Benefits from condition monitoring techniques: a case study on maintenance scheduling of ball grid array solder joints

Abstract

In the recent decade, more and more efforts in the field of reliability engineering are targeted towards developing condition-based maintenance (CBM) techniques for improving system availability and reducing maintenance cost. To justify the increased condition monitoring cost in a CBM programme, it is necessary to benchmark the performance of a maintenance policy enhanced by condition monitoring techniques with that of the conventional preventive maintenance (PM) policies. However, such studies are very limited in literature. This article presents a comparative study on maintenance scheduling of ball grid array (BGA) solder joints in a maintenance policy enhanced by condition monitoring techniques (policy A) and a conventional PM policy (policy B). Driven by the practical characteristics of the degradation signals of BGAs, a two-phase treatment (stable phase and unstable phase) is adopted in policy A: statistical process control techniques are used on the initial stable condition monitoring data of BGAs to detect the shift from the stable phase to the unstable phase, and trigger a traditional PM policy within the unstable phase. Benefits of the condition monitoring techniques are demonstrated in terms of a nontrivial reduction of maintenance cost rate in policy A, compared with that in policy B. From anther perspective, the current study also demonstrates the feasibility of incorporating condition monitoring techniques with the conventional PM policies for improved maintenance performance. The modelling approaches herein are generic and can be directly applied for other systems with similar degradation processes, such as bearings, mills, and gears.