Development of a Digital Autotuning PI for First Order Plant Using RLS-PZC

The fact that a real plant can be estimated as a first order and its parameters vary due to the environment has motivated this article to discuss the development of a digital autotuning PI for a first-order plant using Recursive Least Square (RLS) and Pole Zero Cancellation (PZC). Although the focus is only on first order, the methods discussed here hopefully become a basis for developing higher-order plants. Firstly, formulas for calculating PI parameters are derived using PZC and tested by simulation to verify their effectiveness. Then it is organized serially with the RLS and digital PI to form an autotuning PI algorithm. The RLS periodically reads plant input-output to estimate plant parameters. These resulting parameters are fed to PZC and finally, PZC outputs are used by digital PI to control the plant. This design is verified by Matlab simulation, where the controller is realized as an m-function containing a program code for RLS, PZC, and digital PI algorithm. The test was conducted by varying plant parameters, including DC gain and time constant. Verifying controller parameters and their response shows that RLS-PZC can effectively re-tune the digital PI parameters, proved by its response having zero steady-state error and its settling time is maintained. The proposed algorithm can also ensure that the PI controller output is always within the specified maximum limits hence the actual response does not deviate from the designed response.