R.E.M. had a hit single in 1993, in which they asked ‘What’s the Frequency Kenneth?’ Now in 2011 Adam Stock finds engineers designing wind turbines are asking a similar question.
The UK electricity grid runs at approximately 50Hz; that is the alternating current switches from positive to negative 50 times a second. When all is well on the grid, the 50Hz value is maintained plus or minus about 0.1Hz (you can monitor the frequency on the national grid web site 1). This small variation has very little effect for the vast majority of electrical devices, however if there was a big frequency drop (or indeed a large frequency increase) then many devices would be adversely affected.
What causes the frequency to change?
In simple terms the frequency of the grid is directly linked to the speed of the generators supplying the energy. Imagine a typical coal power plant, producing energy by burning the coal to produce steam that in turn spins its generator. The frequency of the electricity produced is directly proportional to the speed at which the generator rotates. For those who like maths, the formula is F=P×S/120 where F is the frequency, P is the number of poles on the generator and S is the speed in rpm. In order to spin the generator at the set speed, a certain amount of torque is required to force the generator around – this value depends on the electricity demand due to the laws of conservation of energy. The electricity demand provides a torque resisting the rotation of the generator. The generator must therefore supply an equal force in the opposite direction (that is the direction of rotation) to prevent the generator from slowing down. The frequency will remain the same so long as the electricity demand remains the same and the torque from the generator is equal to the torque from the electricity demand. If the demand increases then the frequency will start to drop unless more fuel is burnt to increase the torque of the generator.
The best way to think of this is via an analogy of a man on a bike. As he rides along a flat surface he pushes the pedals round with a certain amount of force to go at a certain speed. If the road remains flat he needs to put the same amount of energy in to keep going at the same speed. What if he then starts to climb a hill? He then needs to push harder on his pedals to keep the bike moving at the same speed.
So far only one generator has been considered, however, the grid is made up of many generators all linked together and spinning at the same speed. To extend the bicycle analogy further, now imagine a very large tandem bicycle with lots of people all pedalling at the same speed. Again, if they reach a hill they will all need to pedal a little bit harder to keep going at the same speed. This is equivalent to the electricity demand increasing (such as when everyone turns on their kettle during the adverts on Coronation Street). A case can also be imagined now, where one person stops pedalling (maybe they fell off the bike…) and so everyone else needs to pedal harder to maintain the speed. This is equivalent to a power plant on the grid failing.
It’s important to note that when a person falls off the bike or when a hill is encountered, the bike will gradually slow down and not instantly move slower. There is stored energy in the spinning wheels which is gradually spent as the bike slows. This is an example of inertia and exactly the same thing applies on the grid. When a power plant fails or the demand increases, the inertia of the spinning generators mean that they gradually slow and so the grid frequency gradually decreases, giving the power plants time to burn more fuel and generate more energy to bring the frequency back up.
What about wind turbines?
“I thought this article was about wind turbines?” I hear you cry, “How do they change any of this?” The answer is that most wind turbines being built today are what are known as asynchronous machines (as opposed to the synchronous gas or coal powered generators). This means that they are decoupled from the grid i.e. they can spin at a different speed to the grid frequency. This is great if you are trying to get as much energy out of the wind as you can, as you can make your turbine spin at the most efficient speed, which varies with wind speed. Unfortunately however, it also means that the wind turbine does not supply any inertia to the grid at all. For now this is not too much of an issue as wind energy makes up a small fraction of the total power supply, so there is a limited impact on total grid inertia. As the percentage of wind turbines increases however, this will become more and more of an issue.
What are the solutions?
There are a number of possible solutions to the problem of a lack of inertia from wind turbines. The simplest idea is simply to make sure that future wind turbines that are installed are synchronous machines rather than asynchronous. This is a rather backward way of looking at things however, as the whole reason asynchronous machines were brought in was because they are more efficient. In addition to this, many future installations of wind turbines in the UK will be far offshore (some up to 100km offshore). Over this distance, the preferred method of electrical transmission is likely to be High Voltage Direct Current (HVDC) to help reduce losses in transmission, which requires the turbine to be asynchronous.
