Modeling and simulation of a fuzzy supervisory controller for an industrial boiler

In this paper we compare and discuss the performance of a boiler evaporator system when the system is controlled by a traditional PID-type strategy and when the system is enhanced by using fuzzy logic blocks to provide set-points for the system. The strategy used in fuzzy logic controllers (FLCs) is called fuzzy supervisory control and it generates set-points for the conventional controllers. The boiler under test is a VU-60 industrial system that produces 180,000 pounds of steam per hour. The mathematical model of the plant is a scaled version model of that obtained for a thermoelectric unit. The new model simplifies the large-scale thermoelectric boiler model to an industrial small-scale type VU-60 boiler model based upon first principle mass and energy balance equations. The main change consists of representing only the behavior of the drum-evaporator system, having a partial model of the combustion process, with a simplified combustion control system and a three-element boiler feed-water controller. The control system for combustion and boiler feed-water receives a supervisory signal (or set-point tracking signal) that comes from the FLC to improve the performance of the overall control system. The behavior of the supervisory controller brings some advantages to the system performance, compared with the traditional control schemes. The comparison reflects fuel improvements from 2.5% to 6.5% depending upon the steam load ramp regime. The simulations are performed using the SIMULINK (R) shell running under the MATLAB (R) platform.