The efficiency of large fossil fuel plant generators is almost 99 per cent – but manufacturers are striving to squeeze every extra drop of efficiency out of their machines, writes Paul Breeze
Manufacturers are now making generators that can cope with frequent starts and stops
The generator that is mounted at the end of the gas or steam turbine shaft in a fossil fuel plant or the machine connected to a wind turbine shaft is the element in the generation process that converts mechanical energy into electrical energy.
The technology upon which the modern generator is based dates from the middle of the nineteenth century and, in the intervening years, generator design has been perfected to such an extent that the machine is by far the most efficient element of the whole power generation train.
“The efficiency of large generators is almost 99 per cent,” says Ràƒ¼diger Stein, chief engineer for the Generator Product Line at GE Power Services. With efficiency that high, it is easy to imagine that there is little more to be gained. But, as Stein points out, the losses in a 900 MW generator are still of the order of 9 MW. A ten per cent loss reduction, the equivalent of a 0.1 per cent efficiency gain, would provide nearly 1 MW of additional output.
Improving efficiency by 0.1 per cent is possible, even with such highly efficient machines. Whether it is cost-effective is another matter, depending on the balance between the additional cost and the increased financial return. “It all depends on the market,” explains Dr Gunar Klaus, head of engineering at Siemens generator plant Erfurt.
“Some markets value an increase in efficiency.” These are usually the markets of the developed world where operating costs are important and capex is available. In other markets the investment cost is the bottom line; these markets generally do not support these small gains in efficiency.
There is another area, however, where significant design improvements are important whatever the market, and these relate to machine reliability. While reliability is normally high, when power plants are being used for grid support duties which involve much more frequent starts and stops as well as regular output ramping, the traditionally high reliability can be compromised.
Non-steady-state operation of this type can lead to a variety of stresses that design modifications may alleviate. With more fossil power plants being asked to provide this type of service, all generator manufacturers are now making machines that can cope with this type of duty cycle. Brush is typical. “Our machines are designed for two starts a day,” says Nigel Daft, Senior Electrical Design Engineer at the company.
Large generators of the type used in coal and natural gas-fired power plants can be up to 2000 MVA in generating capacity. While efficiencies for these machines are close to 99 per cent, the size means that any losses are still significant. A 400 MVA generator that might be typical of a combined cycle power plant, with an efficiency of 98.9 per cent, will lose 4.4 MW. Virtually all that power will be dissipated as heat – there might be a little sound energy as well, resulting from vibrations. All the dissipated heat energy must be removed from the generator if it is not to overheat.
There are several sources of loss. One is the resistive loss in copper windings of the stator and rotor of the generator. Then there are magnetic losses from eddy currents within the steel components of the two major components of the machine, rotor and stator. In most generators the rotor is an electromagnet that must be excited using a small amount of electrical energy from the generator, another energy loss. Finally there is the cooling system that is introduced in order to carry away the heat. Depending upon the size, this will involve air cooling, hydrogen cooling, water cooling or a mixture of the three. Cooling can introduce more parasitic losses so optimization of the cooling regime is also important if losses are to be kept as low as possible.
When looking at optimization of a generator, there are complex interactions to consider. Take the resistive losses. The resistance of the copper windings depends essentially on the amount of copper. Adding more copper by optimizing the size of the slots that hold it or reducing the volume taken up by the high voltage insulation tape used to insulate the copper from the stator core will reduce the electrical resistance of the winding and so reduce resistive losses. This will lead to lower heat dissipation in the winding which will reduce its operating temperature, and this will again lower the resistance because resistance rises with temperature.
John Yagielski, consulting engineer for Generator Systems Integration at GE Gas Power Systems, lists several areas where improvements can further reduce losses. Improved insulation materials, for example, offer two advantages. First they can be thinner, allowing more room for the copper windings. Secondly, they offer improved thermal conductivity, which means that cooling efficiency is also improved. Higher strength rotor steels support field windings with increased copper cross section, reducing one of the largest components of loss. New core materials can also provide slightly improved magnetic performance with reduced magnetic losses. Meanwhile, modern analytical techniques can be used to improve the cooling flows.
Cooling system optimization
Cooling a generator is vital in order to keep it within its operating window and to ensure resistive losses are minimized. For smaller generators, air cooling is preferred. “Air cooling is much simpler,” says Daft. Air cooled machines usually employ a mixture of passive cooling systems and airflow fans attached to the rotor of the generator. This keeps the moving parts to a minimum, reducing life-cycle costs. When air cannot provide sufficient cooling then hydrogen cooling can be introduced. Hydrogen is up to ten times more efficient than air but requires an extensive hydrogen cooling system to be installed as well as casing design changes to ensure safety. This entails a cost in terms of energy loss and reliability.
For this reason all manufacturers have pushed the boundary of air cooled generators as far as they can. “Forty years ago a 50 MVA machine would have been built with hydrogen cooling,” says Stein. Today the main generator makers are producing air cooled machines of up to 400 MVA. These modern machines can be used to replace many older hydrogen and even water cooled generators, reducing both lifecycle costs and reliability.
