Fuel cells: Leak testing tasks in the production of FCEV components.
Alternative drives are the future of the industry. In addition to pure electromobility, fuel cell technology is also becoming important. Manufacturers of the corresponding components - from bipolar plates to hydrogen tanks - will have to meet new requirements in terms of leak testing and quality assurance.
Fuel cell stacks are the heart of fuel cell vehicles. These fuel cell stacks consist of two end plates with several bipolar plates sandwiched between them. These are each separated by membrane electrode assemblies (MEAs). The electrically conducting bipolar plates have the task of connecting the anode of one cell to the cathode of the other cell. Each bipolar plate contains two cavities for the process gases hydrogen and atmospheric oxygen, and usually an internal cooling loop. Starting from the process gas flow cavities, the process gases hydrogen and atmospheric oxygen are fed over a large area to the membrane of the membrane electrode unit via the so-called flow field. The corresponding high-temperature cooling circuit has the function of maintaining an optimum process temperature for the entire fuel cell system. Essentially, this results in four failure modes for a fuel cell:
- Hydrogen loss in general.
- Crossover leaks between anode and cathode or overboard leaks at seals - with an uncontrolled reaction between hydrogen and oxygen.
- Loss of coolant, which reduces the efficiency of the fuel cell stack and leads to damage.
- Hydrogen leakage into the cooling circuit has a corrosive effect, impairs the efficiency of liquid cooling because of the gas bubbles and can even damage the pump.
Fuel cell stacks and the tightness of the individual cells
The failure scenarios result in specific requirements for the leakage rate. Against hydrogen leakage - both to the outside and into the cooling channel - the overall system must be protected against leakage rates in the range of 10-3 Until 10-5 mbar∙l/s to be tested. Hydrogen is known to be highly flammable and ignitable in the wide concentration range between 4 and 73 percent hydrogen in air. Some fuel cell manufacturers refer to the DIN EN IEC 62282-2 standard, the latest version of which was published in April 2021. The standard deals with the safety of fuel cell modules, but it does not deal with fuel cell applications in road vehicles. EN IEC 62282-2 specifies a hydrogen limiting leakage rate of 5 cm³/min for an entire fuel cell stack and requires the user to ensure good ventilation of the fuel cell. However, since this cannot always be guaranteed when installed in a road vehicle - think of a vehicle parked in a single-car garage - automotive applications often have more stringent leakage requirements. But regardless of which leakage rate is applied to the entire stack, the following applies: Because a complete stack consists of several hundred individual cells whose leakage rates must be considered in total, these individual components must be tested against limit leakage rates that are two decades smaller again. If, for example, the fuel cell stack consists of 350 cells and the individual cells are to be tested for leak tightness using helium test gas, a helium limit leak rate of approx. 10-4 mbar∙l/s. For more stringent specifications in automotive applications, individual bipolar plates can also have limiting leakage rates up to the range of 10-6 mbar∙l/s may be necessary. However, in the context of development projects and in scientific research, it is already being discussed whether in the future even smaller limiting leakage rates down to 10-7 mbar∙l/s would be more reasonable.
The vacuum method for line production
To avoid short circuits, the cooling medium in the high-temperature cooling circuit of the bipolar plates must have a low conductivity. As a rule, deionized water with an antifreeze additive is therefore used as the cooling liquid. To prevent this liquid from leaking out of the cooling channel, a leakage test against leakage rates in the range of 10-3 Until 10-4 mbar∙l/s is reasonable. This is the usual order of magnitude for liquid tightness, because water itself seals leaks of this size. For this and other leak testing tasks in manufacturing, the test gas-based vacuum method is recommended. It combines high reliability with short cycle times and is therefore particularly suitable for testing tasks in the production line. The test part is placed in a vacuum chamber, first evacuated and then pressurized with helium. The leak rate of the test part is determined on the basis of the test gas that escapes from any leaks into the vacuum of the chamber. In addition to the high-temperature cooling circuit that flows through the bipolar plates, FCEV vehicles also have one or more low-temperature cooling circuits that keep electrical components such as the powertrain, converters, and power electronics within temperature ranges of less than 60°C. They are operated with a conventional water-glycol mixture and must also be tested against liquid tightness.
Testing the bipolar plate for hydrogen leaks
The vacuum method is also used to test the bipolar plates themselves for hydrogen leakage. In this process, the hydrogen cavity of the bipolar plate is sealed, evacuated and filled with helium. In an evacuated vacuum chamber, a leak detector can then be tested against limiting leak rates of 10-4 Until 10-5 mbar∙l/s test. If no helium can be detected in the vacuum of the chamber, no leaks exist - neither from the hydrogen cavity to the outside nor into the cooling channel. However, if the instrument detects a leak, further investigation of the cause is possible. This is done by taking advantage of the fact that the hydrogen cavity of the bipolar plate is still filled with helium and sealed after the test in the vacuum chamber. However, only the cooling channel itself is now connected to a vacuum pump. In this way, it can be verified whether helium penetrates the vacuum of the cooling channel. Otherwise, it is certain that the originally identified leak leads to the outside.
