Power electronics turns electricity into a flexible universal energy
History of Technology - The Discovery of Nanoworlds Enables Renewable Energy Supply for All / Episode 15
Translation from the German original by Dr Wolfgang Hager
The importance of power electronics for the transformation of the energy system is vastly underestimated. It is about to fundamentally revolutionise the way we deal with electricity.
In the last episode, we looked at how, within the framework of electrical engineering, "electronics" first established itself as a field independent of mechanics as our understanding of processes at the atomic level increased, and power electronics then gradually evolved from signal electronics. Just like photovoltaics, modern power electronics — the second central technology for the energy transition — only became possible with quantum theory-based semiconductor technology. It developed much more slowly than signal electronics, on which the entire digitised information processing and communication is based. The importance of power electronics lies firstly in the fact that it allows the characteristics of electrical energy to be adapted almost at will to any given requirement. Secondly, it makes it possible to digitally monitor and control these parameters over long distances and without delay. As we saw in the last episode, the basic technologies for this were available from the 1990s.
My article series on the history of energy technology:
The discovery of nanoworlds enables a renewable
energy supply for allThe episodes so far:
Nuclear fission: early, seductive fruit of a scientific revolution
Where sensory experience fails: New methods allow the discovery
of nano-worldsSilicon-based virtual worlds: nanosciences revolutionise information technology
Climate science reveals: collective threat requires disruptive overhaul of the energy system
The history of fossil energy, the basis of industrialisation
The climate crisis is challenging the industrial civilisation: what options do we have?
Nanoscience has made electricity directly from sunlight unbeatably cheap
Photovoltaics: Increasing cost efficiency through dematerialisation
50 years of restructuring the electricity system — From central control to network cooperation
Emancipation from mechanics — the long road to modern power electronics
This episode now focuses on how these innovations have since had a practical impact and what possibilities they open up. Especially since the turn of the millennium, the development, miniaturisation and spread of power electronics have undergone a breathtaking acceleration. The generation, transport, storage and use of electricity are thus becoming much more flexible and efficient. Electricity is thus emerging as the flexible universal energy.
Staggering acceleration of development since the turn of the millennium
MOSFETs (metal oxide field effect transistors) were invented as the first power transistors in 1969 — twelve years after the very first signal transistors. However, they did not become commercially available in larger quantities until 1976. Their advantage to this day is the high switching frequency of several megahertz. In 1984, building on this, the IGBT (insulated-gate bipolar transistor) was developed, which is widely used today and allows significantly higher currents (up to 6000 A) and voltages (up to 6500 V) to be switched, although at lower frequencies (now up to 120 kHz). Other special components remained less important.
After a slow start involving small quantities, the speed of development has increased dramatically, especially since the turn of the millennium. After the basic technologies became available in the 1990s, wind power plants, decentralised photovoltaic systems and electric cars led to an explosion in demand for sophisticated power electronics in large quantities. Ever greater sums are being invested in R&D. Manufacturing processes are constantly being improved with more and more new factories. The performance of the components is rising, while the unit costs are falling. In power electronics, too, several functions are increasingly being integrated into a standardised component, with sizes decreasing. Modularisation and standardisation are on the rise. Special components are being replaced by programmable standard components — this proved to be an advantage, especially during supply disruptions caused by the pandemic. In some respects, the situation in power electronics is similar to that in computer electronics at the advent of microchips in the 1980s. One measure of the progress in development is the power density, i.e. the processable power per volume. Since the mid-nineties, it has increased by a factor of 1000 (see figure). Only microelectronics has ever achieved this kind of progress.
Through all of these efforts, it has become clear that the possibilities of the silicon technology currently in use have been largely exhausted. New semiconductor materials, especially silicon carbide (SiC) and also gallium nitride (GaN), are rapidly gaining market share. Silicon carbide — whose raw materials silicon and carbon are available in abundance — offers much higher switching frequencies, tolerates higher voltages, has lower losses, can withstand higher temperatures and dissipates heat better than silicon. This is reflected in significantly smaller sizes, higher power densities, longer lifetimes and completely new design possibilities. However, the materials cannot simply be substituted: new components and circuits have to be developed. It is also necessary to optimise the — much more difficult — production and processing of high-purity large silicon carbide crystals, which are almost as hard as diamonds. The construction of new factories for SiC components and the development of ever more optimised circuits has picked up speed recently, leading to the expectation of further rapid improvements in all application areas of power electronics. The possibilities, at all stages, from raw material to electronic application, are far from exhausted.
