Steam Turbine Overhaul Cost: What Drives the Price and the Lead Time

Steam turbine overhaul cost cannot be pinned to a reliable number until the casing is open, because the largest swing in the bill is what the inspection finds. What you can get beforehand is a firm price for the base scope, a rate card for whatever the inspection uncovers, and a dated lead time for each long-delivery part. Published prices are scarce, but two give a sense of scale: a 2016 Power Engineering article says major outages pass $1 million very quickly, and a 2025 public tender by Verdo, owner of the Randers power plant in Denmark, estimated DKK 8 million, about EUR 1.07 million at the krone's euro peg, for the major overhaul of a 48 MW steam turbine and generator.

Neither figure is a budget. The 2016 one is a decade old and an order of magnitude at best. The 2025 one is an owner's estimate before bids, excluding VAT, for a set overhauled in 2017 and given a new vane carrier in 2020, and the work is meant to carry it to an expected end of life in 2036. We give no price per megawatt, because two data points do not make a rate.

Why Nobody Can Quote an Overhaul With the Casing Closed

Inside a closed casing you cannot see blade erosion, seal clearances, diaphragm distortion or rotor cracking. Stage pressures, vibration and bearing temperatures tell you where to look, but they do not measure damage. So the first price a vendor gives is for a scope: open, clean, inspect, measure, report and close. Everything after the report is a second price.

The first price can be fixed. The second can only be bounded, which is why the two belong on separate lines of the budget.

Our steam turbine overhaul guide covers when to open a machine, and the steam turbine repair guide covers fault-driven work. This one starts at the purchase order.

What Drives Steam Turbine Overhaul Cost

Seven things set the price. Each one changes either the work or the time the machine is out.

Inspection level. A closed-casing inspection takes days. A major overhaul lifts the casing, pulls the rotor and takes weeks. Labour and crane time scale with how far you take the machine apart.

Unit size and stage count. Every stage adds a blade row, a diaphragm, seals and a set of measurements. Size also sets the crane, the laydown area, and the balancing machine and workshop that can take the rotor.

Findings. Discovery work is the largest swing in the total. Seals that need renewing are routine. A diaphragm that has dished, or a crack indication in the rotor, changes the job.

Shop or site. Shop work gives controlled heat treatment and machining, and costs freight and a shop slot. Site work saves the freight and limits the process control. EPRI's 2012 summary of repair methods says welding erosion shields in place is an option when erosion is not too severe, but the traditional advice is a factory weld so the heat treatment can be done properly. Decide by the class of repair, not by the day rate.

OEM or independent. The 2016 Power Engineering article says refurbishment is no longer a manufacturer's monopoly, and recommends a facility qualification review in the request for quotation. The manufacturer holds the drawings and the design basis. An independent often has a shorter queue on commodity parts. Ask both to price the same scope in the same line items. MIMAH is independent of any one manufacturer, so we can specify a made-to-drawing part where it will do and the manufacturer's part where it is right.

Location and logistics. In a remote location the larger cost is waiting: for a crane, for customs clearance on a rotor, for a specialist who has to fly in. MIMAH has an engineering and supply chain hub in Egypt and does reverse engineering and local manufacture of critical spares, which can take a part off the customs path.

Lost generation. This can be the biggest number of all, so it gets its own section.

The seven drivers of steam turbine overhaul cost. Each one changes either the work or the time the machine is out. Source: EPRI, Steam Turbine Monitoring, Inspection and Repair, 2012; MIMAH engineering practice.

Lost Generation Can Cost More Than the Work

Take a hypothetical 10 MW set at an 80 percent average load. It generates 10 × 0.8 = 8 MW, so 8 × 24 = 192 MWh a day.

Suppose each megawatt hour the plant does not generate costs it $150, whether in diesel burned, power bought or product not made. That figure is an input chosen to show the arithmetic; use your own. One day out then costs 192 × $150 = $28,800. A six week outage is 42 days, so 42 × $28,800 = $1,209,600. A one week overrun adds 7 × $28,800 = $201,600.

On those inputs the generation lost over six weeks is the same order of magnitude as both published figures above. Buyers underrate the overrun: a week lost to a missing part can cost more than the gap between two quotes. In our view a contractor who is a little dearer per hour but holds the parts and can run a second shift is often the cheaper option.

