A NEW METHOD OF SYMPTOM BASED PROCESSING FOR MACHINERY CONDITION RECOGNITION AND FORECASTING

This paper considers a new method of construction of the average symptom life curve and symptom reliability for machine condition monitoring, as an alternative to the statistical elaboration of these machine characteristics. It was found that by the sorting up operation of the symptom data and extraction of equal symptom readings, similar recognition and forecasting abilities can be obtained, as in statistical Weibull and Frechet characteristics. It is shown that the new method is also good for self-diagnosis, (i.e. self-recognition of a symptom life curve) based on a small set of symptom readings taken during monitoring of a given object. Such a direct construction method of diagnostic characteristics may be of great value in cases where not enough data for reliable statistics are available. This is usually the case at the start of condition monitoring of some machinery set, or critical object.