ZeroAvia Power Loss Could Lead To Tighter UK Rules On Experimental Testing

Tightening of rules under which experimental aircraft are test-flown in the UK has been recommended following investigation of the April 2021 forced landing of a hydrogen fuel-cell propulsion testbed flown by U.S./UK startup ZeroAvia.

The findings could affect the flight testing of new concepts in the UK.

ZeroAvia’s modified Piper Malibu was severely damaged after a loss of power to the electric motors. During a planned switch from battery to fuel-cell power, the windmilling propeller generated a back voltage that triggered the inverter protection system and locked out power to the motors.

Inverter lock-out had occurred on three prior flights, and a previous loss of power that resulted in a dead-stick landing had not been investigated by ZeroAvia, the UK Air Accident Investigation Branch (AAIB) says in a final report issued July 7 that details several safety management failures by the startup.

ZeroAvia has now revamped its test plans, established an aviation-grade safety management system and agreed to actively manage commercial pressure so it does not compromise safety, the AAIB says. This is evident in the more measured pace of the startup’s testing of a more powerful fuel-cell powertrain in a modified 19-seat Dornier 228. The aircraft is expected to fly soon at Kemble, England.

ZeroAvia’s Piper Malibu testbed was modified with two Yasa electric motors, powered by a lithium-ion battery and a hydrogen fuel cell and fitted with three pressurized hydrogen tanks in the cabin. The plan for the accident flight from Cranfield Airport, England—test flight 86—was to take off on battery and fuel-cell power then switch off the battery and fly circuits on the fuel-cell alone.

Power had been lost on flight 52 on Nov. 6, 2020, when the windmilling propeller resulted in the inverters locking out as a result of a high back voltage. That aircraft completed a power-off landing on the runway. A fault clear button was installed after flight 52 to allow the system to be reset, but an inverter lockout occurred again on flight 83 in April 2021 during the descent to land.

There were two other relevant incidents: flight 80 on April 23, 2021, when the fuel-cell shut down in flight and the aircraft landed on battery power alone; and flight 81 on April 26 when the fuel-cell shut down on the ground and the flight was aborted. Both shutdowns were caused by flooding of the fuel cell with condensation caused by excessive cooling.

After the changes were made following flight 52, a series of ground tests were undertaken. This included fast taxi runs but did not involve the change in power source. No testing was ever carried out that might have alerted engineers to the magnitude of voltage the motors could generate when being driven by the propeller. This contributed to the accident, the AAIB says.

On flight 83, the inverters locked out during a steep descent to landing. The power lever was at 0% and propeller at 2,050 rpm with voltage peaking at 825 volts. Fault clear on the runway restored power. “The issue was discussed with the flight test team, but as the flight crew did not consider it significant, a safety investigation was not carried out,” the AAIB says.

But the engineers that developed the procedure for clearing inverter lockout “did not recognize that if the fault clear button was pressed with the power lever in any position other than idle, the system would see a load demand and try to respond,” the AAIB says. The fuel cell might not react rapidly enough, triggering an undervoltage condition at the inverters.

After the lockout on the accident flight, the pilot pushed the fault clear button with the power lever still at 41%. This reset the inverter but, as the lever was not at idle, the step demand for power after reset could not be met by the fuel cell. This led to a voltage drop and an additional lockout.

The Malibu was being flown under UK Civil Aviation Authority (CAA) E Conditions rules for experimental aircraft. These place sole responsibility for safety on the so-called “competent person.” In the case of ZeroAvia’s project, the competent person was contracted as a consultant and was also involved in two other electric aviation projects with his full-time employer.

Changes were made to the flight-test plan without the competent person’s knowledge, and his involvement was restricted in several areas “due to issues within the organizational relationships, the fast tempo of the project and other work commitments,” the AAIB says. ZeroAvia’s CEO and flight-test director took on day-to-day responsibility but lacked the necessary safety and flight test experience.

“Given his availability, and the pace of the program, it would have been prudent for the operator to also have nominated an individual responsible for risk and safety management who was independent of the experimenting and management teams,” says the AAIB. “This accident demonstrates the importance in a complex, fast paced, experimental project in putting in place an appropriate safety management system at the start of the program.”

The AAIB recommends the CAA clarify the scope of projects considered suitable to be carried out under E Conditions and require an independent review of aircraft to ensure the project can be safely managed by a competent person. The AAIB also recommends the CAA requires the individual nominated as the competent person to have the experience and capacity to oversee the experimental test program.

The Malibu testbed was modified and flown with £2.7 million ($3.2 million) in UK government funding support under the HyFlyer program. The crash prevented ZeroAvia achieving the project’s goal, which was to demonstrate a 200-nm cross-country flight on fuel-cell propulsion. The Dornier 228 is being tested under the follow-on HyFlyer II program, supported by a £12.3 million government grant.