Lump Sum, Time and Materials, Target Cost or a Long Term Service Agreement

Ontario Power Generation's 2016 filing to the Ontario Energy Board is a clear public description of how pricing models share risk. It concerns the Darlington nuclear refurbishment and not a turbine outage, but the definitions are the ordinary ones.

Lump sum. The filing calls this fixed or firm price. The contractor does defined work inside a set budget and time, and the price moves only in specified circumstances or when the owner changes the scope. A lump sum suits the base scope and not the findings, because a contractor who must fix a price on unknown damage prices the worst case, and you pay for damage that never turns up.

Time and materials. The filing calls it reimbursable costs or cost plus mark-up: the contractor is paid actual labour and materials, with percentage mark-ups for overhead and profit. It suits the findings, provided the rates and mark-ups are agreed before the casing opens and the owner approves each item as it is found. Left unchecked it rewards slow work, so sign off time sheets daily.

Target cost. The contractor is paid its allowed actual cost plus a fixed fee, and both sides share savings below the target and overruns above it. The filing says high complexity and uncertainty made passing significant pricing risk to the contractor less commercially feasible. For one turbine, target cost is more machinery than most owners need. It suits a multi-unit outage where both sides will open their books.

Long term service agreement. An LTSA ties the owner to a supplier, usually the manufacturer, for a series of inspections and overhauls. It buys budget certainty and priority for parts and crews, and the risk sits in the terms. A 2003 Power Engineering paper by two Troutman Sanders lawyers lists the clauses that bite. It is about gas turbines and it is more than 20 years old, but the contract logic carries across. Look for the right to cancel early. Look at how the term stretches if the unit runs fewer hours than planned: their example turns a term of about 12 years into more than 20. Look at parts fitted near the end of the term, when the supplier's incentive to fit its best parts fades and the owner may be left with a limited parts warranty that can expire in as little as a year. And look at liability caps tied to what you have paid, which shrink when the unit runs less.

Here is what we would do. Sign a lump sum for the defined base scope and a rate card for findings, with the approval routine written into the contract. Take an LTSA only if you can leave it.

How four pricing models share risk. Source: Ontario Power Generation, filing to the Ontario Energy Board, 2016; Power Engineering, 2003; MIMAH engineering practice.

How to Budget a Base Scope Plus a Discovery Contingency

The method has four steps. Get a firm price for the base scope. List the findings your evidence says are possible, each with a probability and a cost as a share of that base scope. Multiply and add to get the expected discovery cost. Carry that as the contingency, and keep a separate reserve for the tail.

Here is an example with invented numbers. Call the base scope price 100 units, and suppose your evidence suggests five possible findings. Seal and gland renewal beyond plan has a probability of 60 percent and a cost of 8 units, so its expected cost is 0.60 × 8 = 4.8 units. Diaphragm repair on two stages: 0.35 × 15 = 5.25 units. Blade row repair or replacement: 0.15 × 30 = 4.5 units. Bearing rework: 0.25 × 6 = 1.5 units. Rotor weld repair: 0.10 × 40 = 4.0 units.

The sum is 4.8 + 5.25 + 4.5 + 1.5 + 4.0 = 20.05 units, so the contingency is about 20 units on a base of 100. Do not copy that 20; the probabilities are invented, and yours come from your own records. An expected cost is a midpoint and not a ceiling: if these five findings are independent, the total lands above 20 units about 42 percent of the time. All five together would cost 8 + 15 + 30 + 6 + 40 = 99 units, and the rotor repair alone, at 40 units, is twice the contingency. The contingency covers the average outage. The decision rules, who may approve what and how fast, cover the tail.

A worked discovery contingency using invented figures. Each probability multiplied by its cost gives an expected cost, and the five add to 20.05 units on a base scope of 100. Source: MIMAH engineering practice.

Give each finding a duration too, because on the earlier inputs every extra day out costs $28,800.

