Robust multivariable tuning methods

Abstract

A large number of tuning methods have been developed for automatic tuning of Single-Input Single-Output (SISO) PID controllers. Many industrial processes are however multivariable. It is possible to design and use decoupling compensators to reduce the loop interactions and then use SISO methods to tune the PID controllers. This will however need to be repeated when the process operating condition changes. Furthermore, decoupling methods work well only for systems with weak loop interactions. Despite the wide popularity and extensive research for SISO tuning methods, the number of tuning methods for Multi-Input Multi-Output (MIMO) systems is limited. The MIMO control loops may also be tuned sequentially using scalar techniques. This is not often sufficient since tuning a loop may detune other loops in the system. Some sequential tuning methods are partially open loop and only applicable to open-loop stable system. Furthermore these methods are time consuming, expensive and not easily accepted by the plant operators. This chapter reviews and discusses some easy to use practical multivariable methods and demonstrates their application to an industrial example.