PID control is composed of proportion regulation, integral regulation and differential regulation. The algorithm formula includes position digital PID control and incremental PID control. This is the basis of all regulating control.
With the continuous development of technology, intelligent PID control is gradually becoming the trend.
Intelligent PID control is the combination of intelligent control and traditional PID control, is self-adapting, its design idea is to use ExpertSystem, fuzzycontrol and neuralnetwork technologies, introduce artificial intelligence into the controller in a nonlinear control mode, so that the system can get better control performance than traditional PID control in any running state. It is independent of accurate mathematical model of the system and on-line automatic adjustment of controller parameters. It has good adaptability to the change of system parameters.
Intelligent PID control mainly includes fuzzy PID controller, expert PID controller and PID controller based on neural network.
Fuzzy PID control is to use the current control deviation, combine with the change of the dynamic characteristics of the controlled process, and practical experience against specific processes, according to certain control requirements or objective functions, ratiocinate by fuzzy rules, adjust the three parameters of PID controller online.
Expert system is a kind of calculator system which can solve the special task in a particular field at the level of human expert, it had a vast amount of expert-level knowledge and experience in a particular field, be able to use the knowledge and problem-solving methods of human experts to solve problems in the particular field. Expert PID control adopts regular PID control form, through identifying system error and system output, we can understand the change of dynamic characteristics of the controlled object, adjust three parameters of PID online until the response curve of the process is the optimal response curve. It is an adaptive technique based on heuristic rule reasoning, the purpose is to deal with the uncertainty which arises during the process.
Neural network system is also called artificial neural network, it’s a network of which connect artificial neurons in a certain way, to simulate the process of the neuronal activity in the human brain, realize information processing, storage, etc. Neural network has the structure of forward network, feedback network and so on. It is different from fuzzy PID control and expert PID control, the PID control based on neural network uses neural networks directly as controllers, instead of using neural network to set PID parameters, PID parameters can be adjusted indirectly by training the weight coefficient of the neural network.
Advantages of intelligent PID control
It absorbs the advantages of intelligent control and conventional PID control, first of all, it has the ability of self-learning, self-adaptation and self-organization, can automatically identify the controlled process parameters, automatically adjust the control parameters, and adapt to the changes of the controlled process parameters. Secondly, it has the characteristics of simple structure, strong robustness, high reliability and familiarity for the field designers.
In practical projects, it is recommended to use intelligent PID control, because it is simple and easy to implement, and general control can be automatically solved. The traditional PID system requires a wealth of practical experience, otherwise, it will take a lot of effort to implement successful adjustment.
It is suggested that the PID automatic loop should be put into the device as much as possible, so that the operator is relatively relaxed, otherwise it will be very tiring.