You were recently hired to oversee the maintenance department. The plant engineer told you that despite a rigorous insulation resistance testing program, the plant keeps experiencing cable failures.
The program is handled inhouse and overseen by Bob, so the plant engineer had you sit down with Bob to review it. Bob has put the whole program into the computerized maintenance management system (CMMS), and together you reviewed some reports.
Bob had set pass/fail parameters for insulation resistance values for each type of cable. Bob arrived at these values by establishing the mean and then calculating one standard deviation to get the “normal range.” Any cable outside of normal was flagged for replacement.
What would you recommend?
Bob’s approach, while impressive from a mathematical ability perspective, is fatally flawed. There’s not a “normal range” you can apply by cable type for predicting cable failure. The only way to predict failure is to trend the measurements over time. If you chart them, you’ll see a “hockey stick” or similar break in the curve when a given cable needs replacement.
The fact that Bob didn’t know this means there are likely other flaws in the program. For example, when they do replace a cable are they staying within the bend radius and cable pulling tension limits? Or, are they replacing good cables with ones they damage due to flawed installation techniques?
A qualified electrical testing firm probably should take over this program.
