That Day the Machine Broke Down, and My Supervisor Only Asked: "Are You Sure Now Is the Best Time to Repair It?"
I still remember a night shift during my early years at the company when a machine suddenly crashed. The screen turned red, and the alarm blared piercingly. Sweating profusely, I immediately called the production line and then summoned equipment maintenance to handle it. The equipment technician arrived, took a look, and said a part needed to be replaced. I thought, "Great, it's fixable!" However, as he prepared to start, my supervisor appeared, glanced at it, and asked, "Are you sure now is the best time to replace this part? We still have two hours until shift change." I was completely bewildered at the time, thinking, "The machine is down, shouldn't we fix it immediately?" Only later did I realize that there's a profound amount of knowledge behind the simple act of "repairing."
Where the Problem Lies: "Can Repair" Does Not Equal "Should Repair"
In essence, optimizing equipment maintenance scheduling means finding the best timing for a "maintenance window." We all know that a crashed machine certainly needs repair, but often, Preventive Maintenance (PM) is the main event. Good PM prevents frequent machine breakdowns. However, PM also requires machine downtime! Downtime means no production, which translates to a loss for the production line. Therefore, we shouldn't just consider whether a part "can be repaired," but rather whether "repairing it now" offers the best overall benefit.
For instance, imagine you have a machine producing products whose DPMO value has dropped from the usual 6210 (approximately Cpk 1.08) to 9340 (Cpk 0.98). Clearly, quality is deteriorating, and the machine might be on the verge of a critical failure. At this point, you would naturally want to perform PM. However, if the production line is rushing to ship orders with immense pressure from urgent demands, forcibly stopping the machine for PM could lead to production delays. Losing money might be minor, but damaging client relationships is a major issue. So, the key is how to find a balance?
How to Do It in Practice: Three Dimensions for Evaluating Maintenance Windows
To decide on a maintenance window, I consider three critical aspects:
- Equipment Health: This is the most intuitive. What is the current status of your machine? Are there any abnormal alarms? When was the last PM? How many units has it processed since the last PM? Are any indicators (e.g., vibration frequency, temperature, estimated consumable life) approaching alert levels? For example, if a part's expected lifespan is 1000 hours and it has already run for 950 hours, then scheduling maintenance promptly is a must.
- Production Line Status: This aspect is highly dependent on timing. Does the production line currently have urgent orders? Are other machines already down? Will your machine's downtime severely impact overall production capacity? If the production line is currently slow, or if other machines can provide support, then the loss from PM at that time would be relatively small. I typically monitor the UPH (Units Per Hour) changes over the previous hour; if the UPH begins to decline by 5% or more, it might warrant attention.
- Manpower and Materials: To perform a repair, you need equipment engineers and spare parts, right? If equipment personnel are stretched thin, with everyone occupied on other machines, or if the spare parts you need are out of stock, then even if you desperately want to repair it, you can't. Therefore, before scheduling a repair, you must first confirm that these resources are in place.
The key point is that these three dimensions are not viewed in isolation; they are interdependent.
The Most Common Pitfall: Penny-Pinching Leads to Pound-Foolishness
The most common pitfall I've encountered is stubbornly delaying PM to save a small amount of "downtime." What's the result? The Cpk drops below 0.8, leading to a batch of products being scrapped directly. Even worse, the machine completely crashes, affecting products from other processes, and then requires more time to repair, with more expensive spare parts.
Another time, in a rush for an urgent order, we knew a certain machine's consumable was nearing the end of its life but pushed it to keep running. Consequently, that consumable failed at the most critical moment, also damaging an adjacent, more expensive part. That repair took two full days, the urgent order was naturally missed, and we incurred even greater financial losses. Wouldn't you say this is a classic case of being "penny-wise and pound-foolish"? Frankly, often, as engineers, we rely too much on a machine's "resilience," thinking it can hold on a bit longer, only for it to fail spectacularly.
One Thing You Can Do Today
For the machines you are responsible for, identify the top three most critical PM items. Then, ask your equipment engineer when they were last performed and when the next ones are scheduled. Record this information.