Skip to main content

Hitachi

Corporate InformationResearch & Development

January 29, 2018

Report from Presenter


Photo 1 Presentation scene

The International Conference on Industrial Engineering and Engineering Management (IEEM) 2017 was held at the Suntec Singapore Convention & Exhibition Centre from 10-13 December 2017. IEEM, sponsored by IEEE, is one of the leading international forums on industrial engineering and management engineering and has been held in various locations in Asia from 2007. IEEM 2017 was attended by about 500 participants from 50 countries. More than half of the participants were from Asia such as mainland China, Hong Kong and Singapore but there were also numerous participants from European countries such as Germany and Norway. At the conference, Hitachi gave a presentation entitled "Robust Inference Traceability Technology for Product Quality Enhancement" (Photo 1).


Fig. 1 Challenges in data preparation
Enlarge



Fig. 2 Outline of proposed method
Enlarge

In the manufacturing industry, to improve product quality and productivity, it is important to identify the cause of defects in individual products as they occur. In general, an individual ID (unique identifier) assigned to each individual product is used to connect and analyze the manufacturing data generated by different equipment. There are however some harsh machining processes such as sintering and cutting where the manufacturing conditions do not allow individual IDs to be attached, and thus preventing the preparation of necessary data for defect analysis (Fig. 1).

This research proposed and developed Robust Inference Traceability (RIT) technology which links manufacturing data generated by different equipment in various manufacturing conditions. RIT is able to connect manufacturing data by accurately estimating the time each individual product passes through different equipment by using machine learning of processing history, such as production time interval and number of items being processed between different processes (Fig. 2). The results indicate that it may now be possible to identify the root cause of defects in manufacturing which may include processes where individual IDs cannot be attached.

  • Page top