The world is changing in many ways, but the electrification of transport is happening at a pace that few would have predicted as recently as 3 years ago. Although this mainly affects our personal transport, other applications for electrification are being developed. While hybrid and full electric cars are now a common sight on our roads, in only a couple of decades from now it is unlikely that any transport sector will not have moved away from burning fossil fuels for propulsion.

Beyond our day-to-day transportation, delivery vans are the next most popular target for electrification but, beyond cars and vans, most forms of transport aren’t moving quickly toward any form of electrification. Instead they are relying on a combination of increased efficiency and the promise of renewable/sustainable fuels.

However, there is a lot of interest in the electrification of air transport, and some big names are involved in developing solutions for powering aircraft by electricity.

This image shows the 16 connections to each cell – eight wire bonds each to the positive and negative terminals – meaning almost 104,000 individually tested connections in Electroflight’s battery (Images courtesy of Electroflight)

In RET 137 (February/March 2022) we reported that the air speed record for electric aircraft had been smashed by a UK consortium of companies funded and supported by aero engine supplier Rolls-Royce, which has a long history of air speed record attempts, dating back to the 1930s and the famous Schneider Trophy. The consortium’s single-seater aircraft, called the Spirit of Innovation, raised the record by 213 kph (132 mph).

The air speed records for electrically powered aircraft for 3 and 15 km respectively were raised to 555.9 kph and 532.1 kph, representing increases of 61.9% and 122.3% over the previous records. These are early days for EV air speed records though, and we can expect rapid progress, but these percentages are impressive by any standards.

Briefly, the Spirit of Innovation’s powertrain consists of a battery,  three inverters and a stack of three axial-flux permanent magnet motors produced by YASA. The motors are YASA’s standard 750R machines, which were designed primarily for automotive use.

The large diameter and short length of these machines, combined with their hollow, internally splined shafts make them ideal for combining in a stack. Each motor can produce a peak torque of around 800 Nm and, at 750 V, that can be maintained at around 2500 rpm at full power.

In many ways though, the most interesting aspect of the powertrain is the battery provided by Electroflight, based at Staverton in England, which some readers might recall it having done double duty as an airfield and racetrack.

While the battery is based on some relatively ordinary lithium-ion cells, an extraordinary amount of engineering has gone into its design and manufacture. Not only did it provide the energy storage for the record attempt, it is an integral part of the aircraft’s structure, as it provides a mounting for the propulsion motors and inverters, and reacts the considerable torque from the three electric machines.

Three battery modules make up the battery. Each module powers one of the YASA axial flux motors

The battery case attaches directly to the bulkhead of the aircraft, and every part of the propulsion hardware is either encased within or mounted onto the structural casing.

According to Electroflight’s technical director Doug Campbell, the YASA motors not only meet the torque requirement but offer an opportunity for redundancy and safety. One or two motors were insufficient to meet the torque demand and with three motors, three inverters and effectively three batteries mounted in a single case, it is possible to shut down one-third of the powertrain without any danger to pilot or aircraft – in fact Campbell says level flight can be maintained with only one motor and inverter.

It was therefore sensible to split the battery along these lines – that is, to have three separate battery modules, with each providing the power to one motor/inverter packaged in a single composite case, rather than a single battery with three connections. That allows any one or two sets of motor-inverter-battery module to be shut down independently and for the aircraft to return safely to the ground.

At the beginning of the project, it was necessary to have some loose requirements based on a planned flight profile involving four passes. Based on that, some calculation and simulation work was done to set the requirements for power and energy, discharge rates and so on.

 

Cell selection and characterisation

The battery as a whole has 6480 cells and is split into three identical modules; each module therefore comprises 2160 cells. They are of cylindrical format, Campbell saying Electroflight had “assessed a couple of pouch cells, but they lacked energy density”.

Pouch cells can have an advantage in terms of packaging, but Campbell says that with his previous experience of motorsport battery engineering he is “comfortable with the limitations of cylindrical cells” and that “pouch cells can be less predictable when they fail”. He adds that the cylindrical shape has certain “structural advantages”.

