Two proposals land on your desk for the same 2 MW steam turbine project, priced within NGN 250 million of each other. The condensing machine makes its full output no matter what the process does. The back pressure machine makes power only when your plant is drawing steam. Most procurement committees pick the first one, and on a plant with a firm process steam demand that is usually the more expensive mistake.
The back pressure vs condensing steam turbine decision is not really a decision about turbines. It is a decision about where the heat in your steam ends up. One machine passes that heat on to your process. The other throws roughly two thirds of it into the air over a cooling tower. Everything else, capex, water consumption, staffing and operating flexibility, follows from that single difference.
What each machine does with the heat you paid for
A steam turbine converts a pressure drop into shaft work. The size of that drop, from throttle conditions down to exhaust, sets how much power you get per kilogramme of steam. That is the whole mechanism, and both machine types work the same way.
A back pressure turbine exhausts above atmospheric pressure, typically between 2 and 15 bar absolute, straight into your process header. It is functionally a pressure reducing valve that produces electricity on the way through. Steam you were going to throttle anyway does useful work first, then goes on to the steriliser, the dryer, the evaporator or the jacketed vessel exactly as before.
A condensing turbine exhausts into a vacuum, normally 0.08 to 0.15 bar absolute, maintained by a surface condenser and a vacuum system. That much larger pressure drop gives roughly three times the power per tonne of steam. The price is that the latent heat in the exhaust, around 2,200 kJ for every kilogramme, has to be rejected to cooling water and lost.
Put numbers on it. Steam at 45 bar and 440 degrees Celsius expanded to a 10 bar process header yields in the region of 70 to 80 kW for every tonne per hour of flow. The same steam expanded to 0.10 bar absolute yields around 220 to 240 kW per tonne per hour after auxiliaries. Three times the electricity, and none of the heat left over for your process.
That is why total fuel utilisation splits so hard. A back pressure installation running against a real steam load commonly reaches 72 to 82 percent of the fuel energy usefully employed. A condensing installation lands at 25 to 30 percent, which is in the same territory as a simple cycle gas turbine. If your site has no process steam demand, that trade is unavoidable. If it does, you are burning fuel to make heat and then paying again to get rid of it.
If you are sizing a steam turbine against an existing boiler and process load and want the arithmetic done on your actual figures, request a technical proposal and we will start from your steam balance rather than a nameplate.
Back pressure vs condensing steam turbine: the numbers that decide it
The table below is indicative for industrial machines in the 1 MW to 10 MW class at a Nigerian site. Costs cover the turbine island, meaning turbine, gearbox, generator, condenser and cooling system where applicable, controls and switchgear. Boiler, fuel handling and civils sit outside it. Figures move with steam conditions, exchange rate and site access, so treat this as a planning envelope rather than a quotation.
| Item | Back pressure | Extraction-condensing | Full condensing |
|---|---|---|---|
| Exhaust pressure | 2 to 15 bar a, into process header | 0.08 to 0.15 bar a, plus controlled extraction | 0.08 to 0.15 bar a |
| Power per tonne/hour of steam | 50 to 90 kW | 90 to 200 kW, depends on extraction split | 200 to 240 kW |
| Total fuel utilisation | 72 to 82% | 45 to 70% | 25 to 30% |
| Cooling water makeup | None | 1.5 to 2.5 m3 per MWh condensed | 1.5 to 2.5 m3 per MWh |
| Auxiliary load | 1 to 2% of output | 5 to 9% | 6 to 10% |
| Installed cost per kW | NGN 550k to 900k (USD 355 to 580) | NGN 1.1m to 1.8m (USD 710 to 1,160) | NGN 1.3m to 2.1m (USD 840 to 1,355) |
| Output follows | Process steam demand | Electrical demand, within extraction limits | Electrical demand |
| Plot area, relative | 1.0 | 1.9 to 2.4 | 1.8 to 2.3 |
| Operating attention | Lowest | Highest | High |
| Best fit | Firm year-round process steam | Swinging steam demand, flat power demand | No process steam demand |
USD equivalents are converted at an indicative rate near NGN 1,550 to the dollar and will move.
Read the cost per kW line together with the output line. A back pressure set is cheap per kilowatt because most of the plant a condensing machine needs simply is not there. No condenser, no cooling tower, no circulating water pumps, no vacuum system, no condensate polishing. That is fewer things to buy and fewer things to fail at 2am.
