Common Causes of Downtime in Power and Energy Plants

In power and energy plants, downtime is measured as forced outages, and the common causes are boiler tube failures, turbine and generator faults, electrical equipment failures, and control-system trips. Reducing them means monitoring the condition of critical rotating and electrical equipment and planning outages around the assets most likely to fail.

How downtime is measured in power

Power generation separates downtime into planned outages, scheduled for maintenance, and forced outages, the unplanned failures that take a unit offline without warning. The industry tracks unavailability as the forced outage rate, most often the equivalent forced outage rate (EFOR), which captures how much of the time a unit could not generate when it was needed, weighting partial deratings as well as full outages. For context, coal-fired units run a weighted EFOR of around 12 percent, and the rate has been climbing as the fleet ages. Forced outages are the costly ones, in lost generation, in higher-cost replacement power bought at short notice, and in grid-reliability and contractual terms. Planned outages, by contrast, are scheduled when demand and prices are low and the work can be staged in advance, so the strategic aim is to move as much failure as possible out of the forced column and into the planned one.

The common causes of forced outages

The causes are well documented. According to the National Energy Technology Laboratory, boiler tube failures account for roughly 52 percent of forced outages at coal plants, followed by balance-of-plant issues at about 15 percent, steam turbine problems at 13 percent, generator failures at 12 percent, and human error at 4 percent. The same pattern shows up in NERC's generating availability data across thousands of units: more than half of forced outages at thermal plants trace to boiler tube failures.

• Boiler and tube failures. In thermal plants, boiler tube failures are consistently the single leading cause of forced outages. Tubes fail through corrosion and corrosion fatigue, fireside and steam-side erosion, creep from sustained overheating, and weld defects, and because a single tube leak can force a unit down, they account for the largest share of lost thermal generation. Repeat failures make up roughly half of all boiler tube failures, so the same tubes tend to fail again unless the root cause is found.
• Turbine and generator faults. The largest rotating assets in the plant take a unit fully offline when they fail. Common modes include vibration from imbalance or misalignment, blade and bearing damage, steam-path problems, and on the generator side, winding insulation degradation and excitation faults.
• Electrical equipment. Transformers and switchgear fail mostly through insulation breakdown, partial discharge, and overheating, and when they let go they can trip a unit hard and cause collateral damage. These failures usually develop gradually, which makes them catchable.
• Controls and instrumentation. A healthy mechanical plant can still be tripped by a control-system fault, a failed sensor, or a protection device that misoperates. Spurious trips from instrumentation are a meaningful and often underappreciated share of forced outages.
• Balance of plant. The supporting systems, feedwater and cooling pumps, fans, fuel handling, condensers, and cooling, are less glamorous than the turbine but stop generation just as completely when they fail.

Why availability matters so much

A generating unit earns only when it is available, and the cost of not being available is steep. A forced outage commonly runs on the order of $125,000 per hour in lost generation and replacement power, before counting the premium on emergency parts and labor. Capacity markets and grid obligations add to that, often paying for availability and penalizing unavailability directly, so every point of forced-outage rate carries a clear and quantifiable cost. The economics of prevention are just as clear: corrective maintenance after a failure runs roughly $17 to $18 per horsepower a year, while preventive and predictive maintenance run about $7 to $13, a saving of up to 45 percent before you even count the avoided downtime. That math is why the assets whose failures cause forced outages, the boiler, turbine, generator, and key electrical equipment, are the natural focus for reliability spend. Tracking MTBF and availability on them shows where reliability is eroding before it becomes an outage.

How to reduce the outages

Much of this is preventable: by one estimate, 43 percent of plant incidents are mechanical failures that monitoring could have caught in time. Reducing forced outages is about catching those failures on the critical assets early enough to deal with them on plan.

• Monitor critical rotating equipment. Vibration and condition monitoring catch the bearing, balance, and blade problems on turbines, generators, and major pumps weeks before they force a unit down.
• Watch the electrical assets. Thermography and oil analysis on transformers and switchgear catch insulation and overheating problems before they trip the unit.
• Prioritize by consequence. Use asset criticality to focus monitoring and spares on the assets whose failure causes the most lost generation, rather than spreading effort evenly.
• Move critical assets to predictive maintenance. On the highest-consequence equipment, shift from calendar-based service toward predictive maintenance driven by condition, and use a reliability-centered approach to target the specific failure modes that actually cause outages.
• Root-cause the repeat offenders. Because repeat failures are about half of all boiler tube failures, fixing the cause rather than patching the symptom is what takes a chronic failure off the list for good.
• Track reliability trends. Watch MTBF and MTTR on critical assets, so a rising failure rate surfaces before it becomes a forced outage.

Converting forced outages into planned work

The single most valuable move in power reliability is shifting failures from the forced column to the planned one. A forced outage happens at the worst time and at full cost; the same repair done during a scheduled outage happens when demand and prices are low and the crew, parts, and sequence are staged in advance. Condition data is what makes the shift possible. When monitoring shows a boiler tube wall thinning, a bearing trending, or a transformer heating, you can fold that repair into the next planned outage instead of waiting for the failure. Industry programs built on exactly this idea, like EPRI's boiler tube failure reduction work, have measurably cut availability losses. The goal is not zero maintenance. It is to make the maintenance happen on your schedule, not the failure's. The broader playbook is in reducing unplanned downtime.

Common mistakes

• Focusing only on the turbine. Boiler tubes, electrical equipment, and balance of plant cause a large share of forced outages and deserve attention too.
• Patching repeat failures. Repeat boiler tube failures are about half of all tube failures, so fixing the symptom without the root cause guarantees the next outage.
• Ignoring slow electrical trends. Transformer and switchgear failures often build gradually and are catchable with thermography and oil analysis.
• Treating planned outages as fixed scope. Condition data should shape what gets done during an outage, so the right components are addressed.
• Under-monitoring balance of plant. The unglamorous pumps and fans still stop generation when they fail.

From outage causes to reliability decisions

Power plant reliability data spans condition monitoring, control systems, and maintenance records. Knowing which critical asset is trending toward a forced outage, and whether to act now or fold the work into the next planned outage, is the hard part.

SteelTree connects to those systems and turns them into decisions: which assets are eroding toward a forced outage, how their risk maps to lost generation, and the next action to take, with the reasoning attached. You keep your existing systems. SteelTree sits on top as the decision layer.

See how SteelTree turns operational data into decisions →