Classification of wear phenomena by specific ultrasonic emission detection for prognostic purposes

Abstract

Tribological effects (e.g., friction) often define the functionality of typical mechanical elements and mechanical engineering structures. If friction processes does not work well due to bad tribological conditions, sliding surfaces may be destroyed and the components functionality may be reduced up to a complete loss of functionality. The definition of this damage level depends on the particular application and the related tolerable level of deterioration. Hence, this is an individual characteristic that has to be quantized and quantified beforehand, so that the related knowledge can be used for automated supervision, for example in the context of condition-based maintenance concepts etc. In the tribological context the surface of the considered individual component is usually evaluated by visual inspection, which is time-consuming and a subjective measure. Furthermore, material displacements, inner cracks,... might not be detected by visual inspection. Therefore, an automated and continuous monitoring of safety relevant structures affected by wear effects may be useful to improve SHM- or CM-related goals. Destructive testing concepts reveal the level of deterioration of a component at a discrete point of time. Though, the progression of fatigue is hardly reconstructable and the process cannot be continued with the same component. This is only possible with non-destructive methods, which observe/measure those signals that indicate the fatigue progression. Hence, the destructive testing provides the reference damage level at discrete operating times while the non- destructive testing fills the gap between those discrete information with the damage progression in between.

This contribution deals with the problem of detecting and monitoring signals indicating tribo- logical effects with a non-destructive concept. Therefore a new sensor technology is applied and first considerations about the related data filtering technique are considered. The main

idea is to monitor ultrasonic emission properties of the tribo-system. For first experiments using this technique a test rig for wear examination with variable lubrication, and normal force has been developed. This tribological system is equipped with several sensors, amongst others several piezoelectric materials. The transducers are used as ultrasonic sensors, measuring the structure-borne noise. The goal is to connect the characteristic signals unambiguously to their unique sources, e.g., abrasive wear and surface fatigue. This contribution details the possibility and application of structure-borne noise measurements, and shows preparative results for determination of deterioration and for distinguishing different wear-related effects.