Consider a palm oil mill in the South South running fibre and shell in a 21 bar boiler. Steam demand for sterilisation and the oil room sits near 20 tonnes per hour during the season, and the mill previously carried a 1.1 MW electrical load on diesel. A back pressure set exhausting at 3 bar delivered about 1.35 MW gross across roughly 4,000 operating hours a year.
At a delivered diesel price near NGN 1,150 per litre and a specific consumption of 0.28 litres per kWh, that displaced load was costing about NGN 322 per kWh, or close to NGN 1.4 billion a year. The turbine island came in at roughly NGN 950 million. The exhaust steam still did the sterilisation. That is the back pressure case in one paragraph, and it only works because the steam demand was already there and already firm.
The condenser is the part that costs you
Buyers underestimate the condensing side of the plant because the turbine data sheet does not show it. The condenser, cooling tower, circulating water pumps, vacuum equipment and water treatment often carry more capital cost and more operating risk than the turbine itself.
Start with water. A condensing turbine rejects roughly two to two and a half times its electrical output as heat. Evaporative cooling then consumes makeup water at about 1.5 to 2.5 m3 per MWh, plus blowdown to control cycles of concentration. For a 3 MW machine running 7,000 hours, that is in the region of 40,000 m3 a year. Confirm your borehole yield and water quality before you sign, not after.
Ambient conditions matter more here than most proposals admit. Condenser vacuum depends on cooling water inlet temperature, which depends on wet bulb temperature. Coastal Lagos and Port Harcourt sit at high wet bulb for most of the year, and a cooling tower that was designed against a 28 degree wet bulb will not hold design vacuum on a humid afternoon. Output falls with it, typically by several percent, and the loss is real rather than a fault.
Air-cooled condensers remove the water problem and introduce two others. They cost considerably more, and they derate as dry bulb temperature rises, which in the northern states means the hottest hours of the day are also your weakest output hours. Harmattan fouling of the finned tube bundles adds a cleaning regime that nobody budgets for in year one.
Then there is vacuum integrity. Air in-leakage through gland seals, flanges and the low pressure joints degrades vacuum, cuts output and raises exhaust temperature. Low flow at poor vacuum causes windage heating in the last stage blades, which is exactly the condition exhaust hood sprays exist to protect against. Minimum flow limits on a condensing turbine are protection settings, not suggestions, and they constrain how far down you can turn the machine.
Condensing plant also brings wet steam into the last stages. Moisture erosion on long last stage blades is a known life-limiting mechanism, and it is one of the conditions that shows up first as a change in vibration signature. If your existing machine is already telling you something, our note on steam turbine high vibration covers what the readings usually mean before you open a casing.
Where a back pressure turbine will let you down
The weakness of a back pressure machine is that its output is hostage to your process. Power follows steam demand, and when a production line goes down, the turbine goes with it. On a single line plant with a monthly maintenance shutdown, that is a scheduled loss of your own generation at exactly the time you still need lighting, instrument air and effluent treatment running.
Seasonality does the same thing more slowly. A mill that runs hard for eight months and idles for four cannot rely on back pressure generation across the off season. Neither can a plant whose steam demand halves at night.
You also need a full capacity pressure reducing and desuperheating station in parallel with the turbine, sized to carry the whole process load on its own. When the turbine trips, the process header must not lose pressure. That station is not an optional line item and it should be interlocked so that a turbine trip transfers cleanly without a steam pressure excursion at the users.
Exhaust steam quality is a further detail worth checking at design stage. Back pressure exhaust is often still superheated, and process equipment sized for saturated steam will not transfer heat as expected. Desuperheating with treated condensate is the normal fix, and it belongs in the heat balance from the start.
Costing this properly means knowing what a lost hour is actually worth to you. If that number has never been calculated for your site, our breakdown of plant downtime cost per hour sets out the method. To have your steam balance, load profile and outage exposure reviewed together, book a plant assessment or call +234 803 000 0000.
Extraction-condensing when back pressure and condensing steam turbines both half-fit
Most real plants do not sit cleanly at either end. Steam demand swings, electrical demand does not, and the two are not correlated. That is the case an extraction-condensing machine was built for.
An extraction-condensing turbine takes controlled extraction at process pressure from an intermediate stage and passes the balance through to a condensing exhaust. Extraction control valves hold header pressure while the condensing tail absorbs whatever the process is not taking. Power output becomes largely independent of steam demand, within the extraction and minimum flow limits set by the machine design.
Consider a food processing plant in the North with a flat 2.5 MW electrical load. Steam demand ran near 12 tonnes per hour on day shift and dropped to about 3 tonnes per hour overnight. A back pressure set sized for the day load would have collapsed to under a third of its output at night, leaving the plant back on diesel for the very hours it was trying to eliminate.
