Development of a fuzzy model for the control of a pump unit with the application of a coloured petri net

Abstract

The article is devoted to fuzzy control modelling based on Petri nets (PN). The aim of the study was to develop a Petri net-based control model with fuzzy logic for information expressed in a linguistic form. The criteria of the water pump’s operation, depending on the changing water consumption, were used to determine various situations and events in the system. Linguistic variables such as “water flow rate” and “pump speed” were used to describe incomplete knowledge about the system’s behaviour. The terms of these variables correspond to their fuzzy values and are denoted with expressions characterizing one of the system’s states. The fuzzification of linguistic variables was implemented in the Fuzzy Toolbox environment of the MATLAB modelling system. A system of rules for the pump unit’s control was developed by means of describing the required behaviour of the system by the relationship between situations and events using the logic «If ... Then ...». All sorts of situations, events, and relations between them formed sets of PN positions, transitions, and arcs. An algorithm for the control of the pump unit was developed with the consideration of the system of the production rules of control and PN structural elements. Based on the developed control algorithm, the functions of the input and output transition incidents were defined and shown in tables. The tables define the matrices of the input and output transition incidents. The PN graph model was developed. The model describes the operation of one pump unit. The visualization of the model was implemented in the CPN Tools system (Coloured Petri Nets Tools). The values of the terms are taken as attributes of the colours of the coloured Petri net (CPN) and by using CPN ML (Coloured Petri Nets Tools Markup Language) are assigned to the network markers. The values of the terms were used to describe the behaviour and the desired reaction of the system. Simulation experiments referring to the system’s situations and model analysis were carried out.