You've got a pump that keeps breaking. You replace the seal, it runs for three weeks, then it fails again. Same seal, same symptoms, same emergency work order. Your team knows the pump is "problematic," but nobody has ever sat down and systematically asked: What exactly fails on this pump, in what way, and why?
That's what Failure Mode and Effects Analysis (FMEA) is for.
FMEA is a structured method for identifying every way a piece of equipment can fail, understanding the consequences of each failure, and prioritizing which failures to address first. It turns tribal knowledge into documented data — so your team stops guessing and starts preventing.
What Is FMEA?
FMEA (pronounced FEE-mah or spelled out as F-M-E-A) stands for Failure Mode and Effects Analysis. It's a step-by-step approach to:
- List every possible failure mode for a system, asset, or process
- Identify the effects of each failure
- Determine the root causes
- Evaluate current controls (what you already do to prevent or detect it)
- Assign a risk priority number (RPN)
- Recommend actions to reduce the highest risks
The key insight: not all failures are equally important. FMEA helps you decide where to spend your limited maintenance budget and time.
A Brief History
FMEA was developed by the U.S. military in the 1940s and later formalized by NASA and the aerospace industry. If you're designing a spacecraft that can't be repaired mid-flight, you need to think hard about every possible failure before launch.
By the 1970s, the automotive industry adopted FMEA — first Ford, then the rest of the Big Three, then suppliers. Today, IATF 16949 (the automotive quality standard) requires FMEA as part of the design and process development cycle.
Since then, FMEA has spread to every industry where failure matters: oil and gas, chemical processing, power generation, medical devices, electronics, food manufacturing, and general industrial facilities. The methodology is flexible enough to apply to a single pump or an entire production line.
The FMEA Process Step by Step
Here's how a typical FMEA session works. Ideally, you do this as a team — maintenance, production, engineering — not alone at your desk.
Step 1: Define the Scope
Start by defining what you're analyzing. Is it a single critical pump? A packaging line? A whole utility system? Draw a boundary and stick to it.
Step 2: Identify Failure Modes
A failure mode is simply the way something fails. For each component in your scope, ask: "What could go wrong here?"
Common failure modes include:
- Seal leak (mechanical seal wear)
- Bearing seizure (overheating, loss of lubrication)
- Impeller erosion (abrasive particles in fluid)
- Motor winding burnout (overload, phase imbalance)
- Cavitation damage (suction pressure too low)
- Vibration/excess wear (misalignment, imbalance)
Don't filter or prioritize yet — just list everything. You'll rank them later.
Step 3: Identify Effects
For each failure mode, describe what happens. This is the "effects" part of FMEA.
| Failure Mode | Immediate Effect | System Effect |
|---|---|---|
| Seal leak | Fluid drips from pump shaft | Product loss, environmental hazard, downtime for seal replacement |
| Bearing seizure | Rotor locks up | Complete pump failure, production stoppage, shaft damage |
| Impeller erosion | Reduced flow rate | Lower production throughput, higher energy consumption |
| Motor burnout | Pump stops entirely | Production line down, replacement motor cost |
Step 4: Identify Causes
Ask "why does this happen?" for each failure mode. Be specific.
- Seal leak: Wrong seal material for the fluid, excessive shaft runout, dry running
- Bearing seizure: Contaminated lubricant, overgreasing, age-related fatigue
- Impeller erosion: No strainer upstream, abrasive particles in fluid, incorrect impeller material
Step 5: List Current Controls
What do you already do to prevent or detect these failures?
| Failure Mode | Prevention Controls | Detection Controls |
|---|---|---|
| Seal leak | Monthly visual inspection | Seal flush line pressure gauge |
| Bearing seizure | Greasing every 500 hours | Vibration monitoring quarterly |
| Impeller erosion | Strainer cleaning weekly | Flow meter trend monitoring |
| Motor burnout | Overload relay protection | Thermal imaging annually |
Step 6: Calculate RPN
This is where FMEA turns subjective discussion into a number you can use.
Risk Priority Number (RPN): Severity × Occurrence × Detection
RPN is a simple multiplication:
RPN = Severity (S) × Occurrence (O) × Detection (D)
Each factor is rated on a scale of 1 to 10.
Severity (1–10)
How bad is the effect?
| Rating | Criteria |
|---|---|
| 1–2 | Minor: negligible effect on operation |
| 3–4 | Low: minor production disruption |
| 5–6 | Moderate: partial production loss |
| 7–8 | High: major production loss, equipment damage |
| 9–10 | Critical: safety hazard, regulatory violation, catastrophic damage |
Occurrence (1–10)
How likely is the failure to happen?
| Rating | Criteria |
|---|---|
| 1 | Almost never: < 1 failure per 10 years |
| 2–3 | Low: 1 failure every 1–5 years |
| 4–6 | Moderate: 1 failure every 3–12 months |
| 7–8 | High: 1 failure per month |
| 9–10 | Very high: 1 failure per week or more |
Detection (1–10)
How likely are you to catch the failure before it causes harm? (Low score = easier to detect = better.)
| Rating | Criteria |
|---|---|
| 1 | Almost certain detection: continuous monitoring with alarm |
| 2–3 | High: regular condition monitoring catches degradation |
| 4–6 | Moderate: manual inspection might catch it |
| 7–8 | Low: difficult to detect until failure occurs |
| 9–10 | Almost impossible: no warning signs |
Putting It Together
RPN ranges from 1 to 1,000. A common threshold: any failure mode with RPN > 200 (or the top 10 highest RPNs) gets an action plan.
