Research Of Controller Parameters Self-Tuning Method

This thesis firstly introduces the basic principles and methods of self-tuning of PID controller parameters. Then several effective self-tuning methods are discussed in view of single variable and multivariable systems, next responding simulations and research are conducted. The following several aspects are included specifically.Firstly, an advanced genetic algorithm is given to have PID controller parameters tuned recursively and optimally, which is obtained by combining several advanced algorithms. The ways of improving standard genetic algorithms include making heredity operation on genetic algorithm chromosome with advanced crossover operator and mutation operator, proposing strategies of adjusting the searching space dynamically and applying precocity judgment strategy based on variance factor and the remaining strategy of optimal and adaptive value. This method requests little prior knowledge of unknown objects, is capable of parameter tuning in a very wide range and can get the result of anticipated response by many times of self-tuning.Secondly, a digital IMC method is extrapolated based on the discrete math model of object identification. This method is suitable for system controlling of first-order and second-order systems with big time lags.