Please enable JS

Enabling predictive quality and predictive production through digital process twins


A global automotive supplier was facing two major challenges in its dashboard panel production: on the one hand, risks in the process chain remain undetected due to a lack of transparency and poor supplier management. On the other hand, the processing of highly sensitive materials means that even the smallest of errors in the production process can lead to the rejection of complete components. ROI-EFESO supported the customer by using a ‘digital twin’ to map the entire production process step by step. This led to an IoT pilot project that pinpointed potential for improvement and triggers improvements in both production and value stream management at suppliers. As a first step, the project team defined process parameters that could affect both the performance of the plant and the quality result. The derivation of these parameters was empirically based and included initially more than a hundred different parameters, which were reduced or supplemented in the further course of the analysis. In the next step, the team verified that the existing process data had been correctly aggregated and processed and gathered data that had not been previously recorded. The team also used additional sensors to capture the new data required for the defined parameters.

The resulting database was then consolidated and analyzed in a cloud application. On this basis, the project team designed a model that reflects the process of improving the production line for dashboards as precisely as possible - in other words, all relevant parameters, their interactions and critical values. Like the process itself, this model can go far beyond the company. In the case described, for example, it can be extended to the logistics company that transports the foam, or even to the foam manufacturer. This is because the reasons behind problems in the foaming process - such as dangerous temperature fluctuations for the sensitive polyurethane - can appear at any point in the value chain. As a result, the company obtained a digital process map that tracks the entire physical process in real time and enables early intervention according to critical process parameters - a "digital process twin".


With the help of this ‘digital twin’, the project team was not only able to significantly reduce the plant’s rejection rate, but also to make the correlations of relevant influencing factors on the quality result more transparent. The company also developed a prediction model of the result at the next quality gate.