A Method for the Generation of Membership Functions using Operating States in a Glass Furnace

Abstract

A fuzzy set is fully defined by its membership function (MF) and the design of the membership function is an integral part of the whole fuzzy system design. This work looked at using operating states in a chemical plant to identify the membership function parameters. There were three operating states discovered, namely, the not normal operating state (NNOS), the normal operating state (NOS) and the transition operating state (TOS). These states were used to design the MFs of the fuzzy variables with specific reference to setting the limits for the MF and describing the linguistic terms and shape of these terms for the MF. The work showed that dropping sharply and rising sharply could be associated with the not normal operating state and dropping gradually could be used for the normal operating states and so on.