Auto-tuning Based On FFT Algorithm On The Practicality

PID controller because of simple structure, good stability, high reliability advantages, so process control is the most extensive application of the control method. In industrial process control, more than 95% control loop PID controller with the structure. However, due to changes in object parameters, the PID controller parameters will change is directly related to the effect of the control system, control system design is the core content, therefore control model of literacy as a key issue. Therefore, since the PID controller tuning parameters technology is a set adaptive control, intelligent control, automated process control as one of high-tech engineering of new technologies, is currently very hot research topic one.For decades, PID control parameter tuning methods and techniques are in a constant development, especially in recent years, at present, the many tuning method, in the actual industry based on the frequency characteristics of the PID controller tuning parameters method (SF Law) innovation is that it is based on the continuous spectrum Recognition, Application for FFT frequency conversion sequence in the frequency domain and then modeling. Although the algorithm has been in MATLAB simulation results, but there are still some algorithm inadequate, such as leading to the main error spectrum distortion factors - energy leakage problems; MATLAB simulation and poor repeatability unable to meet real-time requirements, and other issues.Then address the problem by R (refine) - SF method, the algorithm in FFT (Fast Fourier) before using high-pass filter to filter out the low frequency signal part of the signal changes relatively stable part of the low-frequency zero, then its truncation, on avoiding the energy leakage problems. At the same time ingenious use of the frequency characteristics of expression and molecular denominator division will be mixed in some of the same filter spectrum elimination. The repeatability of the problem is the differential control study found objects from the input signal with sharp peaks in their sampling at the "fence" effect caused by error.Based on the above basis, will be R (refine) - SF law applied to a number of objects, to verify this algorithm in various object of practicality. In real-time systems in the experimental results show that the algorithm was charged with various objects obtained the frequency characteristics similar to the real value can be applied to the need for frequency characteristics of object parameter tuning PID other methods. Therefore, this algorithm has a certain value and promotes the use of practicality.