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

Abstract: 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.

Key words: photovoltaic greenhouse, monitoring system, temperature prediction, low?power wide?area network, wireless sensor network