The battery voltage at maximum state of charge is 756 V, and the pack as a whole is capable of supplying a continuous 480 kW and a short-term peak of 750 kW. Assuming a nominal cell voltage of 3.6 V and a maximum voltage of 4.2 V, some simple calculations tell us that the most likely configuration for the cell modules is for the cells to be connected in groups of 12 in parallel and 180 of these groups connected in series.

The cells are 18650-type, which is to say that they are 18 mm diameter and 65 mm long. Campbell explains that there was a comprehensive cell selection matrix that included not only technical factors but also practical aspects such as cell availability. He says there is “a risk of not being able to source the cells” so it was decided that only cells available to Electroflight direct from the manufacturer would be considered.

The cells are from Murata, and are its VTC650 model, with MCA chemistry that Campbell was already familiar with from previous motorsport projects. He says they “combine good power and competitive energy content”. The VTC6-type cells come in a range of capacities and continuous current capabilities, and the exact cell chosen has a capacity of 3 Ah.

The attraction of the Murata cells versus comparable ones from other manufacturers is that the maximum continuous current is available up to a cell operating temperature of 80 C; other suppliers have products with similar currents but far lower operating temperatures. In a safety-critical application such as flight, Campbell says Electroflight “could not push beyond the data sheet limits of the cells”, so the energy capacity and thermal limit made the Murata cells an obvious choice.

The advantage of higher cell operating temperatures is that cooling becomes much easier and more space-efficient. Low operating temperatures mean less of a temperature difference between hot coolant and ambient air. The larger this temperature difference can be, the smaller the coolers can be through which powertrain heat is rejected.

A further advantage of the Murata cells is their high discharge rate, of around 10C. If we multiply 3 Ah by 10C, we get 30 A, the maximum continuous current available from the cell.

Campbell says that, having chosen the cell which would best suit the project, a lot of work was done to characterise the cell in detail, including aspects such as specific heat and thermal conductivity; he is keen to highlight here the help Electroflight had from Warwick University.

Further characterisation work included DC resistance measurement, ‘drive-cycle’ testing using a single cell and the predicted flight profile, thermal runaway testing, off-gassing (which happens when cells start to fail and vent gases to prevent explosion), internal short-circuit testing and measurement of flame temperatures once the cell had failed.

The capacity of the cell was also measured, and was found to depend on the rate at which the cell is discharged. Also, the cell’s impedance (a function of resistance and capacitance) was found to depend on temperature.

Prop-end view of the powertrain showing the three motors and inverters mounted to the structural battery case. Note how short the motor stack is

Such detailed characterisation work allowed Electroflight to build a high-fidelity simulation model that would accurately predict the performance and thermal behaviour of the cells throughout the planned flight profile, singly and when assembled as a battery.

The cells are of the ‘protected’ type, meaning they have features that should prevent a range of failures by interrupting the flow of current if the thermal limit is breached. However, this was never triggered in any of the flight testing.

With Rolls-Royce backing the programme and playing an important part – it supplied one of its test pilots to fly the plane, for example – safety was critical. It had to be happy with every aspect of the engineering of the plane and powertrain, and Electroflight followed a ‘stripped back’ aero development programme, using aero industry design and project review structures and gates.

As mentioned, the performance and energy capacity of the cells depend on temperature, and the cells are ‘conditioned’. Campbell explains that “for maximum performance, the cells must be heated to 22-25 C, well above ambient, to put the cell into an efficiency sweet spot based on cell impedance”.

For maximum performance, a low safety factor without compromising on safety is needed. In terms of reducing safety factors on a battery, we need to know more about how the cells are behaving. That involves adding temperature sensors to a number of cells and inferring that the temperature measurements are indicative of the unmonitored cells.

This admittedly imperfect strategy means a bit of headroom in terms of temperature is needed. Monitoring only a fraction of the cells is a strategy used in automotive and in most motorsport applications, owing to cost and complexity implications.