A 3.5 MW extraction-condensing unit with controlled extraction at 6 bar held the electrical output steady across both conditions. The turbine island cost roughly NGN 4.6 billion against about NGN 2.8 billion for the back pressure option plus its bridging generation. The deciding figure was not capex, it was the 8,000 hours a year of stable self-generation that the back pressure arrangement could not deliver.
The counter-example is worth stating too. A starch plant inland specified a straight condensing set on the strength of its electrical output figure, then found borehole yield insufficient for the cooling tower duty. The retrofitted air-cooled condenser added roughly NGN 420 million and about 11 weeks to the programme, and the machine still gives up several percent of output on hot afternoons. The water study should have come before the turbine selection.
Where the plant already has a gas turbine and the question is whether to add a steam bottoming cycle at all, the same logic applies one level up and is worked through in simple cycle vs combined cycle.
What to measure before you commit
Begin with a metered steam balance rather than design figures. You want hourly flows at each header pressure across a full production cycle, including weekends and shutdowns, and at least one seasonal swing if your process has one. Design drawings routinely overstate steam demand by 20 percent or more, and sizing a turbine against an inflated number produces a machine that spends its life below minimum flow.
Log electrical demand on the same timebase. The relationship between steam demand and power demand, not their average values, is what selects between the three configurations. If the two curves track each other, back pressure works. If they do not, you are looking at extraction-condensing or at buying flexibility somewhere else.
Confirm the water position early. Borehole yield, water chemistry, blowdown disposal and the cost of treatment all belong in the evaluation before a condensing option is priced. So does plot space, since the cooling tower and its plume clearance are frequently the binding constraint on a congested site.
Fix the performance test basis in the contract. Steam turbine acceptance testing sits under ASME PTC 6 and condenser performance under ASME PTC 12.2, and referencing the code is more defensible than paraphrasing an efficiency figure into a purchase order. The catalogue is published at asme.org.
Check your permitting position at the same time. Captive generation above the statutory threshold requires a permit, and adding or replacing generating plant changes what you previously declared. The current requirements are published by the Nigerian Electricity Regulatory Commission, and lead time there belongs in your project schedule rather than at the end of it.
Finally, be realistic about installation and commissioning. Alignment of the turbine, gearbox and generator train, condenser tube bundle installation, vacuum tightness testing and governor tuning are specialist activities requiring calibrated tooling and a qualified turbine engineer on site. On a retrofit into a live plant, the tie-in work needs a defined outage window, usually three to six weeks depending on header modifications. This is the phase where a good machine gets a bad reputation, and it is the wrong place to accept the lowest bid.
Steam turbine selection, commissioning and subsequent life all sit under steam turbine overhaul and services, and where the question spans the whole generation and process heat arrangement, a power plant audit is the right starting point. To have your steam and electrical load data reviewed against both configurations, request a technical proposal.
Frequently Asked Questions
Can a back pressure turbine be converted to condensing later?
Rarely in a way that makes commercial sense. The blade path of a back pressure machine is designed for the exhaust pressure it was sold with, and expanding to vacuum needs a different last stage, a larger exhaust casing and an exhaust connection the machine does not have. In practice you are buying a new turbine and keeping the foundation. Where flexibility is expected, specify extraction-condensing at the outset instead.
How much process steam demand justifies a back pressure turbine?
The threshold is firmness rather than volume. As a working guide, a steady demand of at least 8 to 10 tonnes per hour for 4,000 or more hours a year across a useful pressure drop begins to support a machine in the 500 kW to 1 MW class. Below that the electrical output rarely repays the switchgear, controls and grid interface. The pressure drop between your boiler and your process header matters as much as the flow.
Does a condensing turbine need more operators than a back pressure one?
Yes, and the difference is significant. The condensing side adds cooling water chemistry, vacuum system monitoring, condensate quality control and cooling tower maintenance, all of which need attention every shift. A back pressure set on a well instrumented header can often be supervised by the existing boiler team. Budget the manning difference in the operating cost comparison, because it does not appear in the capex column.
What happens to a back pressure turbine when the plant shuts down for maintenance?
It stops, and your site loses that generation for the duration. This is why the parallel pressure reducing station and an alternative supply, whether grid, an existing generator or a rented set, need to be planned as part of the scheme rather than treated as a contingency. Schedule the turbine’s own inspection work inside the process shutdown so the two outages overlap instead of adding up.