Research Of Intelligent PID Control For Liquid Level System

The liquid level control equipment is controlled through common PID algorithm generally , the common PID algorithm has the advantages of simple structure,good stability, but it must have certain mathematics model and its feature of anti-disturbance is poor , besides it can not adjust the PID parameters on-line;So ,this paper chooses PSO-PID algorithm,RBF neural net-PID algorithm,fuzzy-PID algorithm and single neuron adaptive PID algorithm to control liquid level ,then the control has the adaptive and self-tuning advantages.The paper firstly introduces the software and hardware structure of liquid level system , then adopts four optimization algorithms such as PSO-PID algorithm to control liquid level system through simulation and real-time control. The simulation chooses Matlab6.5, and real-time control is under the circumstance of Labwindows. The PSO adjusts PID parameters through iterative optimization strategy based on bird swarm preying on food. The fuzzy adaptive algorithm chooses error and error variation to carry out fuzzy inference, then tunes the PID parameters; RBF NN tunes the PID parameters through identifying the Jacobian information of system ;The single neuron adaptive PID algorithm tunes control quantity through modifying weight coefficient on-line. in conclusion, every optimization algorithm which tunes PID parameters is based on some kind of optimization strategy; After every real-time control experiment, there is a discussion about improvement of the algorithm for further research in the future.The above-mentioned methods of liquid level control based on intelligence-PID not only have the advantages of good robustness, high response speed and low overshoot but also have the advantages of common PID such as simple structure and high stable precision, and realizes the self-tuning function of PID parameters. These intelligence-PID algorithms realized in the circumstance of Labwindows are applied in the liquid level control system, the real-time test results have shown the good effects of control of these intelligence optimization algorithms such as low overshoot,high stable precision; I believe after the improvement of these algorithms, their applicabilitys will be stronger and will be suitable for more complex control systems than liquid level control system.