Aluminium nitride has even better properties than silicon carbide, which is fast becoming established. But there is still a long way to go before it is widely used: only recently the production of small wafers has been achieved.
Power electronics revolutionises power generation
Without semiconductor-based power electronics, renewable energies would be many times more expensive, inefficient and difficult to integrate into an overall system.
The triumph of wind energy was in large part made possible by innovative power electronics: today, wind turbines are operated at variable speeds, often even without a gearbox. The frequency and voltage of the current generated by the turbine vary and must be adapted to the conditions of the grid into which the generated energy is to be fed by means of a converter. In the most commonly used design, the converter converts the generated alternating current into direct current and then back into alternating current that matches the requirements of the grid.
Photovoltaic systems are operated in much the same way: they generate direct current, which is converted into alternating current of the grid frequency by an inverter. Newer inverters above a certain size have to fulfil a variety of additional functions. Depending on the changing load condition of the grid, they must be able to regulate, on demand, the reactive power (undesired phase shift between current and voltage in alternating current), limit the feed-in power or precisely regulate the frequency. They can thus replace large-scale power plants as grid stabilisers. Small, stable and resilient mini-grids (microgrids) become possible, which can also function independently of the general supply grid, if necessary.
The ability of power electronics to transform electricity almost at will makes energy sources usable that were previously almost inaccessible due to irregularity and rapid fluctuations. One example is the use of wave energy, where a series of companies compete to develop a viable solution. The Swedish start-up Ocean Harvesting is developing a system where buoys that rise and fall with the waves drive rotating generators on vertical threaded rods mounted on ball bearings. The electricity generated, whose frequency and power vary with the waves, can only be used efficiently with modern power electronics. Such relatively small individual systems can be produced in large series. In this way, the developers hope to create a system that can produce cheaper electricity than offshore wind farms, produces twice as much power per area of sea, requires far less material, is easier to maintain and is not visually intrusive.
Another example is the precisely controlled energy recovery during vehicle braking. In the meantime, the technology has been developed to such an extent that it is even used in electric bicycles. In order to recover almost all the braking energy in electric cars and trucks, high power must be processed over a short period of time. The weight of the electronics then becomes very important.
Power electronics revolutionises electricity use
Power electronics also brings great savings and completely new possibilities on the consumption side.
First of all, this applies to all types of power supply units that convert electricity from the mains into direct current of different voltages, especially for electronic applications. In a smartphone, which today has the computing power of a mainframe computer from the 1970s, different components need different voltages — the converters required for this now measure only one or two millimetres.
Power electronics, in the strict sense, only concerns the digitally and flexibly controlled conversion of electricity into electricity with other properties. But it opens up completely new possibilities for the subsequent conversion of electricity into all other forms of energy and often back again:
electricity ←→ mechanical energy
electricity ←→ heat
electricity ←→ electromagnetic radiation (from radio waves to visible light to gamma rays)
electricity ←→ chemical energy
For the actual conversion, widely differing technologies have been and are being developed. However, almost all of them have in common that controlling with power electronics on the electrical side ensures significantly higher efficiencies and flexibility and often makes the conversion practicable in the first place.
The frequency-controlled electric motors already described in the last episode allow continuous speed control without losses. This brings considerable efficiency gains compared to the previous simple switching on and off. From heating pumps to motors for electric vehicles to large industrial plants, a considerable proportion of the previously consumed energy can thus be saved. In this way, new heating pumps only need about one-tenth of the electricity of the older models. This is due, on the one hand, to the higher efficiency of the pump motor and, on the other hand, to the sensor-controlled variable power control. Electric vehicles would only have a fraction of today's range without power electronics.
Indeed, today, electromobility is the strongest driver for the development of power electronics. A large number of converters are now required for the charging station, batteries, drive/regeneration, high-voltage and low-voltage electrical systems and all the smaller electrical units in the vehicle. The market for power electronics in vehicles is currently worth over $2.2 billion and is growing at a rate of well over 20% a year (power electronics as a whole: approx. 43 billion, a good 8% p.a.). Innovation cycles are short. Investment in new technologies and new factories is high.
Apart from motors and generators, there are also a large number of special applications at the interface between electrical and mechanical energy that have only become possible thanks to power electronics: Linear motors, pressure sensors, electronic scales, electronic vibration dampers, vibration welding, ultrasound, etc.
