Research And Application Of BP Neural Network PID Controlalgorithm Based On Hybrid Optimizatio

In recent decades,great changes have taken place in the field of industrial control.The control object has evolved from a single object to a complex one.PID control is the earliest control strategy applied in process control.Its algorithm principle is easy to understand and its structure is clear.It achieves considerable effect on the control of a single object.However,more and more complex controlled objects are characterized by nonlinearity,Multi-time-varying parameters and difficulty in establishing mathematical models,therefore,the traditional PID control method can not achieve the desired control effect for complex objects.It is urgent to study an intelligent control strategy which does not depend on the model to solve the emerging problems.Neural networks also play a pivotal role in the development of automatic control.The neural network not only has its own strong self-learning ability,but also has the ability to express nonlinearity.It does not depend on the establishment of mathematical models,and can learn and adapt to the uncertain or time-varying characteristics of complex systems.Has good stability and robustness.BP learning algorithm is the most used algorithm in neural network training.It mainly corrects the network weight and threshold during the training process,so that the error is minimized.However,it has the disadvantages of slow convergence,dependence on initial weight,and easy local minimum.Therefore,the optimization of BP learning algorithm is imminent,so that it does not depend on the selection of network initial weights,thus improving the learning performance of BP algorithm.Combining the optimized BP neural network with PID control algorithm,a new PID control strategy is obtained.Aiming at a series of defects in BP neural network self-learning,this paper proposes an intelligent PID control strategy based on genetic and particle swarm hybrid optimization BP neural network.The intelligent PID control strategy optimizes the BP neural networkby using the strong global search ability of the genetic algorithm combined with the strong local search ability of the particle swarm optimization algorithm,effectively overcoming the slow convergence and easy trapping of the BP algorithm in the network training process.Defects in local minima make system performance more stable.In the aspect of experimental research,this paper uses MATLAB to simulate the intelligent PID control strategy,and realizes the greenhouse environment control system which uses the single chip as the core and Lab VIEW as the monitoring platform to apply this intelligent PID control algorithm.The research results show that the hybrid optimization BP neural network PID control strategy is applied to the greenhouse environmental control system,and the system has better dynamic characteristics and control effects.