Optimisation-based sampling time selection for modelling the three interacting tanks

This paper considers parameters and sampling time identification of a three interacting tanks system. A novel tailor-made model is introduced to deal with the real plant in the laboratory. Experimental data are then exploited to perform parameter identification via particle swarm optimisation (PSO) method. In addition to the system parameters, we propose to simultaneously optimize the sampling selection in the optimisation process. Even though several well known sampling time selection are established, a more precise selection is required. It is shown in this paper that a meta-heuristic optimisation method (PSO) can be an efficient procedure for this problem. Moreover, three different error measures for performing optimisation are also discussed.