To reduce that headroom, Electroflight therefore decided that every single cell had to be monitored for temperature. Consider the scale of that task: 6480 temperature sensors per battery pack and every sensor monitored in real time. Campbell notes though that “100% temperature monitoring allows the battery pack to run at maximum performance and safety”.

There is a battery management system from McLaren Applied and vehicle controller analysing data in the aircraft, and the data is simultaneously transmitted to engineers on the ground for monitoring purposes. That helps the pilot manage the battery, and allows them to be informed if a problem should arise and action needs to be taken.

Campbell explains that the optimum position for the sensor was 43 mm from the negative terminal, to give the required accuracy and measurement robustness. That is not a rule of thumb for all cells of this type but it is based on the design of this particular battery pack and the specific cooling system design, which removes heat from the negative end of the battery.

 

Cell joining

In any battery consisting of multiple cells, we have to consider how to connect the cells to give the right voltage and energy capacity. In a TV remote control for example, this is simple, to allow easy replacement.

In an automotive battery though, each connection is permanent and must be reliable. There are a number of options here, one of which is laser welding; Campbell has experience of that but Electroflight chose ultrasonic wire bonding instead. That process worked perfectly with the chosen cooling strategy.

Each cell module is produced with the cells in the same orientation. On each cell itself, there is a flat end and one with a central pip, like the one on a standard 1.5 V battery for appliances such as the TV remote.

The pip of the 18650 cell is the positive terminal and the rest of the case is negative. That means the non-negative end of the terminal has both positive and negative portions to connect to, which is what Electroflight did – it had all the non-negative ends in one direction and made all of the cell connections at that end.

Between each cell and the next, there are eight individual wire bonds, meaning more then 50,000 individual joints. The wire gauge and number of joints were optimised for current density plus a degree of redundancy.

The specialist equipment, supplied by Hass Mechatronics, was designed to perform an individual pull test on every wire bond, ensuring each joint was both functional and sufficiently strong. If a bond problem was found then the testing would be stopped automatically so that cell could be easily cut out and replaced before resuming.

Unsurprisingly, the wire bonding system is CNC-controlled, but it also incorporates a vision system so that the system is adaptive. A camera senses the position of the positive and negative parts of each cell and makes individual adjustments to the welding program to ensure that the bonds are positioned relative to the axis of each cell, rather than being globally positioned relative to a fixed point in space. Taken as a whole, Campbell said that this Hass wire-bonding system is “incapable of putting down a bad bond”.

 

Cell cooling

The strategy of joining all the cells at one end meant heat could be removed from the opposite, negative end of the cell, which is a nice, flat area, and that allowed Electroflight to create a sealed cooling system in which, according to Campbell, “thermal resistance is reduced as much as possible”.

Each cell was individually bonded with a controlled bond gap to a bespoke cooling plate. The structural epoxy material for bonding the cells to the plate was produced by adhesives specialist Lord, and was chosen for its particular combination of properties, one of the most important of which is thermal conductivity.

The material contains glass spheres that are sized to ensure a specific bond-line thickness. Since the aim of the cooling system is to ensure that each cell is cooled optimally, it is important that the bond thickness is consistent, as the rate of heat transfer through a bonded joint where the thickness is too great would be compromised.

Given that the battery pack is fully instrumented, with each cell being monitored and the hottest cell out of 6480 providing the thermal limit, it is critical to have control over the bond thickness, as one poor bond will effectively limit the performance of the whole aircraft.

Rear view of the powertrain on the test stand showing the cockpit dashboard

Campbell describes the cooling plate as a three-layer construction with two flat outer layers and an intermediate turbulator plate in order to increase heat transfer. After the plates are joined by brazing, the face of the plate in contact with the cells was reduced to 1 mm in thickness by machining.

The heat from the powertrain is rejected to the atmosphere by a specially designed cooler manufactured by motorsport specialist PWR, in Australia. It is a multi-core cooler, reflecting the fact that the various parts of the system operate at different temperatures and use different fluids. The battery core uses water-glycol coolant that is at a far lower temperature than the motor and inverter coolant. The motor and inverter coolant is a dielectric fluid, which is not electrically conductive.

