By John McAndrew, Senior Mechanical Engineer, Aurecon
The period from 1970 through to 1990 saw the construction of many very large power stations in Australia, with turbo-generators sized from 350 to 660MW. These power stations require specialised pumping equipment. Whilst there are hundreds of different types of pumps in power stations, the critical large pumps, or primary power station pumps, are the Cooling Water pumps, Condensate Extraction pumps, and Boiler Feed pumps, and this article covers some general aspects of these pumps.
Power stations require high reliability. They run at a high capacity factor, which means that the main pumping plant is running about 6 to 7000 hours per year, and therefore needs to be very reliable. The failure of a critical pump may cause the trip of a Unit and lost generating capacity, which can have huge financial consequences.
Pumping plant installed in power stations is therefore always of a well-proven type and design, and is built to meet both the continuous duty and occasional off-design operating conditions that occur in power plants.
Critical pumps used in the power station steam cycle
In the Rankine steam cycle the steam passes through the turbine and produces electricity, and is then condensed in the main condenser. For a 660MW turbo-generator the heat rejected in the main condenser will be approximately 800MW. To remove this heat, and to condense the steam, requires a very high cooling water flow, which is provided by the Cooling Water pumps.
The condensate, after condensing in the condenser, collects in the condenser hot-well. From here it is extracted by the Condensate Extraction pumps, and pumped to the Deaerator which provides system storage and removes air from the feed water.
The condenser operates at a very high vacuum in order to extract the maximum possible efficiency from the Rankine cycle, and the surface pressure above the condensate in the hotwell is only about 6 to 8 kPa (absolute). For this reason, NPSH is a major consideration when specifying condensate extraction pumps.
From the deaerator the feed water is pumped into the boiler by the boiler feed pumps. These pumps are very high pressure multi-stage pumps. They usually have a booster pump to provide suction pressure for the main pump, and they are usually provided with variable speed drives to provide flow control.
Most boilers have a drum filled half by water and half by steam. There is a distinct water level which must be accurately controlled and maintained, and the boiler feed pumps carry out this function. In power plants with supercritical boilers there is no need for drum separation, and a different method of feed control is used. The trend is towards using supercritical boilers.
On the cooling side, where cooling towers are not used, environmental constraints are applied to ensure that the cooling water from the lake or sea is not heated above a pre-determined temperature, often 35 deg C. This helps to protect the marine eco-system. If the water discharge temperature approaches the allowable limit, either electricity production has to be curtailed, or an attemperating system installed.
An attemperating system consists of large pumps which draw water from the source, either a lake or the sea, and discharges it at the power station outlet to cool the heated water before it re-enters the lake. Attemperating pumps are usually large capacity low head axial flow type pumps.
Particular power station pumps:
cooling water pumps
The cooling water pumps supply large quantities of cooling water to the condenser to condense the steam into water. In a 660 MW Unit, the main cooling water pipes would be 2.0m to 2.6m diameter, and the cooling water flow would be up to 20 m3/s per Unit.
There are two general types of cooling water systems. The first is the once-through type, or open type, which draws water from a lake or the sea, and then returns it back to its source.
The second is the cooling tower type, or closed type, in which the water is recirculated through the condenser and pipework system, and then cooled in an evaporative cooling tower (refer Fig 1).
Pumps on cooling tower duties usually have a developed pumping head of about 20-23m, and those on a once-through system from a lake or the sea would have a developed head of about 10 to 15m, depending upon topography and the length and size of the conduits. The additional head for the cooling tower arrangement is required to pump the water to the hot well at the top of the cooling tower.
Once-through cooling systems are usually designed with a syphon through the condenser. Such syphons are often very high, and can be 8 to 9m of negative pressure at the high point.
Most CW systems would be provided with pumps in a 2 x 50% configuration. In single pump operation, due to the lower friction component, the pumps will run out on the curve. The system must be designed to ensure that sufficient NPSH is provided for this operating case to avoid cavitation, and also that sufficient submergence is provided to avoid air entrainment, swirling and flow disturbances in the sump.
Pumps used on circulating water duties are usually of the mixed flow or axial flow type, and are usually non-overloading with increasing flows beyond the best efficiency point (BEP).
