Technical University of Munich develops conformal cubic tanks using carbon fiber composites to increase hydrogen storage | the world of composites

Standard flat-platform tanks for BEVs and FCEVs use thermoplastic and thermoset composites with a skeleton construction that provides 25% more H2 storage. #hydrogen #trends
After a collaboration with BMW showed that a cubic tank could deliver higher volumetric efficiency than multiple small cylinders, the Technical University of Munich embarked on a project to develop a composite structure and a scalable manufacturing process for serial production. Image credit: TU Dresden (top) left), Technical University of Munich, Department of Carbon Composites (LCC)
Fuel cell electric vehicles (FCEVs) powered by zero-emission (H2) hydrogen provide additional means to achieve zero environmental targets. A fuel cell passenger car with an H2 engine can be filled in 5-7 minutes and has a range of 500 km, but is currently more expensive due to low production volumes. One way to reduce costs is to use a standard platform for BEV and FCEV models. This is currently not possible because the Type 4 cylindrical tanks used to store compressed H2 gas (CGH2) at 700 bar in FCEVs are not suitable for the underbody battery compartments that have been carefully designed for electric vehicles. However, pressure vessels in the form of pillows and cubes can fit into this flat packaging space.
Patent US5577630A for “Composite Conformal Pressure Vessel”, application filed by Thiokol Corp. in 1995 (left) and the rectangular pressure vessel patented by BMW in 2009 (right).
The Department of Carbon Composites (LCC) of the Technical University of Munich (TUM, Munich, Germany) is involved in two projects to develop this concept. The first is Polymers4Hydrogen (P4H), led by the Leoben Polymer Competence Center (PCCL, Leoben, Austria). The LCC work package is led by Fellow Elizabeth Glace.
The second project is the Hydrogen Demonstration and Development Environment (HyDDen), where LCC is led by Researcher Christian Jaeger. Both aim to create a large-scale demonstration of the manufacturing process for making a suitable CGH2 tank using carbon fiber composites.
There is limited volumetric efficiency when small diameter cylinders are installed in flat battery cells (left) and cubic type 2 pressure vessels made of steel liners and a carbon fiber/epoxy composite outer shell (right). Image Source: Figures 3 and 6 are from “Numerical Design Approach for Type II Pressure Box Vessel with Internal Tension Legs” by Ruf and Zaremba et al.
P4H has fabricated an experimental cube tank that uses a thermoplastic frame with composite tension straps/struts wrapped in carbon fiber reinforced epoxy. HyDDen will use a similar design, but will use automatic fiber layup (AFP) to manufacture all thermoplastic composite tanks.
From a patent application by Thiokol Corp. to “Composite Conformal Pressure Vessel” in 1995 to German Patent DE19749950C2 in 1997, compressed gas vessels “may have any geometric configuration”, but especially flat and irregular shapes, in a cavity connected to the shell support. elements are used so that they can withstand the force of expansion of the gas.
A 2006 Lawrence Livermore National Laboratory (LLNL) paper describes three approaches: a filament wound conformal pressure vessel, a microlattice pressure vessel containing an internal orthorhombic lattice structure (small cells of 2 cm or less), surrounded by a thin-walled H2 container, and a replicator container, consisting of an internal structure consisting of glued small parts (eg, hexagonal plastic rings) and a composition of thin outer shell skin. Duplicate containers are best suited for larger containers where traditional methods may be difficult to apply.
Patent DE102009057170A filed by Volkswagen in 2009 describes a vehicle-mounted pressure vessel that will provide high weight efficiency while improving space utilization. Rectangular tanks use tension connectors between two rectangular opposite walls, and the corners are rounded.
The above and other concepts are cited by Gleiss in the paper “Process Development for Cubic Pressure Vessels with Stretch Bars” by Gleiss et al. at ECCM20 (June 26-30, 2022, Lausanne, Switzerland). In this article, she cites a TUM study published by Michael Roof and Sven Zaremba, which found that a cubic pressure vessel with tension struts connecting rectangular sides is more efficient than several small cylinders that fit into the space of a flat battery, providing approximately 25% more. storage space.
According to Gleiss, the problem with installing a large number of small type 4 cylinders in a flat case is that “the volume between the cylinders is greatly reduced and the system also has a very large H2 gas permeation surface. Overall, the system provides less storage capacity than cubic jars.”
However, there are other problems with the tank’s cubic design. “Obviously, because of the compressed gas, you need to counteract the bending forces on the flat walls,” Gleiss said. “For this, you need a reinforced structure that connects internally to the walls of the tank. But that’s hard to do with composites.”
Glace and her team tried to incorporate reinforcing tension bars into the pressure vessel in a way that would be suitable for the filament winding process. “This is important for high-volume production,” she explains, “and also allows us to design the winding pattern of the container walls to optimize fiber orientation for each load in the zone.”
Four steps to make a trial cubic composite tank for the P4H project. Image credit: “Development of a production process for cubic pressure vessels with brace”, Technical University of Munich, Polymers4Hydrogen project, ECCM20, June 2022.
To achieve on-chain, the team has developed a new concept consisting of four main steps, as shown above. The tension struts, shown in black on the steps, are a prefabricated frame structure fabricated using methods taken from the MAI Skelett project. For this project, BMW developed a windshield frame “framework” using four fiber-reinforced pultrusion rods, which were then molded into a plastic frame.
