6 Steady and Transient State Detection

Steady-state (SS) models should only be adjusted to fit SS data. Similarly, coefficients in models that represent the rate of change of transition between SSs should only be adjusted with transient (dynamic) data. Nominally, it is trivial to distinguish SS and transient state (TS) data. If the value changes in time it is a TS; if the value is constant in time it is an SS. However, measurement noise confounds the ideal, making the determination of SS or TS a statistical, probable determination. This chapter presents a method to determine SS and TS. Further, the steady-state identification SSID approach presented here will be shown to be useful as a scale-independent, universal, convergence criterion in nonlinear regression.