基于STM32H750的连栋温室环境协同智能监控系统设计与试验

Design and Experiment of Collaborative Intelligent Environmental Monitoring System for Multi-span Greenhouse Based on STM32H750

摘要:

针对我国大型连栋温室环境感知精度低、控制响应滞后、通信可靠性差、系统集成度弱及运行成本高等问题,本文以北京市通州区某22000m2 Venlo型玻璃连栋温室为试验平台,设计了基于STM32H750的边缘智能控制器,构建感知控制传输应用四层物联网架构,研发多维度感知、分区联动、混合通信自愈、云边协同管控的环境智能协同监控系统。系统采用硬件模块化光电隔离、软件冗余容错、温湿度耦合解耦、分层协同模糊PID控制、分区能耗优化、断网续传与链路自适应切换等关键技术,解决高温高湿强电磁干扰下长期稳定运行难题。72h连续测试与全生育期对比试验表明:系统数据采集周期1s,控制响应时间(0.7±0.2)s,设备执行滞后4.2~8.5s,通信丢包率小于等于0.1% ,温度控制稳态误差在±0.4℃范围内,相对湿度控制误差在±2.7%范围内,昼夜工况下温度波动控制在±0.5℃范围内、相对湿度波动控制在±3%范围内;与传统方案相比,能耗降低18.6% ,人工运维成本下降62.3% ,番茄单产提高12.7% ,病害发生率降低41.3% (P<0.01),静态投资回收期1.7a。研究结果可为大型连栋温室智能化生产提供技术支撑。

Abstract:

Aiming at the problems of low environmental perception accuracy, lagged control response, poor communication reliability, weak system integration and high operation cost in large-scale multi-span greenhouses in China, taking a 22000m2 Venlo-type glass multi-span greenhouse in Tongzhou District, Beijing as the experimental platform, an edge intelligent controller was designed based on STM32H750. A four-layer IoT architecture, including perception, control, transmission, and application, was constructed, and an intelligent collaborative environmental monitoring system was developed with multidimensional perception, zonal linkage, hybrid communication self-healing and cloud-edge collaborative management. The system adopted key technologies such as hardware modular photoelectric isolation, software redundancy and fault tolerance, temperature-humidity coupling decoupling, hierarchical collaborative fuzzy PID control, zonal energy consumption optimization, offline data transmission and adaptive link switching, which solved the problem of long-term stable operation under high temperature, high humidity and strong electromagnetic interference. The 72-hour continuous test and full growth period experiment showed that the data acquisition cycle was 1s, the control response time was (0.7±0.2)s, the equipment execution lag was 4.2~8.5s, the communication packet loss rate was no more than 0.1% , the steady-state error of temperature control was controlled within a ±0.4℃, and the control error of relative humidity was controlled within a ±2.7%. The temperature fluctuation was controlled within a ±0.5℃ and humidity fluctuation was controlled within a ±3% under day-night conditions. Compared with the traditional scheme, the energy consumption was reduced by 18.6% , the manual operation and maintenance cost was decreased by 62.3% , the tomato yield was increased by 12.7% , and the disease incidence was reduced by 41.3% (P<0.01), with a static payback period of 1.7 years. The research results could provide technical support for intelligent production in large-scale multi-span greenhouses.