Prognostic analysis based on updated grey model for axial piston pump

Abstract

Gradual degradation is the pivotal factor in prognostics analysis while the performance degradation is difficult to get with limit or blurry data. Based on the failure mechanism analysis of axial piston pump, the abrasion resulting from uneven contact between valve plate and cylinder barrel is the crux. With measured value, this paper presents a updated grey model and algorithm flow based on fix dimension updated technique and δ filter, in which the grey model overcomes the influence of uncertainty and disturbance and δ filter improves the quality of original data. Through renewing the information of grey model dynamically, the precision of prognostics increases greatly. Application indicates that the updated grey prediction algorithm could strengthen the scoring of new data and improve the prognostic accuracy with limit data.