Tuning and design of single‐input, single‐output control systems for parametric uncertainty

Frequency domain methods, which are for the design and tuning of control systems for processes with uncertain parameters, and are easy to use and interpret, are proposed. The advantages of the proposed methods over other H∞ methods are mainly that they require substantially less expertise on the part of the designer, and that they result in control system performance that closely approximates that which was desired. Several examples illustrate the proposed methods. A stable second-order lag and dead time process, with gain and dead time uncertainties, provides the vehicle for studying controller designs for single and two degree of freedom control systems. Application of Mp tuning to an uncertain, unstable, first-order lag, plus dead time process, shows that there can be a significant performance penalty associated with using a simple feedback control system rather than an internally stable model-based control system. The performance of a controller for an integrating, oscillatory process obtained by Mp synthesis also compares favorably with that of a controller obtained by other investigators using loop shaping.