A data-driven ensemble ranking system of production quality across manufacturers—A case study for risk assessment in the solar industry

In many production industries, quality assurance and risk analysis are important aspects of the procurement of products. Such analysis often takes the form of during and after production inspections performed by quality assurance companies. These inspections give insight into the quality level of each individual production batch; however, it is also often important to understand a manufacturer's overall production quality and how its quality compares to other manufacturers. Therefore, in this paper, we propose a data-driven method to rank the production quality across manufacturers. The method is based on data from quality assurance inspections and consists of three steps. In the first step, a failure mode and effects analysis is applied for each individual manufacturer. This results in a risk assessment for identified failure modes. In the second step, the risk assessments are combined to create an overall failure index for each manufacturer. Step 2 can be designed in various ways. To ensure our final ranking is robust to the selected method, in step 3 we combine all methods into a unified rank of each manufacturer. We validate our proposed ensemble ranking method using a case study from the solar industry, where we compare our final ranking with the experts’ knowledge of manufacturer quality. It is shown that our data-driven method identifies the high-risk manufacturers in accordance with experts’ knowledge.