Development of a decision support system to enable adaptive manufacturing

The drive for Industry 4.0 has allowed manufacturing companies to stay at the forefront of business competition by making vast improvements over the years. Adaptive manufacturing is a key area for industries to explore where production processes can be fully automated using advanced technologies. However, such processes can generate a large collection of data, and there are difficulties with analyzing the data to derive useful information that will be used to assist the end users in making decisions for any disturbance in the machine during production. An expert system was developed in this study that sends offset feedback from the coordinate measuring machine to the computer numerical control (CNC) machine and aligns the tools appropriately within the CNC, thus extending tool life. Critical-to-quality (CTQ) dimensions were identified from scrap history, and decision trees were developed for each CTQ by using heuristic knowledge of the end users. Thus, this novel approach, which uses a rule-based method because of a lack of training data, includes a set of “IF THEN” statements codified from the decision trees, describing the alterations required on the critical tool if the measurements are not within specification. The ruleset was tested by force-failing the tools associated with the rules where the triggered responses were checked against the stated responses. Although the rule-based expert system proved to be successful at making offset changes for correcting tool positioning, further improvements to the ruleset are required to tackle any uncertainty, and operator interaction needs to be assessed to rule out the non-value-adding steps and to codify useful tacit knowledge.