On-Line Identification of a Variable Sampling Period Discrete Model and its Use for the Adaptive Control of Anaerobic Digestion

The concept of using cumulative amount of a quantity generated or consumed or fed, instead of time, as the independent dynamic model variable in sampled data process control or optimization algorithms is introduced. The sampling period becomes automatically varying, decreasing as the process dynamics speed up. The concept is illustrated with the adaptive control of an anaerobic digester, where the independent variable is taken to be the cumulative amount of methane produced. (The dependent variables are the dilution rate as manipulated variable and the methane production rate as measured output.) The algorithm also employs a modified least squares estimator, which excludes a certain number of the most recent data from forgetting.