Research On Fractional Phase Lead Iterative Learning Control/Repetitive Control Algorithms And Application

Phase lead iterative learning control and repetitive control are both learning control with highaccuracy. Both have a common key component, the phase lead part. Limited by the digital sampling,the traditional phase lead step should be integer, which leads to the bad performance, even theinstability of the control system. This thesis enables the phase lead step can be fractional, and realizesthe fractional phase lead with Lagrange interpolation, which improve the performance of Phase leaditerative learning control and repetitive control with new meanings.The phase lead part of iterative learning control is used to compensate the phase lag of system toget larger learnable bands and higher accuracy. But the lead step of iterative learning control cannotbe tuned continuously to achieve the best learning effect.The proposed fractional linear phase lead iterative learning control enlarges the range of lead step,and good for the optimization of the control system. The design condition of lead step and learninggain can be got according to the convergence analysis in the frequency domain. And the learnablebands can be tuned by lead step to achieve high tracking accuracy. Then the simulation of a robot armjoint is presented to do the comparison of integer lead step and fractional lead step is presented. Theresults reveals that, compared with integer lead step, fractional linear phase lead iterative learningcontrol favors wider learnable bands and higher tracking accuracy.Although the traditional repetitive control can deal with the harmonic problem of the inverteroutput, but the system tends to be unstable under low sampling application. While in fact theswitching frequency of power devices, which is sampling rate, is usually chosen to be low in highpower inverter application. So the application of traditional repetitive control meets with problem.In this paper, the phase lead part of repetitive is improved into fractional phase lead. This methodenables the phase lead step can be fractional, thus enlarges the stability range of system, and reducesthe tracking accuracy as well. The detailed parameters design of notch filter, low-pass filter and phaselead is given according to the frequency characteristics. At last, simulation based onSimPowerSystems environment and experiment based on Quanser semi-physical is built. The resultsof simulation and experiment verify that fractional phase lead repetitive control favors better stabilityand harmonic suppression.