The combination of electronic power control, highly sensitive sensors and digital information processing allows technical systems to react "in real time" to external conditions. These days, robots don't just work through pre-programmed movements, but also cooperate sensitively with humans, changing the relationship between man and machine once again. Autonomously driving cars will soon be able to move independently on our roads, opening up the prospect that it will soon be cheaper to use autonomous rentals that operate around the clock than one's own car, which sits idle most of the time. New types of ships can skim across the water in a stable, calm and highly efficient manner, even in waves, with extremely fast-reacting hydrofoils — they are soon to revolutionise public transport in Stockholm.
At the interface between electricity and heat, inductive heating is particularly noteworthy. Quickly controllable induction hobs, which no longer require a hob to be heated but generate heat directly in the ferromagnetic cooking pot base with the help of a high-frequency electromagnetic field, bring significant savings for households. The principle of inductive heating has been used in industrial processes for a long time and will gain enormous importance when gas is replaced by electricity, especially as precise and fast control is now possible with power electronics. The microwave oven goes even further, heating not the vessel but the food itself by causing the water molecules inside to vibrate with microwaves of suitable wavelength.
Here we are already talking about the conversion of electricity into electromagnetic radiation. The replacement of incandescent bulbs by LEDs has brought a huge efficiency gain — through the direct conversion of electricity into light, instead of the diversions via the filament's heat. In every LED bulb that can be screwed into traditional AC sockets, there is a tiny converter that supplies the actual LED with low-voltage direct current. Without power electronics, LEDs, lasers, microwave generators, modern X-ray machines etc., would be inconceivable: in all these cases, electricity is converted into electromagnetic radiation — supplied and controlled by power electronics. (We will discuss the conversion of electricity into radiation in a separate episode.) And the technology that has perhaps changed our everyday life the most in the last two decades would be even more impossible: mobile communications.
Finally, power electronics is always involved when electrical energy is converted into chemical energy and vice versa. In the storage of electricity in batteries, in the production of aluminium, in the electrolysis of hydrogen...
Power electronics revolutionises power grids
As we have already seen with the grid feed-in of electricity from wind energy and photovoltaics and, in the last episode, with transformers: Power electronics, with its wide-ranging possibilities for digitally controlling and changing the parameters of electric current, is in the process of fundamentally transforming the transmission and distribution of electricity and thus the global electricity grid. Transmission grids are the largest technical structures ever built by humans — and their conversion is correspondingly expensive.
Electric utility grids — which we will return to in the overall view towards the end of this series — connect a growing number of different power sources to billions of power consumers. As we have seen, power electronics play an increasing role in feeding electricity into the grids and in connecting consumers. The grid itself has the double and interrelated task of making electricity available over long distances; and ensuring that at any given moment, the same amount of electricity is generated as is consumed.
In the past, this was achieved (as we saw in episode 12) in AC grids with passive transformers that can step up and down the voltage for efficient transport, with schedules for large consumers, by generously sizing the lines for peak consumption and via the control of a few large power plants.
Power electronics, in combination with digital information networks, now enable a tremendous flexibilisation of the system. This throws conventional decision-making mechanisms and institutional arrangements into question: Power electronics makes it possible, in principle, to convert electricity at all nodes of the widely dispersed grid, to measure its properties and to control it directly or remotely over wide areas. This makes it possible to control generation, transport and consumption in time and space in such a way that the system remains stable in the millisecond range as well as in the long term and is operated as cost-effectively as possible. The spectrum of decisions to be made here ranges from short-term automated switching operations and voltage adjustments to long-term politically influenced investment decisions. Somewhere in the middle, partially automated power exchanges for different time horizons come into play.
For the practical operation of the grid, four aspects are paramount:
The stabilisaton of AC grids without large power plants. Today, even without the inert mass of large rotating generators, the grid can be kept stable even in the event of small disturbances in the millisecond range: responsive power electronics in inverters of wind and solar power plants and, if necessary, in special facilities can provide immediate compensation.
High voltage direct current transmission (HVDC). Under water, alternating current cables have high losses due to the resulting magnetic fields. That is why islands were connected with direct current cables very early on. The cost and space required for the rectifiers and inverters at the ends of the cables are considerably reduced with powerful power electronics. To compensate for different weather conditions when generating electricity with sun and wind, it is necessary to build additional transmission lines over long distances, which have lower losses with HVDC. Moreover, unlike high-voltage AC lines, they can be laid in the (wet) ground in a way that is easy on the natural surroundings.
