设施温室物联网智能测控系统研究  

Research on Intelligent Measurement and Control System of Greenhouse Internet of Things

作　　者：章子文 梁思程 张洪奇 吴勇 张大磊 柳平增[1,2,3] ZHANG Zi-wen;LIANG Si-cheng;ZHANG Hong-qi;WU Yong;ZHANG Da-lei;LIU Ping-zeng(School of Information Science and Engineering/Shandong Agricultural University,Tai'an,271018,China;Key Laboratory of Huanghuaihai Smart Agricultural Technology,Ministry of Agriculture and Rural Affairs,Tai'an,271018,China;Agricultural Big Data Research Center of Shandong Agricultural University,Tai'an,271018,China;Taishan College of Science and Technology,Tai'an,271038,China;Shandong Yong guan Agricultural Science and Technology Development,Heze,274900,China)

机构地区：[1]山东农业大学信息科学与工程学院,山东泰安271018 [2]农业农村部黄淮海智慧农业技术重点实验室,山东泰安271018 [3]山东农业大学农业大数据研究中心,山东泰安271018 [4]泰山科技学院,山东泰安271038 [5]山东勇冠农业科技发展有限公司,山东菏泽274900

出　　处：《山东农业大学学报（自然科学版）》2024年第4期633-643,共11页Journal of Shandong Agricultural University：Natural Science Edition

基　　金：山东省科技特派员项目(2020KJTPY078);中央引导地方科技发展专项资金项目(YDZX2022073)。

摘　　要：针对温室环境调控中手动控制和阈值条件调控存在调控精度低、参数超调严重等问题,以设施番茄温室为例,研究提出一种基于环境预测的精准调控策略。首先,构建专用物联网“六域”架构,采用MSP430F5438A作为主控模块,设计感知、通信等功能模块。其次,构建SSA-LSTM预测模型实现对温室环境的精准预测,并根据模型预测结果确定环境调控策略,通过PSO-PID控制模型实现对温室风口电机的精准控制。实验结果表明,相较于传统LSTM模型,SSA-LSTM预测模型的MAE降低58.52%,MAPE降低61.68%,RMSE降低63.84%。同时,相较于传统PID模型,PSO-PID控制模型的超调量降低89.99%,调节时间降低59.85%。系统经过实地部署验证,在保持种植品种和农事管理操作一致的情形下,智能调控的温室产量相较于传统温室提升约8.5%,证明了系统的有效实用性。Aiming at the problems of low regulation accuracy and serious parameter overshooting in manual control and threshold control in greenhouse environment regulation,a precise control strategy based on environmental prediction was proposed by taking facility tomato greenhouses as an example.Firstly,the"six-domain"architecture of the dedicated Internet of Things was constructed,and the MSP430F5438A was used as the main control module,and the functional modules such as perception and communication were designed.Secondly,the SSA-LSTM prediction model was constructed to realize accurate predictions of the greenhouse environment,and the environment control strategy was determined according to the model prediction results,and the accurate control of the greenhouse air outlet motor was achieved through the PSO-PID control model.The experimental results indicated that compared with the traditional LSTM model,the MAE of the SSA-LSTM model was reduced by 58.52%,the MAPE was reduced by 61.68%,and the RMSE was reduced by 63.84%.Meanwhile,compared with the traditional PID model,the overshooting amount of the PSO-PID model was reduced by 89.99%,and the regulation time was reduced by 59.85%.The system has been verified through field deployment,and under the premise of maintaining the consistency of planting varieties and agricultural management operations,the yield of the intelligently regulated greenhouse has been increased by about 8.5%compared with the traditional greenhouse,which proves the effective practicality of the system.

关 键 词：温度预测 控制模型 农业物联网 温室环境 

分 类 号：S625.51[农业科学—园艺学]