Root cause analysis is the structured investigation that starts from a failure and drills down to the cause that, once fixed, prevents recurrence. In maintenance, it is the mechanism that turns breakdowns into learning; without it, the same failure returns with a new date. This article applies the three most used methods (5 whys, Ishikawa and FMEA) to a concrete case, from symptom to countermeasure.
When to open an analysis
Analyzing everything is analyzing nothing. Triggers that work: failures with downtime above a defined threshold (for example, 4 hours), repeated failures on the same asset within 90 days, any failure with safety consequences, and failures on class A assets regardless of duration. A trigger written as a rule takes the decision out of the day's mood.
The case: paint booth exhaust fan stoppage
The scenario to work the methods: the booth's exhaust fan stopped due to a seized bearing, causing 6 hours of line downtime. The replacement solved the symptom. The analysis begins after the replacement.
Method 1: the 5 whys
Ask why until the answer stops being technical and becomes systemic:
1. Why did the fan stop? Seized bearing at the housing.
2. Why did the bearing seize? Insufficient lubrication, dried-out grease.
3. Why was the grease dried out? The point had been off the lubrication route for months.
4. Why was it off the route? The point requires unbolting a guard, the route has a tight time budget; the lubricator skipped it and did not record the skip.
5. Why could it be skipped without a record? The route was on paper, with no per-point checklist.
Note the chain's shape: it starts at the component, passes through the process and ends at management. Each level generates a countermeasure on its own floor: a remote grease fitting at the point (level 4), a digital route with per-point checklist (level 5). Fixing only level 1 (replacing the bearing) is what most call a solution, and it is only replacement.
Method 2: Ishikawa, when the failure has more than one mother
The Ishikawa (fishbone) diagram organizes hypotheses into families (method, machine, manpower, material, measurement, environment) before drilling down. In the fan's case, the fishbone would surface parallel hypotheses beyond lubrication: paint mist accelerating grease degradation (environment), wrong grease for the point's temperature (material), no sensory inspection at the station (method). Ishikawa does not replace the 5 whys: it opens the fan of hypotheses, the whys drill into each one the evidence supports. The golden rule: a hypothesis without field evidence does not become a cause, it becomes an organized guess.
Method 3: FMEA, the analysis before the failure
While 5 whys and Ishikawa react to a failure that happened, FMEA anticipates: it lists each component's possible failure modes and prioritizes by RPN (severity × occurrence × detection, each 1 to 10). On the fan, the mode "bearing seizure from lubrication failure" would come out with severity 8 (for the line), occurrence 6 (a history of skipped routes) and detection 7 (no inspection at the point): RPN 336, top of the list, demanding exactly the countermeasures the reactive analysis found after the pain. A maintenance FMEA does not need to cover the plant: it starts with class A assets and the history's 10 worst offenders.
What makes analyses die on paper
Four known killers: stopping at the culprit (an analysis that ends in "human error" has not ended: the why of the error being possible is missing); countermeasures without verification (implemented, nobody came back to measure recurrence); analysis without data (without a structured failure record, every meeting restarts from memory); and volume without criteria, which saturates the team. The antidote to all four is the same: a clear trigger, reliable execution data and a countermeasure owner with a date.
Where the data that feeds everything comes from
Analysis quality is hostage to record quality: failure mode, cause and component cataloged on the notification, real times confirmed on the order. It is the difference between analyzing facts and analyzing memories, and the reason root cause analysis is one of the operational pillars of reliability engineering.
Frequently Asked Questions
What is root cause analysis?
It is the structured investigation that starts from a failure and drills down to the cause that, once fixed, prevents recurrence, producing verifiable countermeasures at every level of the causal chain.
What is the difference between 5 whys and Ishikawa?
Ishikawa opens the fan of hypotheses into families (method, machine, manpower, material, measurement, environment); the 5 whys drill vertically into each hypothesis supported by evidence. Used together, one structures the width, the other the depth.
Does FMEA apply to industrial maintenance?
Yes: applied per critical asset, it lists failure modes and prioritizes by RPN (severity × occurrence × detection), showing where to reinforce plan, inspection and spares before the failure happens.
When is a root cause analysis worth opening?
With a defined trigger: downtime above a threshold, repeated failure in a short window, any safety consequence and failures on class A assets. Without a trigger, analysis becomes an indignation lottery.