Capacity increase in the distribution grid. The load limits in distribution grids are usually not given by the thermal load capacity of the lines, but by the fact that the voltage exceeds or falls below the permissible values for which the connected devices are designed when the load changes. This becomes particularly relevant when the direction of the current can be reversed due to strong feed-in. The same applies to reactive power. Automatic stabilisation of the values by power electronics enables a considerable increase in capacity without line expansion. In addition, the grids at the lower voltage levels are designed for peak loads that only occur for a short time — power-electronically controlled buffering with the help of batteries or load reduction allows the average utilisation of existing grids to be increased considerably.
Digitally mediated control of generation, transmission and consumption of electricity at all levels, taking account of grid conditions.
However, this would not be possible without extremely fast digital data processing, without measuring equipment with digitally networked sensors and without digital communication networks. Power electronics thus become the crucial link between the information in the control systems and the electrical energy flowing through the lines. This means that the information on the power grid can be processed independently of the energy flows, and the results can be used to control processes at the energy level without time loss.
Expressed in terms of system theory, the new ability to influence the flow at millions of nodes in fractions of a second increases the system's degrees of freedom many times over. This increases the complexity and importance of the system:
Due to its increased versatility (applicability), the importance of the electricity system for energy supply and for technologised societies as a whole is increasing significantly
The multitude of configuration options at each connection and node multiplies the number of decisions to be made
The variety of actors who could make relevant decisions with new choices and control options in the system increases: new electricity producers and electricity consumers from private households to large corporations, manufacturers of hardware and service providers of all kinds, grid operators at all levels, politicians from municipalities up to the EU
The diversity and significance of the potential consequences of the decisions for these actors grow with the rising importance of the electricity system
This development creates great efficiency gains, but also requires a rethinking of the decision-making systems: who decides what and when? The previous episode already dealt with the fact that the historically grown centralised control logic, which can be traced back to the old technologies, is no longer compelling with the new technologies. Considerably more decisions could now be made in a decentralised manner — and this would offer considerable advantages.
Back to the direct current grid
The new possibilities for transforming electricity make networks using direct current increasingly attractive. Around 1890, the "war of currents" was decided in favour of alternating current because it could be transformed to higher voltages with lower losses, which was not possible with direct current at the time.
Direct current, however, has crucial advantages that gain relevance with today's technology: With an evenly flowing current, there are no magnetic fields that can induce voltages in other conductors. This reduces the resistance in cables, especially underwater, and less copper is required. Photovoltaics generates direct current, batteries store and supply direct current, many appliances use direct current. In wind power plants and energy-saving motors/generators, alternating current of variable frequency must always first be converted to or derived from direct current. LEDs, to which all lighting is gradually being converted, need direct current. And, very importantly, the control of grids becomes much easier: no phase shift, no reactive power, the frequency is simply zero. The voltage provides direct information about the output. The digitalisation of the grids becomes easier.
That is why considerable efforts are underway to convert supply systems to direct current at a wide range of voltage levels. This can also be advantageous for residential buildings — especially if they have photovoltaic systems, battery storage and charging stations for electric vehicles. However, many devices are not yet available with a DC connection, so converters or parallel grids are necessary for the time being. For industrial operations, intensive effort is being put into DC systems. The first factories are being equipped with them, and the necessary adaptations are proving manageable. In the meantime, R&D is also underway to develop medium-voltage grids with direct current.
The transition to DC grids will be gradual and flexible. Existing AC lines can be repurposed. Ever smaller and more cost-effective converters simplify the transition. Switching direct current is no longer a problem with semiconductor switches — for a long time, the danger of arcing and long switching times were an obstacle. From a technical point of view, direct current networks are clearly advantageous. In order to advance standardisation and mass production of corresponding components at all voltage levels, intensive efforts are underway to develop corresponding standards.
Superconductivity — the next revolution in the power sector?
Superconductivity and semiconductors are physically different phenomena. But the detailed understanding of current conduction in semiconductors has significantly advanced superconductivity, which was discovered by Heike Kammerlingh Onnes in 1911. It was not coincidental that one of the inventors of the transistor, John Bardeen, together with L. N. Cooper and J. R. Schrieffer, succeeded ten years after his first big invention in fully explaining the superconductivity of metals (BCS theory, 1957). For this he received his second Nobel Prize.
A wide range of metals and compounds show no electrical resistance at extremely low temperatures. Until 1986, according to the BCS theory, it was believed that superconductivity was not possible above 30K (-243°C). Then, however, the so-called high-temperature superconductivity, which has not been satisfactorily explained to this day, was discovered in ceramics, initially below 35K. This was followed by other materials that no longer needed liquid helium, but only liquid nitrogen (77K = -196°C) for cooling.