The previous issue of RET (issue 140, July 2022) contains an article on EV fluids, and deals with dielectrics in some detail. It is enough to say here that such fluids are optimised for their heat capacity, low electrical conductivity and, ideally, low viscosity. Cooling of EV system components is a rapidly expanding business, so a lot of effort goes into finding fluids that remove heat from electrical components more effectively and with lower pumping losses.

The coolant pump for the Electroflight battery was produced by specialist supplier Sobek. It is a single unit with one drive motor powering three separate pumps, one per battery module.

Despite the lower coolant temperatures compared to an IC engine, the cooler for the Spirit of Innovation is smaller than that fitted to this same basic model of aircraft when powered by an engine. The lower overall cooling requirement also means that the cowl directing air to the cooler is smaller than for an IC-engined aircraft, reducing aerodynamic drag. Campbell explains that with low temperature differences, a refrigerated cooling concept had been considered but was rejected.

The Spirit of Innovation comprehensively smashed the existing electric airplane air speed records; it is shown here with one side of the powertrain cover removed.

The battery coolant is a water-glycol mix of an undisclosed concentration, and was optimised for this application. Campbell says a lot of effort was put into “trying to convince Rolls-Royce to flow water through the battery. It was a shock for them, but quite usual for automotive”.

The advantage of water-glycol over a special dielectric fluid is its low viscosity and high specific heat capacity. It allows for a system wide approach for the best compromise between heat transfer and pressure drop (and therefore pumping power requirements).

 

Cell screening

Every cell received by Electroflight is tested to measure voltage, energy capacity, internal resistance and state of health, all of which is logged with each cell’s serial numbered so that they have their own ‘passport’ containing this important information.

The precise location of each cell in the battery module is chosen according to it its internal resistance, with those cells that tend to run slightly hotter positioned on the cooling plate to take advantage of higher temperature differentials between cell and fluid – that is the cells with the highest internal resistance are positioned closer to the battery coolant inlet. According to Campbell, that allows Electroflight “to get the absolute best performance out of the battery”.

The result of all of this work – from screening each cell and its precise placement, careful design and management of the bond gap, clever design of cooling and monitoring the temperature of every cell – means that the maximum cell temperature spread over the entire pack of 6480 cells is never more than 4 C.

 

Battery safety

We have all seen the stories of lithium-ion batteries exploding or catching fire in mobile phones, to laptops, cars and aircraft. Many will remember that the US Federal Aviation Authority grounded every Boeing 787 in 2014 because of malfunctioning lithium-ion batteries in the plane’s auxiliary power unit: a cell had overheated and gone into a thermal runaway.

A cell in thermal runaway is very dangerous. It causes neighbouring cells to do the same (called thermal runaway propagation), causing the venting of fumes and, eventually, battery fires.

Owing to this propensity to release gas close to cell failure, cells are vented so that they release this pressure before the cell explodes. It also means there is a specific location on a cell where a flame comes from when it catches fire. In the case of the cells used by Electroflight, this is at their positive end.

That end of the cell is not cooled, so there is no benefit to having a casing wall in close proximity, other than for packaging reasons. In a battery with circulating coolant, gaps between cell ends and the casing effectively control coolant speeds, but that is not relevant in this battery. Campbell explains that the battery modules were designed with a 45 mm gap between the positive terminal and the inside of the case.

This is the multi-layer cooling plate, which is vital to the success and safety of the battery. It also serves as a mounting for the cell temperature-monitoring PCBs. 6480 cells are monitored in real time

In the event of a cell fire, what Campbell calls “the umbrella effect” can make matters far worse. The effect is where a flame hits a nearby wall, spreads, cools slightly and gas or flame that is still very hot therefore contributes to the rapid thermal failure of nearby cells.

Moreover, the battery case is a structural part of the Spirit of Innovation and therefore needs to be protected from anything that might damage it; a very hot flame definitely falls into this category. The aircraft is no glider – it needs to maintain a degree of power in order to fly in a controlled manner to a safe landing.

