6.4 Reviewing Reliability Performance & Adjusting Tactics

Reviewing Reliability Performance & Adjusting Tactics

Quick Answer: Reliability strategy requires a closed feedback loop: compare performance with expectations, investigate meaningful deviations, implement justified changes, and verify results. Those decisions support CMRP Function 3.7. When assets fail, organizations deploy a closed-loop Failure Reporting, Analysis, and Corrective Action System (FRACAS) and Root Cause Failure Analysis (RCFA) to identify not just the physical failure mechanism, but the latent organizational roots (procedural, cultural, procurement, and training deficiencies) that allowed the defect to exist. Verifying corrective action effectiveness and updating CMMS job plans ensures permanent defect elimination.

The Continuous Improvement Mandate in Asset Management

The Plan-Do-Check-Act (PDCA) cycle is one useful model for continuous improvement:

• Plan: Establish asset reliability baselines, conduct RCM/FMEA analyses, and formulate maintenance strategies (Chapters 5 and 6.1).
• Do: Implement precision lubrication, execute scheduled PMs, and deploy predictive condition-monitoring routes (Chapter 6.3).
• Check: Continuously measure performance metrics, track Mean Time Between Failures (MTBF), monitor unreliability costs, and log failure events.
• Act: Execute Root Cause Failure Analysis (RCFA) on significant failures, adjust maintenance tactics, update CMMS master data, and eliminate systemic organizational defects.

A common failure in maintenance organizations is treating equipment strategy as a static, one-time exercise. If a plant performs an RCM analysis in 2020 but never updates its maintenance tasks when operating contexts change or failure patterns emerge, reliability inevitably degrades. Reviewing performance and adjusting tactics is the mechanism that keeps asset strategies aligned with business goals.

---

The Closed-Loop FRACAS Process

A Failure Reporting, Analysis, and Corrective Action System (FRACAS) is a disciplined, closed-loop engineering workflow designed to capture failure data, prioritize investigations, execute root cause analysis, implement corrective actions, and track long-term effectiveness.

The 6 Stages of an Industrial FRACAS Workflow

1. Failure Reporting & Incident Capture: The moment an asset experiences an unscheduled breakdown, functional failure, or significant condition-monitoring alert, a failure record is generated in the CMMS. Standardized failure coding is mandatory: technicians and operators record the specific Object Part, Failure Mode, and Symptom rather than vague notes like "machine broken."
2. Screening & Criticality Triage: Not every breakdown warrants a 20-hour engineering investigation. The reliability team applies predetermined RCFA Trigger Criteria to screen incidents. Formal RCFA is triggered if an event:
   • Incurred an OSHA recordable injury, safety near-miss, or environmental permit exceedance.
   • Inflicted unscheduled production downtime exceeding a specific threshold (e.g., $> 4.0$ hours or $> $25,000$ in lost margin).
   • Generated direct maintenance repair costs exceeding a set limit (e.g., $> $15,000$).
   • Represents a recurring chronic failure (e.g., the same pump seal failing three times in six months).
3. Multidisciplinary RCFA Investigation: A cross-functional investigation team—comprising the reliability engineer, maintenance technician, equipment operator, original equipment manufacturer (OEM) technical rep, and metallurgist if needed—gathers physical evidence and reconstructs the failure timeline.
4. Corrective Action Formulation & Feasibility Review: Developing SMART (Specific, Measurable, Achievable, Relevant, Time-bound) corrective solutions. Solutions must focus on error-proofing, engineering redesign, or procedural changes rather than merely telling technicians to "be more careful."
5. Implementation & CMMS Alignment: Modifying physical machinery, altering PLC control logic, updating CMMS job plans, adjusting PM/PdM frequencies, and training personnel.
6. Closed-Loop Effectiveness Verification: Track the asset over a defined operating exposure appropriate to the failure opportunity. Keep the case open until the selected evidence shows implementation and the intended risk reduction; absence of recurrence alone does not prove permanent elimination.

