微型植物工厂内部环境调控试验平台研制及试验

为解决植物工厂研究中存在的控制对象模型缺失问题,该文介绍了一种微型植物工厂内部环境调控试验平台。该平台舍弃了常用的参数设定间接调控模式,赋予控制算法对加热、降温等执行器的开关控制权,从而为更加直接地观测植物工厂内部环境的控制响应创造了条件。实际测试表明,该试验平台的加热、降温调控功能运行良好。同时,初步的数据分析确定了系统的延迟特性和平台内部温湿度因子间的相关性特征,为采样周期的选择以及后续开展温湿度智能优化调控奠定了良好的基础。

Development and test on experimental platform for inner environmental control of micro plant factory

Abstract: As a promising agricultural production mode, plant factory has been a hot topic for the last decades. Though a large quantity of efforts have been made on the research on the models of plant factory's inner environment, few attentions have been paid on developing control object models from the prospective of control theory. To solve this model-lacked problem, this paper proposes a novel experimental platform in the form of the micro plant factory, hoping to bring a powerful research tool for control object modeling. In the platform, control algorithms are endowed with direct control authorities for the actuators, such as the heating and cooling equipment. That is, control algorithms can influence the object parameters (micro plant factory's inner temperature and humidity for example), by directly opening or closing the actuators other than indirectly sending expected parameter values to the actuators' onboard controllers. In such a manner, the platform can present the control responses of the inner environment more directly than other existing plant factories, which will be a great advantage in control object modeling. Also, the platform has a more convenient interface for the application of intelligent control algorithms. Among the dominant inner environmental parameters, temperature and humidity are chosen as the main control objects whose model is to be analyzed, considering their control response features and monitoring cost. Aiming at the control and monitoring for the inner temperature and humidity, the control components, sensing components and operating components of the platform are chosen carefully. A computer and a controller are selected as control components. Control algorithms will run in the computer and generate control orders. Then the orders will be sent to the actuators through the controller. The temperature and humidity sensors are chosen as the sensing components, which act as real-time monitors for the control objects. The electro-thermal membrane, the refrigeration system, the lamps and the fans are selected as operating components. The electro-thermal membrane can give a uniform heating performance for the inner environment of the plant factory. The refrigeration system is used to lower the temperature. The lamps are mainly used for artificial lighting, which may lead to trifling heating effect. The fans are mainly used as air exchange tools, which may lead to obscure cooling results. All the components are integrated into a whole system under proper hardware and software design. The platform's planting experiments employing lettuce as the test crop were conducted in the lab from April to June, 2014. During the experiments, the feasibility of the platform in inner environmental control operations, such as heating and cooling, is verified. By analyzing the temperature and humidity changing curves along the time axis, the delay time of the system is determined to be 30 seconds, which helps to carry out the proper sample duration of 3 seconds for further precise control. Also, the correlation of the inner temperature and humidity factors is obtained using the Pearson correlation factors. All the results above will lay a stable foundation for the application of intelligent control algorithms on the platform in future. In the next step, we hope the real control object model for temperature and humidity can be established through a system identification algorithm.