The ability to recreate the inertia effect using the wind turbine controller, known as ‘synthetic inertia’, is a more promising idea. There are various different ways of doing this that are currently being developed. These can be broken down into three broad categories:
- Power electronics at the converter
- Using the controller to utilise the pitch mechanism
- Using the controller to change the torque
The first of these options has a major flaw in that it is only applicable to Doubly Fed Induction Generator (DFIG) machines. DFIG machines would not be an option if HVDC is used. If HVDC is not used (which may be the case for a few, though by no means all sites) then utilising the power electronic controller can be used to give some ‘synthetic inertia’.
By utilising the controller in the converter, low levels of ‘synthetic inertia’ can be provided. Using a model of the Irish grid network Lalor, Mullane and O’Malley investigated what would happen in the case of a large dip in input power. Whilst the results of this technique were not as good as a simulation with no wind turbines, nor quite as good as those produced with synchronous wind turbines, there was a considerable improvement in frequency response compared to a simulation without the control technique.
Using the Controller to Utilise the Pitch Mechanism
One way of looking at the inertia supplied by synchronous power plants is to think of it as ‘extra’ kinetic energy that is added to generate power. If this ‘extra’ power was generated by the wind turbines then it could replace the inertia that is supplied by synchronous power plants.
The question then becomes ‘how do you get extra power output from a wind turbine on demand?’ The main problem is that the wind turbine is generating the maximum amount of energy that it can at all times when the wind speed is below what is known as the ‘rated’ wind speed. Above this wind speed the turbine changes the angle of its blades to reduce the amount of power it takes from the wind, because if it took more than this amount for a prolonged time it would damage the turbine.
One answer is fairly straight-forward then. If the wind speed is above the ‘rated speed’ of the wind turbine and the grid frequency drops, reduce the angle the blades are pitched at and increase the power output for a short time. This can allow the turbine to provide an artificial inertia response when the wind speed is high enough.
An additional advantage is that this technique could also provide some of the so called ‘primary response’ required by the grid as well as the inertia required. In the bicycle analogy the inertia is the spinning wheels that mean that the bike does not instantly drop its speed, the primary response is part of the boost in energy provided in the analogy by the cyclists pedalling harder. In order to prevent the frequency from dropping too much, additional primary response can be provided to help cover for the lack of inertial response as it could be argued that the grid simply sees both of these as additional electricity supply. The two do have slightly different effects however, with inertia acting instantly and slowing the drop in frequency, whilst primary response is used to limit the drop in frequency and to help the frequency return to an acceptable value. It can be argued that the response of the variable pitch method is not be quick enough to provide an ‘inertia response’ and is instead a method for providing primary response.
Additionally, below the rated wind speed the only way to allow the wind turbine to produce more energy on demand by this technique is to run the turbine with a slight pitch angle already on the blades. That is to run it in a less efficient manner most of the time. Obviously, many energy companies would not be hugely in favour of deliberately producing less energy.
Using the controller to change the torque
A final method for generating ‘synthetic inertia’ is to use the wind turbine controller to increase the torque demand on the generator. Ordinarily torque demand is used to control the speed the wind turbine rotates at. As the wind speed increases the torque demand is increased, which restricts the speed of the wind turbine. If the torque demand is too low the turbine will speed up and rotate too fast and will not be as efficient. If the torque is too high then the turbine will slow down and spin too slowly.
If changing the torque away from its optimum value decreases the efficiency then how can we produce more energy by changing it? The answer is that when we increase the torque, the inertia of the spinning blades means that they do not instantaneously begin rotating at a slower speed, they gradually slow down. This means that for a short time the blades are spinning at approximately the same speed as normal but we are extracting more torque at that speed. The power of a wind turbine is equal to the torque times the speed, so for a short period of time we can generate additional power.
Once the grid need for the synthetic inertia has passed (the frequency has stopped dropping) the energy of the spinning blades that was used up can be replenished, that is the turbine can speed up again. This requires the turbine to reduce its electrical power output slightly, however this process happens over a longer period and so the standard grid frequency regulation techniques would be more than capable of coping with this reduction in power supply without causing a frequency drop.
University of Strathclyde’s contribution
Whilst all of the above methods have been investigated and research is ongoing, there are still some barriers to applying the above techniques. Work being undertaken at the University of Strathclyde under the Supergen wind project is looking at ways to successfully implement controllers that can achieve the goal of synthetic inertia. Hopefully in the near future, this work will let wind turbine engineers stop asking “What’s the Frequency Kenneth?” and listen to the rest of their R.E.M. CDs.