Tests of assembled fuel cell stacks
After the bipolar plates have been assembled into complete fuel cell stacks, end-of-line tests are required - although tests can also be useful after preceding intermediate steps. For all these tests on assembled fuel cell stacks, helium is also used as the test gas. If hydrogen were used instead, there would be a risk that the fuel cell would already be producing electricity unintentionally. Hydrogen is also prohibited as a test gas for safety reasons, because major leaks in the hydrogen circuit could quickly lead to ignitable hydrogen concentrations of more than 4 percent in air. Typical helium boundary leak rates for leak testing of assembled fuel cell stacks in practice are in the range of about 10-3 Until 10-5 mbar∙l/s. The tolerable leakage rate for the complete fuel cell stack also depends decisively on the specific installation situation in the vehicle. The leakage rate at which an ignitable hydrogen concentration of 4 percent can occur in air is not only a question of the leak tightness of the fuel cell stack, but also of the volume surrounding it in the vehicle and the air exchange in this environment. These factors must also be taken into account when determining a sensible leakage rate.
Tightness of the hydrogen recirculation
Further leak tests are required on components such as the media distribution plate of a fuel cell (which conducts hydrogen, air and coolant), its various valves, pumps and its hydrogen recirculation. Fuel cells supply hydrogen and atmospheric oxygen superstoichiometrically to the membrane electrode units of their bipolar plates. This means that residues of each of the two gases remain when they react to form water. For this reason, fuel cells require hydrogen recirculation. The process gases first pass through a water separator, and the hydrogen fraction is then recirculated and used again. It is also advisable to test the hydrogen-carrying components of the hydrogen recirculation system against leakage rates in the range of 10-4 Until 10-6 mbar∙l/s.
Permeation limits for hydrogen tanks
The hydrogen tanks installed in FCEVs are mostly so-called Type IV tanks, made of composite materials. Such tanks for passenger cars are usually designed to withstand operating pressures of up to 700 bar. The much larger hydrogen tanks for buses are expected to withstand operating pressures of 350 bar. The leak tightness requirements for hydrogen tanks arise from a set of international standards that define maximum permissible permeation rates. For a passenger car hydrogen tank with a capacity of 30 l and a pressure of 700 bar, for example, the permeation limits of ISO 15869 B.16 translate into a helium boundary leakage rate of 2.3 ∙ 10-2 mbar∙l/s. In practice, however, hydrogen tanks are often not merely tested according to the standards, but against leakage rates in the range 10-3 mbar∙l/s. This is because any measured leakage rate that exceeds the unavoidable permeation of the material itself is necessarily indicative of a real leak.
Accumulation test on the hydrogen tank
When the necessary fittings and valves are attached to a hydrogen tank, the original tank body becomes the so-called tank module. Both the vacuum method with helium and the accumulation method with forming gas are suitable for preliminary testing of the tank body. In the latter, the test part is exposed to a non-flammable mixture of 5 percent hydrogen and 95 percent nitrogen, the commercially available forming gas. The leakage rate is calculated from the quantity of test gas that then escapes from the test part in a simple test chamber and accumulates there over a defined period of time. Because production figures are not yet high enough to make vacuum testing worthwhile because of its shorter cycle times, this accumulation method is often still used. The large hydrogen tanks of buses in particular, which have volumes of up to 1,700 l, are tested in accumulation chambers with up to 4,000 l chamber volume. Because of the lower test gas costs, the test specimen is filled with the less expensive forming gas. However, at a pressure of 700 bar, because the otherwise much smaller leak rates would not be detectable in the very large accumulation chamber. Because of the high test pressure, there is also an emergency outlet in the accumulation chamber in this special case, which opens in the event of overpressure.
Sniffer leak detection on complete tanks with all fittings
Leak tests are still required even after the tank body has been assembled with all fittings - filling and outlet valves as well as pressure sensors. However, the so-called sniffer leak test is usually used here. The finished tank is filled with either helium or forming gas as the test gas and sealed. A sniffer tip is then moved along the surface of the tank. The focus is on the neuralgic points, i.e. the connection points to the fittings. Automated, dynamic sniffer leak detection, in which a robotic arm guides the sniffer tip, avoids any errors made by a human inspector and guarantees maximum throughput. However, this requires leak detectors that have a particularly high gas flow. Otherwise, the robotic arm would not be able to move the sniffer tip across the test part fast enough or with the required safety distance. Typical limit leak rates for these end-of-line tests on finished hydrogen tanks are in the range 5∙10-2 mbar∙l/s.
Tightness of electric and hydrogen components
Ultimately, it is electric motors that move a fuel cell vehicle. The lithium-ion batteries that feed the motors are also the same in principle as in electric vehicles - although the traction battery in the FCEV is far smaller and acts only as a buffer. Leakage testing tasks exist here as well. No electrolyte may leak from lithium-ion cells, for example, and no humidity may penetrate the cells. Otherwise, the electrolyte could react with the water to form hydrofluoric acid. The testing tasks for the batteries, control modules and electric motors of FCEVs are the same as for electric vehicles. But the specific components of fuel cell vehicles also require very reliable leak testing. Especially since the term hydrogen is quickly associated with the word danger in the public's mind. Consistent quality assurance is therefore indispensable. Test gas-based methods are the way to go.
Author:
Sandra Seitz is Market Manager Automotive Leak Detection Tools at Inficon. Further information can be found in the e-book "E-Mobility: Leak Testing for Alternative Drive Vehicles". It covers the wide range of testing tasks involved in the industrial production of components for Battery Electric Vehicles (BEV), Plug-in Hybrid Electric Vehicles (PHEV) and Fuel Cell Electric Vehicles (FCEV). The e-book is available for free download here: https://www.inficon.com/de/maerkte/automobilindustrie/dichtheitspruefung-emobilitaet-elektroauto-brennstoffzelle
