Self-tuning through use of inherent conflicts

Abstract

Analysis of a priori design information often justifies the use of self-tuning of simple, application-specific controller structures. This justification arises from indications that good performance with respect to a primary performance index can be maintained through the slight adjustment of only a few controller parameters. The optimality of this adjustment is often dictated by the level of degraded performance of the system with respect to measurable, secondary performance indices. These multiple, countermanding indices realize a controller tuning conflict with respect to the self-tuning controller parameters. The author describes how a priori knowledge of such an inherent conflict can be used in the design of a sensitivity-based self-tuning mechanism. The proposed design methodology begins with descriptions of the concept of inherent control conflicts, sensitivity-points techniques, and application of integral manifolds and averaging to self-tuning control. The preliminary design tasks entailed in the checking for feasibility of self-tuning, which includes consideration of the input and disturbance spectra are described. Averaged integral manifold computations are then utilized in the design of the pseudosensitivity estimator. An example of the design of a power system stabilizer of a single-machine-to-infinite-bus power system illustrates the application of the design technique.<>