基于物联网的光伏温室温度预测与环境监控系统  

Photovoltaic greenhouse temperature prediction and environmental monitoring system based on Internet of Things

作　　者：海涛[1] 招兴业 陆剑锋[2] 王钧 Hai Tao;Zhao Xingye;Lu Jianfeng;Wang Jun(College of Electrical Engineering,Guangxi University,Nanning,530004,China;China Tobacco GuangxiIndustrial Co.,Ltd.,Nanning,530004,China;Hualan Design Group Co.,Ltd.,Nanning,530004,China)

机构地区：[1]广西大学电气工程学院,南宁市530004 [2]广西中烟工业有限责任公司,南宁市530004 [3]华蓝设计(集团)有限公司,南宁市530004

出　　处：《中国农机化学报》2025年第2期75-82,共8页Journal of Chinese Agricultural Mechanization

基　　金：国家自然科学基金资助项目(52277138);广西科技计划项目(桂科AB22035037)。

摘　　要：针对大型温室群普遍存在耗能高、监测困难及温度调控滞后等问题,设计集光伏发电、低功耗广域物联网和长短期记忆网络预测技术于一体的温室监控系统。根据广西桂南地区的气候特征,通过Ecotect仿真得出屋顶光伏组件覆盖率在25%或33%时可兼顾光伏发电和温室内部采光效果。监控系统利用LoRa和NB—IoT技术混合组网实现环境参数的无线采集,上位机结合云平台及物联网技术对温室环境进行远程监控,并运用采集数据训练WOA—LSTM模型为温度预测提供支撑。测试表明,系统通信距离在500 m内,丢包率不超过3%,满足大型温室群对环境信息采集和稳定传输的需求,温度预测模型的均方根误差和平均绝对误差分别为0.476℃、0.367℃,可为温度预测和提前调控提供参考。该系统能够实现温室环境的实时监测、温度预测与调控,可为进一步提高温室种植作物的产量和质量提供借鉴。Aiming at the problems of high energy consumption,difficult environmental monitoring,and delayed temperature control in large-scale greenhouse clusters,a greenhouse monitoring system.integrating photovoltaic power generation,low-power wide-area Internet of Things,and long short-term memory network prediction technology is designed.Based on the climate characteristics of southern Guangxi,Ecotect simulations indicate that a roof photovoltaic panel coverage rate of 25%or 33%can balance both photovoltaic power generation and internal greenhouse lighting.The monitoring system uses LoRa and NB—IoT technology to realize wireless collection of environmental parameters.The host computer combines cloud platform and IoT technology to remotely monitor the interior of the greenhouse environment,and uses the collected data to train the WOA—LSTM model to provide support for temperature prediction.Test results show that the communication distance of the system is within 500 m with a packet loss rate of less than 3%,meeting the requirements of large-scale greenhouse clusters for environmental data collection and stable transmission.The root mean square error and mean absolute error of the temperature prediction model are 0.476℃and 0.367℃respectively,providing reference for temperature prediction and advance control.This system enables real-time monitoring,temperature prediction,and control of greenhouse environments,serving as a reference for improving the yield and quality of greenhouse crops.

关 键 词：光伏温室 监控系统 温度预测 低功耗广域物联网 无线传感网络 

分 类 号：S24[农业科学—农业电气化与自动化]