Autonomous In-Process Quality Control

Intense global competition has forced many US manufacturers to examine their current business practices as well as evaluate how to meet these challenges and remain competitive.  Major emphasis has been placed on disruptive innovation and manufacturing research with recognition of the need to improve product and process quality, automation, and decrease scrap.

The traditional manufacturing process is to produce a part on machine, transfer it to the coordinate measuring machine, check for quality, and make decisions.  If the quality is found unacceptable, the production line is stopped to locate and alleviate the root cause of the defect. In such Post Production Quality Control practice, not only is the part inspected scrapped, but also a considerable amount of work in progress (WIP) is purged. This preventative quality control practice, which barely manages to satisfy customer quality requirements, is not sustainable in today’s competitive market place.

As a better solution, manufacturers have embraced the concept of Post Process Quality Control utilizing on-machine probing technologies. In this reactive methodology, the part is measured for quality attributes on the machine itself after completion of a single/multiple machining process ensuring quality at source. It leads to lesser WIP scrap and the need for CMM evaluation. However, it still suffers the drawback of measuring the non-conforming part after it’s produced. Additionally, this method does not support the philosophy of first part correct. It also consumes expensive machine time for quality checks (reduced throughput at the given station) and may require special labor skills to operate the probe and probing routines.

The goal is to develop a MTConnect-enabled, autonomous, in-process quality control system for manufacturing first part correct with integrated user-initiated alarm management. The proposed In-Process Quality Control predictive methodology monitors critical, real-time process/machining parameters, checks them against boundary limits, and then raises alarms/alerts. The philosophy is fundamentally different from Post Production Quality Control and Post Process Quality Control because we measure and constrain the in-process machining parameters like power, forces, torque, thermal effects, vibration level, etc…, which cause quality issues, instead of reacting to the problem after it has already occurred. The proposed system will ensure zero scrap and minimum downtime thus boosting the overall equipment effectiveness and process capability indices.  Another major benefit of the system will be to reduce the number of inspected parts in a batch of final product which will directly translate into reduced lead time and inspection cost.

The current alarm management systems are not customizable and user-initiated.  Some problems include too many alarms on display, high alarm rates (flooding), false alarms, repeating alarms, excess grouping of alarms, no defined action, and wrong recipient.  Also, the current alarms are based on the equipment health/condition and not process conformance. The proposed system will eliminate these issues and have an integrated, user-initiated alarm management system that will notify the user of real-time faults and process violations.  

Lastly, the system will be MTConnect standard compatible. The current manufacturing systems are proprietary. Vendors bundle together software and hardware for competitive purposes. The MTConnect Initiative was started to address the need for a standard that will help realize a "seamless manufacturing pipeline" from design to production. The goal of this pipeline approach is to allow for “universal” capture of data from the machine tool which would transfer this captured data into other control systems thereby facilitating a seamless method for managing and analyzing data for process and product optimization. MTConnect compatibility will ensure industry-wide applications, easier integration, and commercial viability for the proposed In-Process Quality Control system.