Using Special Filter with Membership Function in Biomass Combustion Process Control

Abstract

The paper deals with the special filtration method using a filter with membership function. The paper presents a model of a filter, its specific characteristics and some parameters that have an impact on quality of filtration. A filter with different membership functions (Gauss, Bell, Power and Triangle) was designed and tested for specific demands, which followed from the experience with the realization of a biomass combustion control system. Data obtained from the combustion process were extremely noisy (influenced by various transfer errors, disturbances and external interferences) and, therefore, had to be properly filtered. The paper also describes some results of filter simulation in the Matlab Simulink environment and its implementation into an on-line monitored process control system of biomass combustion. It was proven by implementation that such a filter can be useful for the signal filtering of oxygen concentration and carbon monoxide emission sensing and it can be very useful in reducing signal interferences arising in biomass combustion.