A Product Process Adaptive Design Method Based on Manufacturing-Related Big Data

Abstract

As digital and smart production methods being applied widely in manufacturing, enterprises should pay more attention to the values of manufacturing-related big data, which is important to the innovation of product process design. This study aims to propose a product process adaptive design method based on manufacturing-related big data. This method is proposed based on the requirement of enterprises for data–business deep integration and it is used for solving the insufficient utilization of manufacturing-related data among enterprises. To this end, a product process adaptive design model “data+ knowledge+ decision” is proposed and the data mining and utilization processes are summarized for the model, namely, multi-source heterogeneous data fusion; data cleaning and preprocessing; data conversion and dimensionality reduction; data mining; data visualization; and design decision. Subsequently, the automobile welding process is used as an example. A prediction model of the relationship between welding parameters and welding defects is established, aiming to improve welding quality and realize welding process adaptive design. This research reveals that manufacturingrelated big data contains rich knowledge and patterns and thus can guide product design decisions and support the product process adaptive design under different manufacturing environments. In the future, to enhance the integration of manufacturing-related big data with product process design, the integration of big data with 5G technology should be promoted and investment should be increased in the development of big data and algorithm design platforms.