Then check the plan from the top down. The US EPA's 2015 catalogue of combined heat and power technologies says steam turbine maintenance typically costs below $0.01 per kWh. Its table gives $0.009 for a 3 MW back-pressure set and $0.006 for a 15 MW one, at normal service intervals over five years and excluding parts. The hypothetical set, at 8 MW for 8,000 hours a year, generates 64,000 MWh, or 64 million kWh. At those rates that is $384,000 to $576,000 a year, about $1.9 million to $2.9 million over five years before parts. That is a 2015 US check on the scale of the whole maintenance budget, not the price of one overhaul.

Better evidence narrows the probabilities. Our steam turbine inspection checklist lists what to measure, a condition monitoring programme turns it into trends, and plant shutdown and turnaround planning covers building the outage around the scope.

Lead Times by Component Class

Nobody who has not seen your drawings can give a lead time in weeks for a diaphragm or a blade row. The useful logic is to treat each class of part by what limits its delivery.

Stock items. Gaskets, fasteners, oil, standard bearings and instrument spares. Order against the base scope.

Parts made to drawing. Seal strips, packing rings, bearing shells and bushings. The limit is whether drawings exist. If they do not, add the time to measure the worn part and reverse engineer it.

Fabricated stationary parts. Diaphragms and nozzle rings are welded assemblies that need stress relief and finish machining, then a check of the throat openings. Drawings, material and a shop slot set the date.

Blades. Machining is no longer the slow step. The 2016 Power Engineering article says blade rows can be made in days rather than weeks if the right bar stock is on site, and that material lead time and production capacity now drive the schedule. In our view a vendor can only commit either one against a purchase order, so the clock starts on the day you order.

Forgings and large castings. A rotor forging is usually the longest wait. A 2010 POWER magazine case, written by staff of the contractor that did the work, describes a utility whose 100 MW unit showed ultrasonic indications near the rotor bore in June 2008. The contractor built a replacement in six months by welding two stock forgings together, to avoid waiting for a custom one. Six months is that project's number, not a quote.

So order now what the outage cannot proceed without and what your evidence says you will probably need. Leave the rest until the report, and put a price on the delay that waiting would cause so the choice is visible. Ask each vendor for a dated delivery in writing, counted from a stated purchase order date, with its assumptions listed: drawings, material and shop capacity.

Warranty and Acceptance Terms to Agree Before You Sign

Start date. Warranty can run from completion of the work, return to service or first synchronisation. Take the latest one the vendor will accept.

Length and scope. Ask for workmanship and parts warranties written separately, and for the reused parts to be named. We found no published norm for an overhaul warranty. The 2003 paper's one year applies to end-of-term parts under a gas turbine LTSA, so read it as a prompt to check and not as a benchmark.

Acceptance tests. Set vibration and alignment criteria for mechanical condition, and name the test code and edition for performance. ASME PTC 6-2004, reaffirmed in 2014 and now on stabilised maintenance, is the steam turbine performance code, and IEC 60953-3:2022 covers thermal verification tests of retrofitted turbines where the work changes hardware.

Liability cap. A cap smaller than the cost of opening the machine a second time protects the vendor and not you.

Handover documents. As-found and as-left clearances, non-destructive test reports and balance records. Without them the next overhaul starts blind.

Questions Buyers Ask About Overhaul Quotes

Can a vendor supply seal kits and send people to site? Yes. Ask for the seal set, the supervisor day rate and the technician day rate as separate priced lines; they double as the rate card for findings.

What does it cost per stage to refurbish LP blades? There is no figure we would stand behind before inspection. It depends on how far the erosion goes, the repair route and what crack testing shows. Ask for a price per blade for each route and for replacement.

How far ahead should we plan? Start about 18 months before a major overhaul. Siemens Energy's service page calls for a planning meeting on scope, spares and schedule at about that point, as our overhaul guide notes.

What to Do Before You Ask for a Quote

Gather the evidence first: the last inspection report, vibration and bearing temperature trends, stage pressures, start count, oil and steam chemistry records. Then freeze the base scope and ask every vendor for the same line items, plus a rate card for findings. Order the long-lead parts the outage cannot proceed without. Last, write down who can approve a finding on site, and how fast.

MIMAH's industrial engineering team is independent of any turbine manufacturer, and has an ongoing rehabilitation of a 14 MW steam turbine in Nigeria. If you have an outage to scope or quotes to compare, get in touch.