Development of an Intelligent Diagnosis System for Detecting Leakage of Circulating Fluidized Bed Boiler Tubes.

The detection of leaks in circulating fluidized bed boiler tubes is an important issue in thermal power plants, because leaks lead to enormous economic and social losses. To address this issue, a time-based maintenance (TBM) method has been employed in power plants, but it cannot detect leakage of a tube during operation. Instead, a condition-based maintenance (CBM) method is required to detect unexpected leakage of a boiler tube in real time. This paper proposes an acoustic emission (AE)-based diagnostic system for detecting leakage of a circulating fluidized bed boiler tube using a data acquisition (DAQ) system and support vector machines (SVMs) for data acquisition, data analysis, and leakage detection. Experimental results show that the proposed diagnosis system perfectly detects leakage of a boiler tube on the simulation testbed.