On grid stability
Readers of this site will be more aware than most of the difficulties involved in eliminating greenhouse gas emissions from electricity grids. The policy preference for wind turbines and solar photovoltaic panels renders the objective still more taxing. Those sources’ intermittency is well understood, and is generally regarded as the primary obstacle to large-scale adoption and reliance. However, it is not the only problem that they present for grid operation.
Relatively soon after they were first developed, electricity grids consolidated on alternating current (AC) technology, despite the fact that grid operation would be rather simpler with direct current (DC) systems. The adoption of AC was not a result of eccentricity on the part of physicists and engineers in pursuit of technical thrills: modern grids could not exist without it, and the disadvantages have long been accommodated and managed as a matter of necessity. Whatever might be said of its disadvantages, an AC grid is what we have, and an AC grid there shall be; modern DC transmission lines are a supplement, not a replacement.
Along with a basic requirement for balancing energy supply and demand, AC grid operation requires precise and consistent control of frequency and voltage. Deviation from acceptable values is more than a nuisance: connected equipment can operate improperly, and can be damaged. A grid has a certain highly strung and binary quality, in that it either works correctly or it does not work. Hence the practice of disconnecting portions of a network in response to serious instability; doing so is preferable to a systemic collapse. This unavoidable fact has fundamental implications for the design of generating plant that can safely be connected to a grid.
Grids have developed on a foundation provided by large synchronous alternators driven by a smooth and controllable source of mechanical power (water, steam or combustion gas). These machines produce an alternating voltage in time with their mechanical rotation. They can readily be connected together so as to operate in parallel, with each machine turning in time with the others. Such an arrangement can be a few metres apart, or separated by hundreds of miles. In other words, they can readily form an AC grid.
There are additional qualities that render them ideal for grid formation. They can vary their active power output, i.e. they can vary the amount of energy that flows out and is consumed. This enables the matching of supply and demand, and allows for control of grid frequency. They can vary their reactive power settings, which facilitates control of grid voltage in response to an unavoidable consequence of feeding AC power to and through certain devices. Their nature as large rotating machines turning in time with the grid provides synchronous inertia; a desirable characteristic that slows variations in grid frequency.
All of these qualities enable stable grid operation, and grids, as currently configured, could not function without them.
But what of wind turbines and photovoltaic panels?
They have the substantial disadvantage of being unable to increase their active power output beyond what the weather conditions allow at any given moment. That much is obvious on a time scale ranging from hours to weeks; hence the attention given to intermittency and energy storage. However, it also applies to shorter timescales: maintaining grid stability requires continuous adjustment. With appropriate control arrangements, wind turbines and solar arrays can reduce their output as required, but they cannot necessarily increase it. Furthermore, wind speed and solar irradiance can be volatile, even at short timescales.
Wind turbines use rotating alternators. However, for various reasons, they are unsuited to synchronous operation. Accordingly, irrespective of the variety employed, they must be operated asynchronously. Solar arrays are a DC source, and are therefore not alternators, whether synchronous or otherwise. Inherently, therefore, wind turbines and solar arrays offer no synchronous inertia whatsoever.
That wind turbines and solar arrays can be connected to an AC grid is due principally to the application of modern power electronics, particularly inverters. These devices shape the electrical output of asynchronous generators so as to mimic that of a synchronous alternator. This enables them to connect to a grid. If the conditions permit, they can supply active power. If their specifications and controls permit, they can vary their reactive power settings. However, there is a caveat. Simple mimicry is tolerable at sufficiently small scales, but is a potential liability at larger ones. This is a consequence of the limitations of the inverters currently employed.
Reliance on inverters extends beyond wind turbines and solar arrays. Batteries are the currently favoured technology to compensate for wind and solar intermittency. As with solar arrays, they are a DC source. Similarly, the importing end of a DC transmission line is a DC source that must feed an AC grid. This is an important consideration for the British grid, given that it is becoming festooned with interconnectors. If a battery installation has sufficient charge, or if the exporting end of a DC transmission line has power to spare, active power can be supplied reliably. In that regard, such installations are rather better than wind turbines and solar arrays. Nevertheless, the same inverter limitations apply, with adverse implications for grid stability.
The greater the number of wind turbines, solar arrays, batteries and DC lines connected to a grid, the less it resembles a network of synchronous alternators. It becomes inherently less stable. Attempting to operate a grid using asynchronous sources exclusively, or nearly exclusively, would lead to a swift collapse, irrespective of overall weather conditions. This is an engineering problem that is well understood in the relatively small world of grid operation, but is far less appreciated outside. Unless it is solved, even daydreaming about vast wind, solar and battery installations will be a waste of time.
Some aspects of a possible solution are reasonably plausible. For example, there are specialised devices that can improve synchronous inertia and reactive power compensation. However, a complete solution is currently not assured. The behaviour of inverters is an evident vulnerability. Work is under way to improve them, namely to make them act more like synchronous alternators, but success is not guaranteed. It is a case of a newer technology struggling to catch up with an older and more effective predecessor; perhaps an appropriate metaphor for the pursuit of decarbonisation as a whole.
The simplest way to solve or avoid the problem is to limit the capacity of asynchronous sources, relative to that of synchronous alternators. That, however, places a firm limit on asynchronous capacity addition, thereby undermining the current policy preference. On the other hand, given the vast cost and potential disruption of adding the envisaged asynchronous capacity, a technical limit on the scale of deployment could be seen as a blessing. After all, doing nothing is sometimes best.