Root Cause Failure Analysis (RCFA) Methodologies

Root Cause Failure Analysis (RCFA) is a structured, evidence-based engineering problem-solving process that seeks to determine the fundamental cause of a failure so that appropriate corrective actions can prevent recurrence. SMRP BoK 3.7 highlights four primary analytical methodologies:

1. The 5 Whys Methodology

Originally developed by Sakichi Toyoda for the Toyota Production System, the 5 Whys is an iterative interrogative technique that explores the cause-and-effect relationships underlying a particular failure by asking "Why?" five successive times:

• Strengths: Rapid, intuitive, requires zero specialized software, and highly effective for straightforward, single-root mechanical issues.
• Limitations & Pitfalls: Tends to isolate a single linear causal path, failing to capture complex multi-variable interactions. Unskilled facilitators often stop questioning prematurely when a human error is identified (e.g., "The operator failed to open the valve"), completely missing the underlying systemic reasons why the operator committed that error.

2. Ishikawa (Fishbone / Cause-and-Effect) Diagram

Developed by Kaoru Ishikawa, the Fishbone Diagram visually maps all potential contributing causes leading to a specific problem (the "fish head"), categorizing potential causes along the 6Ms of Manufacturing:

• Machine: Equipment design, metallurgy, mechanical wear, piping vibration, sensor calibration.
• Method: Standard operating procedures, maintenance job plans, lubrication routes, permitting processes.
• Material: Raw material viscosity, seal elastomer chemical compatibility, lubricant additive packages, spare parts quality.
• Manpower (Personnel): Craft skill levels, training gaps, fatigue, communication handovers between shifts.
• Measurement: Inaccurate pressure gauges, misplaced thermowells, sensor drift, lab test turnaround delays.
• Mother Nature (Environment): Ambient temperature extremes, humidity, coastal salt air, airborne abrasive dust.

3. Fault Tree Analysis (FTA)

Developed by Bell Telephone Laboratories for aerospace systems, Fault Tree Analysis is a deductive, top-down logic diagram that models the pathways leading to an undesired system-level failure (the top event) using Boolean Logic Gates:

• AND Gate: The output event occurs if and only if all input events occur simultaneously. In reliability engineering, AND gates represent redundant or multi-barrier protection systems.
• OR Gate: The output event occurs if any single input event occurs. OR gates represent single points of failure.
• Quantification: By assigning individual probability failure rates to the basic initiating events at the bottom of the tree, FTA calculates the overall mathematical probability of the catastrophic top event.

4. The Apollo Root Cause Analysis Method

Developed by Dean Gish, Apollo Root Cause Analysis is grounded in the Principle of Causality: every effect has at least two causes that must exist simultaneously—an Action (a dynamic event or trigger) and a Condition (an existing environmental or physical state):

$\text{Effect} = \text{Action (Moment in Time)} + \text{Condition (State of Being)}$

• Rather than forcing causes into rigid categories, the Apollo method constructs a visual Reality Chart where every cause-and-effect link must be supported by verifiable physical evidence. Cause-and-effect links should be supported by available evidence and tested as hypotheses. Uncertainty, missing evidence, and alternative explanations should be documented rather than hidden.

---

The Three Levels of Root Causes: Physical, Human, and Latent

A useful taxonomy separates physical mechanisms, human actions or decisions, and latent organizational conditions. Not every event will contain one cause in each layer, and the categories are an analysis aid rather than an official three-cause requirement:

1. Physical Root Causes (The Technical Mechanism)

• Definition: The tangible physical, chemical, or electrical mechanism by which the physical component degraded or fractured.
• Examples: Severe fatigue beach marks on a 4140 steel shaft; abrasive quartz three-body cutting wear on an impeller vane; thermal embrittlement of an ethylene propylene diene monomer (EPDM) O-ring; electrical insulation dielectric breakdown due to moisture tracking.
• Remedy: Physical roots are resolved by component redesign, material upgrades, or physical repairs. Addressing only the physical mechanism may restore the asset without controlling the conditions that allowed the mechanism to develop.

