Etra: Machine Performance (CONT’D FROM PAGE 58)
What were more useful were the comments, albeit even they required some deciphering. In this case a fair amount of effort went into producing a report of limited value. This issue is not limited to manual and semi-automatic collec- tion systems, as I have seen similarly inaccurate data from machines that automatically record stops, and even from those with the capability of identifying and logging elec- trical or mechanical issues, since the crew still needs to identify non-machine related downtime. So how can we simplify downtime data collection and improve its utility? Implementing more complex schemes and collecting more granular data across the board re- quire significant effort with limited return inasmuch as they still require suspect crew input. A simpler approach is to start at a high level: estab- lish a target downtime percentage for each machine and focus only on machines exceeding your targets, starting with the worst offenders. For these, create a limited num- ber of downtime codes which will facilitate collection of high-level data which separate crew, machine, and what I call service issues. Then, having determined which category or catego- ries are the main offenders, create additional sub-codes allowing crews to record more detailed information for the major issues. Also, don’t overlook the obvious; if you just ask your crews to tell you off the top of their heads what is causing the most downtime, they can usually give you surprisingly accurate responses. When you actually drill down into the data, you may be shocked by how much downtime can be attributed to one or two causes. Although everyone may be aware that something is a problem, they may not realize the totality of time is being lost. I have seen downtime of forty hours in a single month caused by a single repetitive equipment malfunction. Because reporting was flawed, this was bur- ied in the reports and not identified; sure, everybody knew that this was a repetitive problem, but when the scope was identified, management and crews were astounded that it accounted for an entire shift for a whole week. Not realiz- ing how serious this problem was, and not wishing to in- terrupt production or spend the money to deal with it, they had been content to tolerate it. Once the magnitude was determined, they realized that this had been a mistake. Sometimes, identifying the true culprit can and should change the calculus of where you should spend your maintenance time and dollars. Replacing expensive com- ponents that entail significant maintenance hours looks a whole lot more attractive if they account for forty hours of lost time per month versus what you heretofore believed was three hours. One clue in analyzing excessive downtime is when it almost always occurs on the same shift. That is a red flag suggesting that the problem is caused by the crew or by supporting personnel, regardless of what the crew reports. Is the crew operating the equipment improperly? Are inks,
they dash off the sheets at the end of their shifts? Perhaps you are feeling smug because paper logs are so outdated, and your equipment automatically captures the number and duration of stops. But is this useful? Inherently it’s only of value to the extent that it alerts you to excessive downtime but does nothing to identify the causes. You are still at the mercy of the crew to record the reasons for interruptions, but how accurately are they selecting and entering the downtime codes? Furthermore, do they take unscheduled smoke, texting, or other breaks and disguise them by enter codes for legitimate down- time? That brings us to another issue: how extensive are your downtime codes? If, for example, you have only a single code for mechanical downtime, it is not specific enough to identify a malfunctioning component or section. On the other hand, if you have twenty-five codes, does your crew — when focusing on getting the machine run- ning — spend the time to select the correct code, assum- ing they even understand what the codes mean? In one plant I visited, each crew recorded downtime each shift on a hard copy log, ostensibly selecting from no less than twenty-two codes. The data were then entered into a spreadsheet by a clerk who summarized them by code. When I began analyzing the data, it became obvious that the code entries were inaccurate and of limited value.
Suppliers: Leverage Our Reach
Board Converting News is read by more independent AND integrated decision-makers than any other industry publication.
Len Prazych at 518-366-9017 lprazych@nvpublications.com Advertise in BCN in print and/or online to leverage our expansive reach!
CONTINUED ON PAGE 62
60
www.boardconvertingnews.com
August 28, 2023
Made with FlippingBook Digital Publishing Software