Example: FMEA for an Industrial Centrifugal Pump
Let's apply this to a real pump. Here's a simplified FMEA worksheet for a centrifugal pump handling cooling water:
| Component | Failure Mode | Effect | Cause | Current Controls | S | O | D | RPN |
|---|---|---|---|---|---|---|---|---|
| Mechanical seal | Leak | Fluid loss, environmental spill | Seal wear, dry running | Visual check monthly | 7 | 5 | 6 | 210 |
| Bearing | Seizure | Rotor lock, pump failure | Lube depletion, contamination | Grease every 500h, vib check yearly | 8 | 4 | 7 | 224 |
| Impeller | Erosion | Reduced flow, efficiency drop | Abrasive particles in water | Strainer cleaning weekly | 5 | 6 | 4 | 120 |
| Motor winding | Burnout | Complete pump stoppage | Overload, single-phasing | Overload relay | 9 | 3 | 3 | 81 |
| Coupling | Misalignment | Vibration, coupling wear | Installation error, settling | Visual alignment at install | 4 | 3 | 5 | 60 |
| Suction piping | Cavitation | Impeller damage, noise, low flow | Blocked strainer, low NPSH | Pressure gauge monitoring | 7 | 4 | 5 | 140 |
The two highest RPNs are bearing seizure (224) and mechanical seal leak (210). These get action plans first.
Recommended Actions
For bearing seizure (RPN 224):
- Install online vibration monitoring with alarm (improves Detection from 7 to 2)
- Switch to automated greasing system (improves Occurrence from 4 to 2)
- Target RPN after actions: 8 × 2 × 2 = 32
For mechanical seal leak (RPN 210):
- Install seal flush line pressure transmitter with low alarm (improves Detection from 6 to 2)
- Upgrade seal material from standard carbon to silicon carbide (improves Occurrence from 5 to 3)
- Target RPN after actions: 7 × 3 × 2 = 42
How FMEA Feeds Into Your Maintenance Strategy
FMEA is not a one-time exercise. The real value comes when you use the results to drive your maintenance decisions.
Preventive Maintenance
FMEA tells you what to check, how often, and why. If the analysis shows that bearing failure is driven by lubrication degradation every 500 hours, you schedule regreasing at 400 hours. If seal leaks are caused by pump deadheading, you add a minimum flow recirculation line and a procedure to never close the discharge valve fully during operation.
Every PM task in your system should trace back to a failure mode identified in FMEA. If a PM doesn't prevent or detect a specific failure mode, why are you doing it?
Predictive Maintenance
FMEA identifies the best condition monitoring technique for each failure mode:
| Failure Mode | Detection Technique | Sensor / Tool |
|---|---|---|
| Bearing wear | Vibration analysis | Accelerometer |
| Seal leak | Pressure trend monitoring | Pressure transmitter |
| Impeller erosion | Flow rate degradation | Flow meter |
| Motor winding degradation | Thermal imaging | IR camera |
| Cavitation | High-frequency vibration / noise | Ultrasonic sensor |
Instead of guessing which PdM technology to buy, FMEA tells you exactly what to monitor — and why.
Spare Parts Strategy
FMEA also informs your critical spares holding. High-RPN failure modes with long lead-time parts get added to your spare parts inventory. Low-RPN failure modes with easily available parts can be ordered on-demand.
How a CMMS Stores and Uses FMEA Data
FMEA generates a lot of information. A CMMS is the natural home for it.
Asset-Level FMEA Records
In a CMMS like OpexMX, you can attach FMEA worksheets directly to each asset:
- All failure modes are listed
- RPN scores are recorded and tracked over time
- Current controls are documented
- Recommended actions become work orders
- Target RPNs and actual results are compared after action completion
Triggered Work Orders
When FMEA identifies a preventive action — "replace seal every 6 months" — that becomes a scheduled PM in the CMMS. When it identifies a condition monitoring threshold — "alarm if vibration exceeds 4.5 mm/s" — that becomes a meter-reading task that auto-generates a work order when exceeded.
Failure Reporting Integration
When a technician closes a work order in the CMMS, they can tag the specific failure mode from the FMEA. Over time, this creates a feedback loop:
- The system tracks how often each failure mode actually occurs
- Actual occurrence rates are compared against the FMEA estimates
- If a failure mode was rated Occurrence = 3 but happens twice a month, the FMEA is updated with real data
This is how static FMEA becomes a living document. Without a CMMS, most FMEAs end up in a PDF on someone's desktop — never updated, never used.
The Bottom Line
FMEA is not a paperwork exercise. It's a practical tool for answering one question: what should we work on next?
Without FMEA, your maintenance strategy is driven by whoever complains loudest, or by whichever machine failed most recently. With FMEA, you have a data-backed priority list. You know which failures cost the most, which are most likely, and which you can actually detect before they happen.
Run one FMEA session on your most critical asset. You'll be surprised what you find — and how obvious the next steps become.
See how OpexMX helps you manage FMEA data and turn it into real work orders — built for Indonesian factory floors, not boardroom consultants.