Predictive modelling for 3D inkjet printing processes

摘要

The importance of predictive modelling in addressing quality and reliability issues of 3D printing technology is quite significant. The aim of this paper is to develop and demonstrate a data-driven approach for monitoring and forecasting the quality of electronics structures manufactured by a 3D inkjet printing process. In order to realise this predictive capability, data-driven prognostics techniques that use information for process parameters are employed. The proposed quality assessment of 3D printed electronics products is based on the predictive capabilities of computational intelligence algorithms. The ability of neural network models to assess the quality characteristics of printed electronic products caused by dimensional deviations is evaluated and discussed. Performance capabilities of selected training algorithms for the adopted neural network models is also studied and reported.