| 摘要: |
| 【目的】 探索解决植物工厂内部温湿度环境控制中环境变量存在的非线性、强耦合等问题，实现植物工厂温湿度环境的自适应控制。【方法】 文章采用阶跃响应实验法建立植物工厂温湿度系统控制模型，采用前馈补偿解耦方法对植物工厂内部温湿度解耦；采用天牛须算法（Beetle Antennae Search algorithm，BAS）对径向基函数（Radial Basis Function，RBF）神经网络PID（Proportion Integration Differentiation）控制器的PID初值进行优化，使用优化后的RBF神经网络对PID参数整定，实现对温湿度控制。【结果】 采用BAS优化RBF神经网络PID的解耦控制系统，与单神经元PID解耦控制系统相比，温度和湿度曲线变化更平顺，超调量几乎为0，系统基本无震荡，温度和湿度的超调量分别降低了12.5%和5.5%，调节时间缩短了80%和84%。【结论】 经过BAS优化后的RBF神经网络进行PID参数整定的解耦控制器，超调量更小，温湿度到达稳定时间更短，不仅解开了植物工厂内部环境温度和湿度的耦合，还提高了系统的调控速度与精度。 |
| 关键词: 植物工厂 温湿度解耦控制 RBF神经网络PID算法 天牛须搜索算法 |
| DOI：10.12105/j.issn.1672-0423.20240202 |
| 分类号: |
| 基金项目:政府间国际科技创新合作重点专项“面向果蔬及其植物工厂的智慧管控系统及机器人装备研发”（2019YFE0197700） |
|
| Design of temperature and humidity decoupling control system in plant factory based on RBF neural network PID |
|
Xu Chenyi1, Wang Yulong1, Wang Yuchao1, Yu Fenghua1,2,3
|
|
1.College of Information and Electrical Engineering，Shenyang Agricultural University，Shenyang 110866，Liaoning，China;2.National Sub-centre for Regional Innovation in Digital Agriculture （North-East），Shenyang 110866，Liaoning，China;3.Liaoning Key Laboratory of Intelligent Agricultural Technology，Shenyang 110866，Liaoning，China
|
| Abstract: |
| Purpose To explore and solve the problems of nonlinearity and strong coupling of environmental variables in the control of temperature and humidity environment inside plant factories，and to realize the adaptive control of temperature and humidity environment in plant factories.Method The step response experimental method was used to establish the control model of the plant factory temperature and humidity system，and the feedforward compensation decoupling method was used to decouple the temperature and humidity inside the plant factory；the Beetle Antennae Search algorithm （BAS）was used to optimize the initial value of the PID controller （Proportion Integration Differentiation）of the Radial Basis Function （RBF）neural network，and the optimized RBF neural network was used to adjust the PID parameters to achieve the control of temperature and humidity.Result The decoupled control system using BAS optimized RBF neural network PID，compared with the single neuron PID decoupled control system，the temperature and humidity curves changed more smoothly，the overshoot amount was almost 0，and the system was basically free of oscillation，and the overshoot amount of the temperature and humidity was reduced by 12.5% and 5.5%，respectively，and the regulation time was shortened by 80% and 84%.Conclusion The decoupled controller with PID parameter tuning by RBF neural network optimized by BAS has smaller overshooting and shorter time for temperature and humidity to reach stabilization，which not only decouples the coupling of temperature and humidity of the internal environment of the plant factory，but also improves the regulation speed and accuracy of the system. |
| Key words: plant factory decoupled control RBF neural network PID algorithm tennyson search algorithm |
