Data-driven prognostics for smart qualification testing of electronic products

A flaw or drift from expected operational performance in one electronic module or component may affect the reliability of the entire upper-level electronic product or system. Therefore, it is important to ensure the required quality of each individual electronic part through qualification testing specified using standards or user requirements. Qualification testing is time-consuming and comes at a substantial cost for product manufacturers. Many electronics manufacturers have access to large historical sets of qualification testing data from their products which may hold information to enable optimization of the respective qualification procedures. In this paper, techniques from the domain of computational intelligence are applied. The development of data-driven models capable to forecast accurately and in-line the end result of a sequence of qualification tests is discussed and presented. Data-driven prognostics models are developed using test data of the electronic module by Neural Network (NN) and Support Vector Machine (SVM) techniques. The performances of the models in predicting the qualification outcomes (pass or fail) are assessed.