Global Conduit Manufacturer Saves $5M Per Month in Quality Failures

Summary

Even with a great deal of investment, a global conduit manufacturer was experiencing $5 million in non-conformance costs per month. They needed a better way to solve quality failures. They worked with Oden to put an industrial analytics solution in place to reduce defects, improve responsiveness and identify perfect line speed targets without sacrificing first pass yield. 

Situation and Background

A global conduit manufacturer lost a key multi-million-dollar customer due to recurring shipments of out-of-spec pipe which could cause safety-related issues in the field.

Even with significant investments in advanced technology and measurement solutions in place they realized they were experiencing over $5 million in the cost of non-conformance each month. While the measurement systems were collecting rich sets of data, accessing that data took manual intervention. If the data was not manually pulled within 24 hours, it was no longer accessible resulting in extended periods of manufacturing defective product.

The team realized they needed a better way to process data and proactively identify and address potential quality failures.

They needed a better way to process data and proactively address potential quality failures.

Learn how you can increase throughput while maintaining quality with Oden.

Optimizing Production Process Controls

The team engaged Oden Technologies to provide a product traceability solution that would reduce defects and perfect line speed targets without sacrificing first-pass yield.

The process began by collecting and analyzing all relevant data for the ongoing measurement of pipe being produced, as well as the associated process data leading to those results which:

    • Provided greater product traceability for plant managers and quality assurance teams
    • Helped identify and address suspect product to reduce the likelihood of defects being shipped to customers
    • Gave supervisors and process engineers access to all data necessary to properly conduct root cause analysis

Oden Technologies helped to perfect line speed targets without sacrificing first-pass yield.

Moving Toward Predictive Performance

With faster analysis of process data, engineers were able to move from simply diagnosing problems to proactively improving performance and product quality.

Oden evaluated historical measurement data to find the optimal output rate when pipe measurements had the tightest control. The best case segments were then compared to other segments using machine learning to identify which process conditions were the most important along with optimal control limits to drive the best performance.

The out-of-the-box recommendation engine prescribed optimal settings for controllable variables to replicate these runs more consistently. Real-time alerts when critical process conditions went outside of optimal parameters allowed engineers to quickly take action and prevent quality failures from happening.

The out-of-the-box recommendation engine prescribed optimal settings to replicate the best production runs.

Winning the Customer Back

In the end, the conduit manufacturer was able to improve product quality and recover their lost customer.

Additionally, they were able produce orders more efficiently as recommendations identified over 200 production hours that could be saved by adjusting settings.

This allowed them to lower operating costs associated with individual orders as well as fill subsequent orders faster by reducing the opportunity cost of machine time.

About Oden Technologies

Oden is the fastest and least resource intensive way for manufacturers to get the analytics and guided insights their team needs to be problem solvers. Our software unifies siloed machine, production data, and prebuilt machine learning applications to provide real-time insight and context. Teams are enabled to make faster, more proactive decisions, and reduce firefighting. Our customers have deployed in 42 days, and have achieved 5x ROI in 12 months.