The type of pump usually used are vertical spindle wet pit pumps, with a directly coupled motor on top as shown in Fig 2. Sometimes multi-pole motors are directly coupled, and in other cases a four-pole motor and reduction gearbox are used, although the gearboxes can add unnecessary complication. The pump would have a suction bell, mixed flow impeller and a diffuser for head development. The pump shaft would have a radial and thrust bearing at the top end, and a product-lubricated ‘cutless” rubber bearing at the lower end.
Depending upon the shaft length, there may be one or more line-shaft bearings, which are either grease or product lubricated. Bearings with oil systems are usually avoided where possible to avoid any possibility of pollution, particularly if the water discharges into a lake or the sea.
Vertical wet pit pumps are proven technology suitable for continuous operation, and usually run at fixed speed. They are very reliable and have the advantage that they can be pulled out without the need for isolation on the suction side. This is a distinct advantage.
Vertical wet pit pumps work best when there is only a small suction side water level variation, otherwise the column length would have to be increased to accommodate the lowest operating water level.
Most of the large CW pumps used at power stations have high efficiencies in the region of 88 to 90%, and for this reason should be preferred where possible.
Wet pit pumps can be susceptible to poor intake design, which may lead to rough running, higher power consumption and vibration. Sufficient space should be provided in the initial design layout of the power station to accommodate the intake if wet pit pumps are to be used, and the design requirements of the pump supplier should be followed.
Other pumps for cooling water duty
Both double-suction pumps and concrete volute pumps are also used for power station cooling water duty in Australia. Double suction pumps in horizontal or vertical configuration may be used up to approximately 5000 L/s, and a typical horizontal pump is shown in Fig 5. This type of pump has a long history of use on cooling water duty. For the larger pump sizes a smooth and straight flow profile into the pump is essential, and an adequate margin of NPSHA over NPSHR is essential, to suppress any potential cavitation.
A concrete volute pump is a bottom suction single stage volute pump with the volute cast directly in the concrete works. These are slow speed pumps that can attain high efficiencies and flow rates up to 30 m3/s. They are very reliable, and there are many instances where they are used in a single 100% installation with no standby pump, although this is not common practice in Australia.
Attemperating, as previously described, is the process of cooling the heated water before it leaves the power station boundary. The aim is to keep below the required maximum discharge temperatures. These limits apply to water discharging to lakes, estuaries and to the sea, where strict environmental requirements need to be met.
Attemperating pumps are a part of the cooling water system in a power station. They do not usually pump through the condensers, and would therefore be of lower head than normal cooling water pumps.
Some power stations provide attemperating pumps as part of the initial station design, by providing up to 50% or more additional cooling water capacity to operate during the summer months.
Attemperating pumps are usually only needed through the summer period, and for NSW this means from the period from 1st December to 31st March each year.
In more tropical areas, attemperating pumps may be in operation for a substantial portion of the year.
Various types of pumps could be used for attemperating, although the basic requirement is for high volume pumping rates. They are usually, but not always, low head pumps, and they are very likely to be axial flow (propeller) type pumps. Such pumps are very susceptible to poor inlet conditions, and care must be taken with the pump intake design.
One suitable type are submersible pumps, which can be provided in an axial flow design. This type of pump has several good features as follows:
• Inexpensive, and proven design,
• Short manufacture period,
• Ease of installation and handling,
• Available in flow rates up to 5 m3/s or more.
As indicated earlier a condensate pump draws water from the hot-well of a condenser. The water is usually warm, and at an absolute pressure of about 6 to10 kPa. The NPSH available to the condensate pump is generally in the range of 0.6m to 1.5m. Condensate pumps are usually multistage pumps, and fitted with a low NPSH first stage impeller, or a double-suction first stage impeller which has a lower NPSH requirement.
Many different types of pumps or pump combinations could be used on condensate duty, such as horizontal multi-stage pumps, vertical turbine pumps, and ring section pumps. All would have a low NPSH requirement on the first stage.
Condensate pumps are critical power station pumps, and would usually be installed in a 2 x 100% capacity arrangement.
Historically, condensate pumps were horizontal pumps with multiple stages, operating at low speeds. Due to the low pressure at the pump glands, air ingress was usually a problem, and water injection was used to seal the glands. Modern condensate pumps use mechanical seals, but still require water injection to avoid air ingress.
