1. Maintenance is an economic decision wearing overalls
Every maintenance philosophy is a bet about money. Spend too little and the plant fails at the worst commercial moment; spend too much and the O&M budget eats the margin the plant exists to earn. The maintenance manager's real job is to hold availability, reliability and cost in a deliberate balance — and to be able to show, with records, why the balance sits where it does. This guide walks through the components of that balance as they are actually practised in operating plants: the mix of maintenance types, the overhaul rhythm, spares, systems, condition monitoring and the safety framework that governs all of it.
2. The maintenance mix
No serious plant runs on one maintenance type. The craft is in the proportions, chosen per asset by criticality and failure behaviour:
- Corrective (run-to-failure) — legitimate for cheap, non-critical, redundant assets where failure has no safety or production consequence. Deliberate, not lazy.
- Preventive (time-based) — servicing and replacement at fixed intervals. Right for age-related wear with a predictable life; wasteful and occasionally harmful (infant mortality after intrusive work) when applied blindly.
- Predictive / condition-based — intervention when measurements say the asset is degrading: vibration trends, oil analysis, thermography, performance deviation. The highest-value work in most plants, because it converts surprises into planned jobs.
- Proactive — eliminating the failure cause itself: alignment precision, lubrication discipline, root-cause analysis that ends repeat failures.
The full argument — including what a sensible percentage split looks like — is in The Maintenance Mix.
3. The overhaul rhythm
Major rotating equipment lives on an hours-based calendar set by the manufacturer and hardened by experience:
| Machine | Typical rhythm | Scope highlights |
|---|---|---|
| Reciprocating gas engines | Minor overhaul ~30,000 h; major overhaul ~60,000 h | Top-end and power-unit work at minor; crankshaft, bearings, liners and full inspection at major |
| Aeroderivative gas turbines | Hot section ~25,000–30,000 h; major per OEM programme | Hot-section exchange, HPC work; Level-3 scope at authorised facilities |
| Steam turbines | Level B (minor) and Level C (major) inspections on multi-year cycles | B: valves, bearings, auxiliaries. C: full casing lift, rotor and diaphragm inspection |
| HRSGs / boilers | Statutory and condition inspections, e.g. to API 510 for pressure equipment | Tube condition, headers, drums, refractory and safety valves |
An overhaul is a project, not a work order: scope frozen months ahead, spares staged and verified, contractor and OEM support contracted, a critical path built, and return-to-service criteria agreed before the first bolt turns. Inside a 60,000-hour overhaul describes one run end to end — the complete 32-engine fleet programme at SGTPS — and the LM6000 hot-section exchange covers the aeroderivative case. The bench tools for overhaul work are in the mechanical calculator set: bolt torque, alignment shims, sling tension and thermal expansion.
4. Spares and inventory: the quiet budget
A mid-size plant's warehouse quietly holds more capital than most single projects — in my own case, responsibility for over PKR 500 million of inventory. Inventory discipline has two failure modes: stockouts that extend outages (a single missing gasket set can hold a unit), and overstock that fossilises cash in slow-moving parts for machines that may never need them. The working system is criticality-led: insurance spares for catastrophic, long-lead failures; min-max levels for consumables driven by actual usage; periodic slow-mover reviews; and overhaul BOMs verified against the warehouse before the outage, not during it.
5. Condition monitoring: seeing failure early
Condition monitoring is the sensory system of the maintenance strategy. The core techniques in a power plant are vibration analysis on rotating machines (severity judged against ISO 20816 zones — see the Vibration Severity Evaluator), lubricating oil analysis, thermography of electrical and hot equipment, borescope inspections of gas turbine hot sections, and performance trending — output and heat rate corrected for conditions, watched for drift (the Degradation Tracker does this job). Bearing decisions rest on calculated life versus condition evidence together: the Bearing L10 Life Calculator gives the statistical expectation that monitoring then confirms or overrules.
6. Systems: the CMMS and SAP PM
None of the above survives on clipboards. A computerised maintenance management system carries the asset register, the preventive plans, the work-order flow and the permit interface — and it is only as good as its structure: a functional-location hierarchy that mirrors the plant, equipment records with the right technical data, and maintenance plans people actually follow. Having led a plant-wide SAP PM implementation, I can state the lesson plainly: the software project is 20% of the work; the hierarchy, the data discipline and the work process are the other 80%. The story, including what we got wrong first, is in SAP PM: The System Behind the Wrenches.
7. Asset management and ISO 55001
Above the work-order level sits asset management: the organisation's system for extracting value from assets across their whole life — acquisition, operation, maintenance, renewal, disposal. ISO 55001 formalises it: an asset management policy tied to business objectives, a strategic asset management plan, and life-cycle decision-making with documented trade-offs. For a plant engineer the practical meaning is simpler: maintenance stops being a cost centre defending its budget and becomes a managed contributor to the business case — with disposal, like the BQPS-I units I managed through sale and dismantling, treated as a planned life-stage rather than an afterthought.
8. Safety and process safety
Maintenance is where process safety is won or lost, because maintenance opens the plant up: breaking containment, lifting loads, working on energised systems, entering confined spaces. The framework in serious plants combines permit-to-work discipline, lock-out/tag-out, lifting plans with calculated sling tensions (not estimates), and a process safety management system of the DuPont type — operating procedures, management of change, mechanical integrity and contractor control. Every technique in this guide is subordinate to that framework; an overhaul finished fast and hurt someone is a failed overhaul.
9. Measuring whether it works
What gets measured gets managed, and maintenance has honest metrics: equivalent availability and forced outage rate for the plant; planned-versus-unplanned work ratio for the strategy; schedule compliance and backlog age (in crew-weeks) for execution; repeat-failure rate for quality; and maintenance cost per MWh for economics. Reviewed together, monthly, they tell you which part of this guide your plant is currently weak in.
10. Frequently asked questions
What is the difference between preventive and predictive maintenance?
Preventive maintenance acts on a schedule — time or running hours — regardless of the asset's actual state. Predictive maintenance acts on evidence: a measurement trend (vibration, oil debris, temperature, performance loss) that shows degradation in progress, so work is done when the asset needs it, neither early nor late.
How often should a gas engine be overhauled?
For large gas engines of the Jenbacher 620 class, the typical programme is a minor overhaul around 30,000 operating hours and a major overhaul around 60,000 hours, adjusted by condition findings, fuel quality and starts history. The manufacturer's schedule is the starting point, not the whole decision.
What is a Level B versus Level C steam turbine overhaul?
Level B is a minor inspection: bearings, valves, governors and auxiliaries, without lifting the main casing. Level C is the major: casing opened, rotor lifted and inspected, diaphragms and blades examined and repaired — a multi-week outage with a correspondingly larger scope, crew and budget.
What does ISO 55001 add beyond a good CMMS?
A CMMS manages work. ISO 55001 manages the system around the work: policy, life-cycle planning, risk-based decisions and alignment between asset spending and business objectives. Plants usually need the CMMS first; 55001 is how its output gets connected to the boardroom.
11. Work with me
I have managed maintenance for reciprocating fleets, an aeroderivative combined cycle complex and multi-plant portfolios — including LM6000 Level-3 and steam turbine Level B & C overhaul programmes at a 248 MW combined cycle complex, and PKR 500M+ of inventory. The services page describes the maintenance audits and overhaul support I take on independently, or reach me via the contact page.
Saad Hafeez is a Chartered Mechanical Engineer (CEng MIMechE), Deputy General Manager – Power Plants at OMS, formerly Maintenance Manager at K-Electric and Senior Manager Engineering & Operations at Omni Group.
Engineering calculators | Guide: Performance Testing | Services