For the moment, 400 MVA appears to be the practical limit for air cooling. This is already a massive machine in size and weight, and removing the heat using air is a challenge. The cooling of the stator windings in these machines is normally indirect, meaning that the cooling air never comes into direct contact with the copper, but instead cools the insulated surface and the core components in contact with the insulated winding bars. Direct air cooling is possible, but it would add to the complexity of the design and may not currently be cost-effective. So, for larger machines, hydrogen cooling will remain common up to around 700 MVA, after which direct water cooling of the stator is used to allow ratings up to 2000 MVA.
The refurbishment of older generators to extend their lifetimes represents a significant business for many of the main manufacturers. A major refurbishment will normally upgrade the performance at the same time.
By using newer materials not available when the machine was originally built, it may be possible to optimize the stator slots – for example, to add more copper. Replacing the windings is common during refurbishment. Cooling flows can be optimized too with advanced design techniques, while diagnostic maintenance can reveal vibrational issues that can be resolved by some structural redesign. “This is a pretty large business”, admits Dr Klaus.
However, there are limits. For example, the steel core of the stator cannot be changed. Or, as Stein puts it: “The iron is fixed for an old generator.” There will often be cases where it is better to install a completely new generator, in the same footprint as the old machine but providing both higher efficiency and a higher rating.
Generators are being constantly refined
Credit: GE/Voith Hydro
One area in which all power plant machinery designers are being challenged is adapting existing technology for grid support. This is true for generators as much as for gas turbines and coal plants with complex steam cycles.
Improving efficiency by even 0.1 per cent is possible with today’s highly efficient machines
Credit: Voith Hydro
Fossil fuel plants, particularly those based on gas turbines, can provide grid support to back up renewable generation from wind or solar power, but this involves many more startups and shutdowns and more changes in output that would have been typical in a similar plant ten or 20 years ago. These increase the levels of stress within machinery, with the potential for early failure, increased wear rates and higher maintenance costs.
“The biggest challenge is to cope with the thermal stress,” explains Daft. For example, it is necessary to introduce slip planes to accommodate the thermal expansion of the copper in the stator as the load changes and the temperature rises.
Another area of concern is the rotor end cap which holds the windings in place. Expansion of steel components represents a challenge too. The generator shaft in a 100 MVA generator might grow up to 4 cm as its temperature rises; there must be room for that. In addition, when there is a turbine attached to one end, its shaft will expand too.
Vibrational analysis is important to ascertain where the natural frequencies of generator components fall. When load is changing frequently and rapidly, a natural frequency is more likely to be excited, adding stress and wear to components. Analysis allows designers to take measures to ensure that the natural frequency of the generator support structure is a long way from that of the rotor.
Specific components may also be strengthened to reduce the potential for movement. Vibrations in stator end winding can also be an issue, particularly with two-pole machines where the size means that the natural frequency is close to twice the grid frequency.
An additional problem, Yagielski notes, results from the large power factor and voltage variations found when units are acting in a grid support role. Faults in the core materials can lead to both losses and stresses which are exaggerated by these variations, so it has become important to eliminate – as far as possible – any faults within the core.
Wind turbine generators
While the development of large generators takes place in very small steps, one market where major advances in generator design are taking place is in the wind turbine market. As Raimo Sakki, product manager for ABB wind generators, points out, “Wind is still a young segment.” Technologies have not become completely standardized yet and manufacturers are still experimenting with different drive train configurations.
Many options are possible, and each has an impact on generator choice. The drive train may include a gear or do without a gear. For direct drive systems with no gears, the slow rotational speed means that a synchronous generator needs many poles. This can be achieved with a permanent magnet generator or with a conventional multipole synchronous generator, similar to a slow speed hydropower generator. If the drive chain includes a gearbox that increases the rotational speed of the shaft, other options open up.
A large three-stage gearbox will provide a gear ratio of up to 100:1, allowing a high speed asynchronous generator to be used. Alternatively a medium speed gearbox and generator may be chosen.
Then there is the matter of speed variation to take into account. The wind is rarely constant and the drive train must be able to cope with gusts. This is most commonly handled today with doubly-fed induction generators, a relatively old-fashioned technology that was considered a dying breed during the 1990s but which has found a new lease on life as manufacturers have learned how to make better versions of these machines.
Today, Sakki says, these have become relatively standardized with 4-pole machines in the 1.5 MW-2 MW range and 6-pole machines at 2.5 MW and above. The alternative is to use a full power frequency converter with an induction or permanent magnet generator, a solution that is more expensive but potentially offers greater reliability and the ability to satisfy grid requirements.
The efficiency of wind turbine generators is somewhat lower than large generators. “Siemens direct drive wind turbines achieve an efficiency of around 90-95 per cent, dependent on specific working point,” says Arwyn Thomas, head of electrical design for wind generators in the Technology Department of Siemens Wind Power and Renewables Division.
He believes permanent magnet generators are the key to higher efficiencies in both direct drive and geared drive trains. “In offshore wind turbines, permanent magnets generators are already the dominant generator type,” he points out. The potential efficiency advantage is already demonstrated by medium speed offshore permanent magnet generators which can reach 98.3 per cent efficiency.
The variety of options makes wind a much faster-moving market in terms of technology that the large fossil fuel market. But manufacturers in both are finding they have to adapt to changing conditions and demands. While the gains that can be made in the large machine market are tiny, each improvement in efficiency might reap an increase in output that is equivalent to a medium sized wind turbine.
Paul Breeze is a journalist specializing in the power sector.