Therefore, with the remote possibility that a flame with a temperature of about 1200 C could be inside the case means thermal protection of the case needs to be considered carefully. As part of the testing stage, one cell was deliberately failed in order to measure flame temperatures using thermocouples.

Also, part of the safety justification was that any venting of the cells must not result in these gases escaping in an uncontrolled manner. The battery is a sealed compartment, so escaping gases would cause an increase in pressure, and that in itself constitutes a significant safety risk if it is not mitigated.

The solution is to incorporate burst discs into the battery casing, which must be vented to the atmosphere behind the cockpit and in a position where they can do no damage. That means venting them into the heat exchanger exit duct. The battery case not only needs to be protected for structural reasons, but also to ensure that any vented gas is expelled safely.

The case also controls electrical creepage and clearance distances, which are the distances by which high-voltage components need to be kept apart from earthed components. The distances are increased in highly polluted or wet atmospheres, so it is important that the case is structurally sound and remains that way.

In the end, protecting the battery casing in the event of a cell fire was quite simple. Campbell says “Mother Nature provided the solution” in the form of a very surprising material: a thin layer of cork.

There were a number of concerns about microporous ceramics that ruled them out from use as the casing protection, including their response to chemicals, pressure and debris. Other than cork and ceramics, intumescent fire-protection materials (those that swell up when heated, protecting the material underneath or sealing a gap) were investigated, along with silicones.

Surprisingly, cork is used in some very high-tech applications and, as Campbell puts it “not just for sealing wine bottles and making pinboards”. Cork comes from the bark of the cork oak tree, which is noted for its ability to recover from forest fires.

Most trees regenerate through seed or perhaps re-sprouting from the remaining base of the tree. Cork oak main trunks however often survive owing to the highly insulating properties of the cork, and simply need to regrow the canopy.

Cork insulation is used in rocket booster protection and re-entry protection for hardware returning from space. Importantly, it does not intumesce, so the design clearance is maintained in case of a single cell fire.

The 2-3 mm insulating layer for the battery was provided by specialist aerospace supplier Amorim in the form a cork-phenolic composite consisting of granules in a phenolic resin matrix. The composite is cast into sheets, cut to shape and bonded into the composite case of the battery.

The composite proved capable of preventing the epoxy matrix burning, and indeed was shown to prevent the battery casing reaching the glass transition temperature of the epoxy.

The battery case structure is made from Hexply 8552 carbon-epoxy composite, which was chosen for its damage tolerance. In the same way that Electroflight minimised safety factors for cells by thoroughly characterising them and thereby maximised their performance, the composite chosen for the case was similarly characterised in detail. That allowed design safety factors to be reduced based on accurate information, with the result being that the battery case is as light as possible.

 

Summary

In an earlier conversation, Campbell had noted that the project leaned heavily on motorsport suppliers and engineers to ensure that the right skills were used and to get the project completed in a very tight timescale. Campbell himself has experience in Formula One and Formula E, and has worked on a number of interesting electrification projects in motorsport and automotive concept cars. The result of applying that experience and using motorsport suppliers on this novel aerospace project with an adventurous timescale and ambitious target was very successful.

At present the speed record seems beyond the reach of any competitors, while there could still be more to come from the Spirit of Innovation should the record be in danger. Surprisingly, the world’s fastest EV is a land speed record car, but there seems little doubt that over a single measured mile in two directions, the aircraft would prove to be faster. Even so, the record over the 15 km course was more than double the previous air speed record for an electric plane.

Electroflight has taken some fairly ordinary cells, added some novel cooling and safety concepts and produced an extraordinary energy store. The level of engineering is worthy of any top-level motorsport project, and the attention to detail in optimising every aspect of the battery is above anything I am aware of in any form of motorsport.

The development of negative-end cell cooling using water, and the application of cork composite for fire protection, are concepts that Electroflight will develop further for future projects. They should also help to maintain the Spirit of Innovation’s position as the fastest electric plane for some time to come.