Manufacturing process reliability evaluation based on Granger causality test and Cox model

To evaluate the impact of manufacturing process on the product reliability quantitatively, a comprehensive evaluation method is put forward together with integration of the Granger causality test (GCT) and Cox proportional hazards regression model (PHRM). This method facilitates the identification of the weak manufacturing processes and the key process characteristics that affect the product reliability. Firstly, Granger causality test is conducted on the variable series, i.e. process characteristics and product characteristics, the proposed model could be used to determine the quantitative relationship between the deviation of machining process and the satisfaction of product characteristics, which is provided as a more accurate reliability evaluation model. Then, the Cox PHRM is constructed based on the Granger test result. While the process characteristics series and the product characteristics series is the Granger cause for each other, the Cox model is established with cross term. Otherwise, the Cox model would be built without cross term. Meanwhile, the hazards ratio is given as an index of Cox model to judge the criticalities of different processes. Afterwards, the product reliability and life expectation are calculated under the weibull Cox PHRM assumption. Finally, the case study on reliability evaluation of the manufacturing process of radar phase shifter is demonstrated to prove the feasibility and effectiveness of the proposed method.