2. Human Root Causes (The Inappropriate Human Action or Omission)

• Definition: The inappropriate action, omission, decision, or procedural deviation committed by an individual (operator, technician, or engineer) that triggered the physical mechanism.
• Examples: A maintenance mechanic torqued cylinder head bolts using an uncalibrated impact gun without following the cross-star sequence; an operator bypassed an automated high-temperature alarm to prevent a nuisance trip; a buyer selected a non-OEM mechanical seal to save budget.
• Management Trap: Traditional reactive organizations stop here. They blame the mechanic, issue a disciplinary reprimand, and declare the RCFA "closed." Stopping at a label such as human error usually leaves system conditions unexplored; examine the task, interface, procedure, competence, incentives, supervision, and environment.

3. Latent Root Causes (The Organizational / Systemic Flaws)

• Definition: The underlying management systems, organizational policies, cultural norms, procurement rules, or training deficiencies that permitted the human error to occur or failed to catch it.
• Examples: The company does not provide calibrated torque wrenches or torque specifications in CMMS job plans; the training program does not certify technicians on precision bolting; the procurement incentive bonus rewards purchasing low-bid spare parts without engineering approval; management penalizes operators for production stops, driving them to silence alarms.
• The CMRP Principle: True root causes are always latent organizational roots. Sustained defect elimination is achieved only when management systems, procedures, tools, and cultural governance are fundamentally re-engineered to make the human error difficult or impossible to commit.

---

Root Cause Taxonomy Table: Physical vs. Human vs. Latent Roots

---

Verifying Corrective Action Effectiveness & Closing the Loop

A major vulnerability in maintenance programs is the "premature closure" of root cause investigations. Corrective actions are written on paper, filed in the CMMS, and marked "complete" the day a work order is signed off. Six months later, the exact same failure recurs.

To test whether corrective action was implemented and reduced risk, an organization can use a formal Corrective Action Verification Protocol combining leading and lagging evidence. The following review windows are illustrative:

• Leading Verification Indicators (Short-Term Auditing): Audited within 14 to 30 days of implementation to verify that the physical and procedural solutions were actually enacted in the field:
  • Have the updated CMMS job plans and torque specifications been loaded into the system?
  • Are technicians utilizing the new calibrated torque wrenches and recording verified torque values on work order closeout forms?
  • Has the automated PLC low-flow flush interlock been physically tested and trip-certified?
• Outcome verification indicators: Track over risk-appropriate operating exposure or site-defined milestones:
  • Has the asset operated without functional failure or unpredicted stoppage throughout the evaluation period?
  • Are predictive vibration amplitudes and ultrasonic decibels remaining within pristine baseline limits?
  • Has the Mean Time Between Failures (MTBF) for this asset class demonstrably increased?
• Final Sign-Off: The FRACAS case is formally closed only after the cross-functional reliability committee reviews post-implementation performance data and confirms that the failure mechanism has been eradicated.

---

Feeding RCFA Findings into the Living Maintenance Strategy

The ultimate measure of a mature reliability organization is its ability to institutionalize learning. RCFA findings must not remain buried in engineering PDF reports on a network share; they must dynamically feed back into core plant systems:

1. Updating CMMS Asset Data & Job Plans: Revise standardized job packages with verified bolt torque sequences, laser alignment tolerances, and required gasket materials. Attach photo-illustrated Standard Operating Procedures (SOPs) directly to work orders.
2. Recalibrating PM & PdM Task Frequencies: If an RCFA reveals that bearing degradation developed over an 8-week P-F interval, but the vibration monitoring route was scheduled every 12 weeks, a monthly route could be evaluated as a starting response. The final interval must account for P-F variability, measurement sensitivity, consequence, and the time needed to plan and act; half the observed interval does not guarantee detection.
3. Refining MRO Storeroom Inventory: Update the equipment Bill of Materials (BOM) in the CMMS. Eliminate low-grade non-OEM parts, re-specify upgraded materials (e.g., fluoroelastomer O-rings instead of Buna-N), adjust Min-Max stocking levels, and establish proper parts preservation protocols (e.g., motor shaft rotation routines in storage).
4. Updating Engineering Design & Procurement Standards: Feed lessons learned back into the capital project procurement standards (closing the loop back to Section 6.1). If a specific pump casing design proved prone to chronic cavitation erosion, update the plant procurement specification so that future capital expansions specify upgraded metallurgy and improved suction piping geometries from day one.