,云南昆明650092 [2]云南师范大学物理与电子信息学院,云南昆明650092

出　　处：《可再生能源》2013年第12期11-16,共6页Renewable Energy Resources

基　　金：国家自然科学基金项目(11072211;51266016);高校博士点专项基金项目(20105303110001);云南省应用基础研究计划项目(2011FA017)

摘　　要：受地域、季节和气候的影响,在户外进行太阳热水器热性能测试工作很难按计划开展。文章通过建立符合热水器集热条件的神经网络对热水器进行试验,用多云天气条件下的测试数据训练网络,再把标准天气条件下获得的测试数据输入该网络中,显示出该网络具有较好的外推能力,其预测值与试验值吻合较好。文章建立了真空管热水器的人工神经网络循环预测模型,实现了对水温的连续预测,其平均相对百分误差为3.899%,可在不具备测试条件的情况下,对太阳热水器的性能进行预估。Due to various influences such as geographic locations, seasons, and climates, it is usually hard to carry out a thermal performance test of solar water heater under an all-weather outdoor condition. In this paper, an appropriate neural network model that complies with the requirements for the heat gain of the heater was founded. Experimental measurements have been made on a heater and the neural network is trained with the measured data obtained under cloudy climate conditions. The measured data obtained under standard clear climate conditions are then put in the trained neural network to produce predicted results that are found to be very close to the measured data, which indicates the network possesses good extrapolating ability. The artificial neural network is then established for evacuated-tube solar water heaters, which can be used to produce a continuous prediction of the water temperature. It is found that the predictions are excellent, with the deviations within 4% between the predicted data and the measured data for the heater, It is therefore concluded that the neural network can be used to accurately predict the thermal performance of a solar water heater by using the measured data obtained under the climate conditions that are deemed to fall short of the standard clear climate conditions required by the testing.

关 键 词：神经网络 物理模型 家用太阳热水器 循环预测模型 

分 类 号：TK515[动力工程及工程热物理—热能工程]