Predicting Your Next Failure

Every production system tells you what will break next — if you are capturing the right signals. This post covers a recent engagement with a high-throughput automotive customer: quantifying their biggest system pain points, then deploying cloud software that predicts and prevents failures before they stop the line.

The Cost of Unplanned Downtime

Unplanned downtime has a cost far beyond the part and labor costs for an emergency field service response. The fact it is unplanned means firstly that the cause for this downtime is not understood. Secondly the work capacity of the downed machine is not just zero, it's going negative every hour. A fast repair might be possible to get production going again, but if the root cause was never determined it's just a matter of time until the system fails again for the same exact reason, or worse a brand new failure happens adding to the cost and confusion.

This can result in hundreds of thousands, if not millions, of dollars in lost production, overbought spare parts, and expensive field labor.

Finding the Real Pain Points

At a recent customer visit, this was exactly their situation. Multiple lasers experiencing similar failure modes, without any clear understanding from the OEM or the client's staff of what was really happening. The client contacted us in desperation, the line needed to be up, but they also needed to understand why they kept running into these failures.

We jumped into action, gathering system data, interviewing key personnel and physically tearing down the broken system on the line. Within the first couple days of analysis, we found the smoking gun. In this case there was an unflagged chiller flowrate that was far too low for the 24/7 production load on these systems. The result was optic temperatures constantly spiking above the intended operating temperatures- more thermal cycling on the optical components, more costly failures.

We corrected the clogged/contaminated cooling lines and heat exchanger in the system leading to the poor cooling performance, restoring optic flow on these systems from almost nothing to over 2.0 L/min.

The result: cool optics, more stable processing, and no more surprise failures.

From Reactive to Predictive

These particular systems operated on a legacy software package that was built for German engineers, so ultimately it wasn't accessible to a level that would enable intelligent system monitoring and quick responses.

That's where Phos Cloud came in, cloud software that can turn arcane legacy equipment into smart, remotely monitored systems accessible from anywhere with predictive maintenance alerts and machine learning applied to enable clear operational visibility, fast maintenance response, and greatly reducing the chances of surprise equipment failures.

Results

The result for this client was full visibility into their laser operation, enabling deep understanding and a predictive posture towards future failures.

Get Help Now

Where downtime is unacceptable, and failures are still common, contact Phos Industries to get to the root of the problem.

Email support@phos-ind.com or text/call 913-547-5317. We typically respond within 24 hours, sooner if you have an urgent need we can address.