Temperature And Humidity Control Of Greenhouse Based On Particle Swarm BP Neural Network PID Control Algorithm

China is the country with the largest greenhouse area in the world.With the continuous development of China’s agricultural economy and the continuous expansion of the greenhouse area used for cultivating crops,the importance of improving greenhouse control algorithms has been highlighted.The time when foreign countries developed the technology to control the greenhouse environment is relatively earlier,with more advanced greenhouse hardware foundation,integrated control system and control algorithm,which makes the cultivation efficiency of foreign greenhouses very high.Compared with foreign control technology,the domestic research started late and the technical level reached is not high enough.One reason for the failure to improve control performance is the lack of a combination of intelligent algorithms and control methods.PID control is widely used in greenhouse control,but due to the fixed internal parameters of conventional PID controllers,it is difficult to obtain satisfactory performance in all equipment in a greenhouse with a wide range of dynamic characteristics.Therefore,this paper used a BP neural network with adaptive and learning capabilities to optimize PID control to achieve real-time tuning of internal parameters,and used a particle swarm optimization algorithm that can globally optimize to solve the shortcomings of the BP neural network itself that is easy to fall into local optimization.This paper effectively used the three algorithms to complement the advantages and disadvantages to better control the complex nonlinear system of the greenhouse with the main work includes:The first work was the establishment of temperature and humidity control experimental greenhouses.Firstly,PLC and HMI were introduced and selected,and PLC control program design and HMI interface design were carried out.Next,a distribution box was designed,integrating PLC and HMI,making it a control center for the greenhouse.Then the temperature and humidity control devices and sensors were introduced and selected.Finally,the above work was summarized and the overall design and construction of the greenhouse were carried out.The second work was the temperature and humidity sensor accuracy compensation.Firstly,the principle and training process of BP neural network were analyzed.Then the temperature and humidity data needed for accuracy compensation were collected.Then the temperature and humidity data were used to train the BP neural network.Finally,the trained BP neural network was used to compensate the accuracy of the temperature sensor and humidity sensor,and the effect was compared with the traditional linear algorithm.By comparing with the experimental results of the traditional linear algorithm,it can be concluded that after using the BP neural network to compensate the temperature and humidity sensor,the absolute value of the maximum relative error of the temperature sensor was reduced by 4.65%,the average value of the absolute value of the relative error was reduced by 1.32 %,and the average value of the absolute value of absolute error was decreased by43.1%.The absolute value of the maximum relative error of the humidity sensor was reduced by 0.21%,the average value of the absolute value of the relative error was reduced by 1.05%,and the average value of the absolute value of the absolute error was reduced by 59.5%.The third work was the research on PID temperature and humidity control of particle swarm BP neural network.Firstly,the principle and process of particle swarm optimization algorithm and PID control were analyzed.Then the method of particle swarm optimization of BP neural network weights and thresholds and the method of BP neural network optimization of PID control parameters were studied.Finally,the new algorithm was used in the greenhouse system designed in this paper,and compared with the traditional PID control effect.By comparing with the experimental results of the traditional PID control algorithm,it can be concluded that the temperature control rise time of the particle swarm optimization BP neural network PID control algorithm designed in this paper was shortened by 16.7%,the overshoot was reduced by 33.3%,and the maximum absolute error was reduced by 37.5%.The rise time of the humidity control was shortened by 11.9%,the overshoot was reduced by16.4%,and the maximum absolute error was reduced by 33.4%.The purpose of optimizing the temperature and humidity control performance of the greenhouse was achieved.