Condition Monitoring and Diagnosis of a Drum Boiler Using Algorithms of the Paraconsistent Annotated Logic: Condition Monitoring Model of Drum Steam

Abstract

In this work, a condition monitoring model of drum steam using Paraconsistent Annotated Logic (PAL) was investigated to provide the mathematical and logical foundation for the application of PAL to monitor condition of steam in boiler drum. Three methods are proposed in order to calculate evidence degrees from steam pressure and temperature and evaluate steam state in the drum. The Paraconsistent Analysis Node (PAN) takes evidence diagrams as input to analyze and process the inconsistency and uncertainty of signals corresponding to pressure and temperature. The PAN outputs the resultant real evidence degree, the Contradiction Degree and the Normalized contradiction degree. A network of paraconsistent analysis (PANnet) which is combined with these methods is also proposed in order to analyze and treat the contradiction. And based on this, the application of PAL for condition monitoring of drum steam, one of the local systems of the Fault Detection and Diagnosis (FDD) for Thermal Power Plants (TPPs), is demonstrated. For two faults, the fault detection times are reduced by 60 and 64%, respectively, compared to the conventional ones. The condition monitoring and diagnosis of evaporators as well as boiler drum can be improved by using this steam condition monitoring model.