STUDY ON FAILURE DIAGNOSTIC METHOD AND INTELLECTUAL DIAGNOSTIC SYSTEM OF RECIPROCATING COMPRESSOR

摘要

Three categories of the failure diagnostic methods are classified according to the signal adopted by the diagnosis. They are the parameter method, the vibration method and the oil analysis method. In this paper, the applicable range and operational difficulties of these methods are discussed on the basis of analysis and induction upon normal failure. It's proposed that compressor's normal failure can be divided into thermodynamically property failure and mechanical function failure, as to the former, the parameter method that takes cylinder pressure signal as main diagnostics signal my be applied, as to the latter, the vibration signal frequency spectrum can be used to diagnose. At the same time, the stutter based on artificial intellectual neural network diagnostics system is introduced and its general chart is given.