Technically, superconductivity has been used primarily for the construction of extremely strong magnets: for particle accelerators in nuclear research, for medical magnetic resonance devices (MRI) and for experimental facilities aiming at nuclear fusion. There is also increasing interest in strong magnets for a variety of automation concepts in industry. For operation at the highest possible temperatures, the focus has been on complicated metallic alloys containing barium, yttrium or gadolinum - all expensive, extremely rare elements. However, the required superconducting layer is only micrometre thick: only one thirty-thousandth of the copper in a cable of the same power is required of the superconducting material. For high-volume use, other materials are being explored.
For the transmission of electricity, two projects are worth mentioning:
In Munich, the city's utility company is planning a 12km long power cable with a capacity of 500 MW using superconducting GdBaCuO. This offers considerable space advantages despite nitrogen cooling. A 14cm cable diameter retrofitted in an existing cable duct provides a multiple of the power, a drastic reduction of line losses, lower voltage and thus less expenditure on transformer stations, and no heat generation. Many other cities are interested in the project because not only can it realise significant capacity increases in a given limited space, but it can also save costs if mass-produced.
In 2014-2018, the EU's Best Paths project developed and tested, among other things, a DC cable with the superconductor magnesium diboride (MgB2) for long-distance HVDC lines with a capacity of 3.2 GW. However, the low-cost material MgB2 requires cooling with liquid helium or liquid hydrogen (-253°C). Nevertheless, it is becoming clear that HVDC cables with MgB2 will be significantly cheaper than conventional HVDC cables.
Major obstacles to efficient electricity transport over longer distances are lengthy permitting procedures and high costs. With superconductors, laying underground cables becomes much easier and cheaper. Exposure to magnetic fields and heat is minimal, and the cross-section is small. This makes for much simpler permitting and cost-effective deployment.
For the development of electricity grids in densely populated regions and over long distances, developments in superconductivity are thus emerging — even without sensational breakthroughs — which, in combination with increasingly efficient power electronics, massively increase grid capacity, range and flexibility. This will make it increasingly easy to achieve full supply with only renewable electricity. For the US, it was calculated that a full supply of renewables with an optimised nationwide transmission grid would cost 46% less than with grids limited to individual states. For Europe, a similar study (funded by a superconductor company) showed 32% lower electricity supply costs compared to the current approach. In order for new components to be widely accepted by the network operators, however, it has so far been an essential requirement that pilot plants prove their reliability over many years. This is not yet the case today, and the number of pilot projects is small.
Despite all the current activities in this field, I cannot help feeling that in the old industrialised countries, the topic of superconductivity is underfunded compared to other, much less promising technologies (hydrogen, CCS, fusion, nuclear energy...) and is making much slower progress than would be possible. Advances in superconductors have largely resulted from the huge research effort put into the mirage that is nuclear fusion. That could change: China has been increasingly active in superconductor technology for years: in basic research, as well as in initial applications in power grids (cables, transformers, energy storage in magnetic fields) and in transport technology (maglev trains). China has not hesitated to take advantage of high-voltage direct current (HVDC) transmission: after ABB and Siemens introduced the technology there in the 1990s, around 60% of the world's capacity was installed in China in 2018, mostly of national production. A similar level of commitment can be expected for superconductivity.
Korean researchers recently announced a breakthrough: superconductivity at room temperature and atmospheric pressure with a new class of materials made up of unproblematic elements to boot.
A global hype was followed by disappointment: Hundreds of scientists around the world tested the material but could not confirm the effect. On the one hand, this story shows how easily the public is prepared to believe that a single new technology will solve almost all energy problems. On the other hand, it shows what possibilities the new material sciences can open up in principle. Superconductivity without additional cooling is currently unlikely but not impossible, according to what we know today. Much smaller cable cross-sections with no resistance or heat generation would make the entire electricity system much more efficient — not only the grids but also almost all electrical devices and machines. But even if such a material were found, it would take many years to produce wires and cables from it and decades for the producers of the essential applications to make the switch... So there is no reason to wait for it. Even without breakthroughs, a much more efficient system based entirely on renewable energy is already possible with the technologies available today.
In the previous episode and in this one, we saw how fundamentally nanoscience-based power electronics has already changed the way we deal with electricity and our technological possibilities. Increasingly rapid advances in materials science, now helped by artificial intelligence, will continue to drive this development. Power electronic functions are increasingly being miniaturised and integrated into other components such as PV modules, motors or electricity storage systems. In the next episode of this series, we will look — after photovoltaics and power electronics — at the third nanoscientific innovation fundamental to the energy transition: battery technology.