Power Plant O&M KPIs: How to Judge the Team Running Your Plant

Availability: The Headline Measure

Availability measures the share of time a unit or plant could produce power, whether or not it was asked to. It is the first number owners and lenders look at, and the one most often argued over, because the definition decides who carries the blame for each hour lost.

A sound definition separates planned outages, agreed in advance and carried out on schedule, from forced outages caused by failures. It also deals with events outside the operator’s control, such as fuel not delivered by the owner, grid failures or force majeure. Many contracts use equivalent availability, which counts partial derates as well as full outages, so a unit running at half its rating is not reported as fully available.

Reliability: Forced Outages and Failed Starts

Availability can look healthy on a plant that trips often but restarts quickly. Reliability indicators catch that pattern.

For standby and peaking plants, starting reliability is often more important than availability, because the plant earns its keep in the minutes after a call to run. For baseload plants, forced outage rate and time between failures tell the owner whether maintenance is preventing problems or only reacting to them.

Efficiency: Fuel and Lubricating Oil

Fuel is the largest running cost of most thermal plants, so efficiency indicators carry real money. Specific fuel consumption, the fuel burnt per unit of electricity, should be tracked per unit and for the whole plant, corrected for fuel quality and ambient conditions where the contract allows.

Lubricating oil consumption is a sensitive early warning for engine wear. A rising trend can point to worn rings, liners or valve guides long before a failure. Combined with oil analysis results, it gives a clear picture of engine condition. Our article on oil analysis in African service programmes shows how the two are read together.

On hybrid plants, add solar yield against expected yield, battery round-trip losses and the amount of fuel the renewable share saved. Those figures show whether the control system is using the cheapest energy first. USP&E’s approach to hybrid plant operation relies on exactly that kind of reporting.

Maintenance Discipline

Indicators of maintenance quality show whether the provider is protecting the asset for the long term or running it hard to hit this month’s numbers.

• Planned maintenance compliance: the share of scheduled tasks completed on time.
• Backlog: open work orders, ageing and priority.
• Repeat failures: the same fault on the same equipment within a set period.
• Spare parts availability: critical spares on the shelf when needed.
• Overhaul adherence: major inspections carried out at the maker’s intervals.

These need a working maintenance management system. Multi-site owners benefit from one system across all plants, as discussed in our page on a shared CMMS for African fleets.

Safety, Environment and People

No performance report is complete without safety. Lost time injuries, recordable incidents, near misses reported and permit-to-work compliance show the safety culture on site. A rising number of near-miss reports is often a good sign, since it means people are speaking up.

Environmental indicators cover spills, emissions exceedances, waste oil disposal and noise complaints. Many plants also report local employment and training hours, which matter to host communities and governments. Our article on training local operators explains how skills transfer is planned and measured.

Setting Targets Without Creating Bad Incentives

Indicators shape behaviour, sometimes in ways nobody intended. A provider paid only on availability may defer maintenance to keep units running, storing up problems for the next contract period. A provider penalised heavily for every forced outage may hesitate to take a unit offline when a fault first appears.

1. Use a balanced set of indicators so no single number can be gamed.
2. Agree baselines from real plant history before setting targets.
3. Define exclusions precisely, especially for fuel supply, grid events and owner instructions.
4. Link bonuses and penalties to indicators the provider genuinely controls.
5. Review targets each year as the plant ages and the duty changes.

A contract that is fair to both sides survives the bad months as well as the good ones.

Reporting That People Actually Read

Monthly reports should open with a single page: the key indicators against target, the trend over the past year and a short explanation of anything outside limits. Supporting detail follows for those who need it. Raw data should be available to the owner, not only the provider’s summary.

Daily and weekly reporting suits operations staff; monthly and quarterly reviews suit management and lenders. Remote monitoring makes near real-time dashboards possible, but a dashboard without a regular review meeting changes little.

Linking KPIs to the O&M Contract

Measures only change behaviour when they have consequences. In many O&M contracts a share of the operator’s fee is tied to availability or reliability, with bonuses for beating targets and deductions for missing them. That works well when the definitions are precise: what counts as an outage, which outages are excluded because they were caused by the owner or the grid, and how the figures are verified.

Disputes usually arise from vague wording. Agree the data source, the calculation method and the reporting period before the contract is signed, and give both parties access to the same raw records. An independent review once a year keeps confidence high on both sides.

Data Quality Behind the Numbers

A KPI is only as good as the data under it. Running hours, start counts, fuel meter readings and outage logs must be captured consistently, ideally straight from the control system rather than typed in by hand at the end of a shift. Meters should be calibrated on a schedule, and any gap in the record should be flagged rather than filled with an estimate.

Simple cross-checks catch most errors: energy generated should match fuel burned at a believable efficiency, start counts should match the alarm log, and outage hours should add up with running hours to the hours in the period. When a figure fails a check, fix the data source first and the report second.

What to Ask a Prospective O&M Provider

When comparing power plant service companies, ask each one to show real reports from plants it runs today, with client names removed if necessary. Ask how it defines each indicator, which system holds the data and who checks it. Ask what happened on a plant where performance fell short and how it recovered.

USP&E runs plants under operations and maintenance contracts built on these measures, often after delivering the plant itself under its EPC service, which keeps design knowledge with the team that runs the station.