For flow regulation, fixed speed multistage pumps were regulated either by discharge control valves, or by using cavitation to control the flow. This latter type of control is generally referred to as ‘submergence control’ or self-regulation, and was once very common on condensate systems.
Submergence control for the condenser hot-well level is inherently automatic, since it requires no control equipment. The level in the condenser hot-well will be pumped down until the available NPSH is just equal to the NPSH required by the pump, and then the pump flow reduces.
This type of condensate flow control is not suitable for modern day condensate pumps.
Today large turbo-generators would most likely use the vertical ‘can’ type multi-stage pump (refer to Fig 7). The ‘can’ allows the pump to capture the NPSH available, by inserting the pump into the basement floor into which the steel ‘can’ is inserted.
The sizing of condensate pumps requires some consideration. Whilst the flow required by the turbo-generator is known, there are occasions when the feed heating plant is bypassed, and under these conditions additional condensate goes to the condenser. The condensate pumps have to be able to pump these additional flows, which results in them often being oversized for their normal maximum duty by 30% or so.
Boiler feed pumps
The boiler feed system is the highest pressure section of the steam cycle. The boiler feed pumps draw water from the Deaerator storage tank, and pump it to the boiler.
Modern boiler feed pumps are multistage pumps, and use mechanical seals, and usually have axial balancing devices such as a balance drum to help take the hydraulic thrust.
The larger turbo-generators very often use barrel type boiler feed pumps, with a bolted discharge head (refer to Fig 8). Such pumps would usually have a removable inner cartridge assembly, which can be pulled out of the casing and replaced with a spare, making repairs and overhauls more efficient. The cartridge contains the impellers, diffusers, and shaft assembly.
There are, however, many different arrangements of boiler feed pumps, some of which use volutes instead of diffusers, and some with double suction first stage impellers, and opposed impellers arrangements for inherent balancing.
Smaller turbo-generators and industrial plants also use ring section pumps, where the entire cartridge assembly is held together by a series of exterior bolts (refer to Fig 9).
Many installations use 3 x 50% boiler feed pumps, with two of the pumps being driven by steam turbines, and the third pump driven by electric motor. In such a case, the electric pump would act as the standby pump, and also be used for unit start-up from cold. On a 660MW turbo-generator, the electric motor for a 50% pump would be approximately 10MW in size. Steam turbines are variable speed, and drive the pump directly at its rated speed up to approximately 6000rpm.
Some power stations use a 100% steam turbine driven pump, with one or two 50% standby electrically driven pumps. There are many different combinations, and a detailed reliability and efficiency study is needed to determine the most cost effective arrangement.
For motor driven pumps, the motors would be 4 pole, and a step-up gearbox or variable speed hydro-coupling would be used to attain normal operating speed.
Speed control of motor driven boiler feed pumps using hydro-couplings is limited to the range 25 to 100%, and a control valve may be required if lower flows are required, for example during start-up, or periods of low load operation.
Boiler feed pumps require a booster pump to provide sufficient suction pressure, and the booster pump is usually driven by the same motor or turbine drive. Sometimes the booster pump may be appended to the non-drive end of the existing motor, or driven by an extension shaft from the non-drive end of the pump. In the latter case a reduction gearbox would be needed for the booster pump, to bring it back to 4 pole speed.
In some cases the booster pumps can have their own drive motor, but this is not usual in Australia where it is customary have the booster pump driven by the same turbine or motor that drives the main pump.
Each boiler feed pump would have its own separate leak-off line (minimum flow protection line), fitted with a throttling control valve. The leak-off line would open at all flow rates below approximately 20% of full flow.
Boiler feed pumps have close internal clearances, and require protection from grit and debris, by the provision of a filter usually situated before the booster pump.
The standby pump would need to be in readiness to start at any time, and would have its suction and discharge isolating valves open (and non return valve closed). The water temperature is about 190 deg C at this point in the cycle, and the pump must be pre-warmed and ready to operate.
About the author
John McAndrew has extensive experience in the design of cooling water systems, pumping stations, and associated plant for both Power Stations and Major Water Supply Projects including production of associated specifications. His experience includes feasibility studies, specification of plant and equipment, hydraulic design and waterhammer studies of all types of pumping systems including power station feed and condensate pumping systems. He is currently a senior mechanical engineer with Aurecon.