The frame of an experimental cubic tank. Hexagonal skeletal sections 3D printed by TUM using unreinforced PLA filament (top), inserting CF/PA6 pultrusion rods as tension braces (middle) and then wrapping the filament around the braces (bottom). Image credit: Technical University of Munich LCC.
“The idea is that you can build the frame of a cubic tank as a modular structure,” Glace said. “These modules are then placed in a molding tool, the tension struts are placed in the frame modules, and then MAI Skelett’s method is used around the struts to integrate them with the frame parts.” mass production method, resulting in a structure that is then used as a mandrel or core to wrap the storage tank composite shell.
TUM designed the tank frame as a cubic “cushion” with solid sides, rounded corners and a hexagonal pattern on the top and bottom through which ties can be inserted and attached. The holes for these racks were also 3D printed. “For our initial experimental tank, we 3D printed hexagonal frame sections using polylactic acid [PLA, a bio-based thermoplastic] because it was easy and cheap,” Glace said.
The team purchased 68 pultruded carbon fiber reinforced polyamide 6 (PA6) rods from SGL Carbon (Meitingen, Germany) for use as ties. “To test the concept, we didn’t do any molding,” says Gleiss, “but simply inserted spacers into a 3D printed honeycomb core frame and glued them with epoxy glue. This then provides a mandrel for winding the tank.” She notes that although these rods are relatively easy to wind, there are some significant problems that will be described later.
“At the first stage, our goal was to demonstrate the manufacturability of the design and identify problems in the production concept,” explained Gleiss. “So the tension struts protrude from the outer surface of the skeletal structure, and we attach the carbon fibers to this core using wet filament winding. After that, in the third step, we bend the head of each tie rod. thermoplastic, so we just use heat to reshape the head so that it flattens and locks into the first layer of wrapping. We then proceed to wrap the structure again so that the flat thrust head is geometrically enclosed within the tank. laminate on the walls.
Spacer cap for winding. TUM uses plastic caps on the ends of the tension rods to prevent the fibers from tangling during filament winding. Image credit: Technical University of Munich LCC.
Glace reiterated that this first tank was a proof of concept. “The use of 3D printing and glue was only for initial testing and gave us an idea of ​​a few of the problems we encountered. For example, during winding, the filaments were caught by the ends of the tension rods, causing fiber breakage, fiber damage, and reducing the amount of fiber to counter this. we used a few plastic caps as manufacturing aids that were placed on the poles before the first winding step.Then, when the internal laminates were made, we removed these protective caps and reshaped the ends of the poles before final wrapping.”
The team experimented with various reconstruction scenarios. “Those who look around work the best,” says Grace. “Also, during the prototyping phase, we used a modified welding tool to apply heat and reshape the tie rod ends. In a mass production concept, you would have one larger tool that can shape and form all the ends of the struts into an interior finish laminate at the same time. . ”
Drawbar heads reshaped. TUM experimented with different concepts and modified the welds to align the ends of the composite ties for attaching to the tank wall laminate. Image credit: “Development of a production process for cubic pressure vessels with brace”, Technical University of Munich, Polymers4Hydrogen project, ECCM20, June 2022.
Thus, the laminate is cured after the first winding step, the posts are reshaped, the TUM completes the second winding of the filaments, and then the outer tank wall laminate is cured a second time. Please note that this is a type 5 tank design, which means it does not have a plastic liner as a gas barrier. See the discussion in the Next Steps section below.
“We cut the first demo into cross sections and mapped the connected area,” Glace said. “A close-up shows that we had some quality issues with the laminate, with the strut heads not laying flat on the interior laminate.”
Solving problems with gaps between the laminate of the inner and outer walls of the tank. The modified tie rod head creates a gap between the first and second turns of the experimental tank. Image credit: Technical University of Munich LCC.
This initial 450 x 290 x 80mm tank was completed last summer. “We’ve made a lot of progress since then, but we still have a gap between interior and exterior laminate,” Glace said. “So we tried to fill those gaps with a clean, high viscosity resin. This actually improves the connection between the studs and the laminate, which greatly increases the mechanical stress.”
The team continued to develop the tank design and process, including solutions for the desired winding pattern. “The sides of the test tank were not fully curled because it was difficult for this geometry to create a winding path,” Glace explained. “Our initial winding angle was 75°, but we knew that multiple circuits were needed to meet the load in this pressure vessel. We are still looking for a solution to this problem, but it is not easy with the software currently on the market. It may become a follow-up project.
“We have demonstrated the feasibility of this production concept,” says Gleiss, “but we need to work further to improve the connection between the laminate and reshape the tie rods. “External testing on a testing machine. You pull the spacers out of the laminate and test the mechanical loads that those joints can withstand.”
This part of the Polymers4Hydrogen project will be completed at the end of 2023, by which time Gleis hopes to complete the second demonstration tank. Interestingly, designs today use neat reinforced thermoplastics in the frame and thermoset composites in the tank walls. Will this hybrid approach be used in the final demonstration tank? “Yes,” Grace said. “Our partners in the Polymers4Hydrogen project are developing epoxy resins and other composite matrix materials with better hydrogen barrier properties.” She lists two partners working on this work, PCCL and the University of Tampere (Tampere, Finland).
Gleiss and her team also exchanged information and discussed ideas with Jaeger on the second HyDDen project from the LCC conformal composite tank.
“We will be producing a conformal composite pressure vessel for research drones,” Jaeger says. “This is a collaboration between the two departments of the Aerospace and Geodetic Department of TUM – LCC and the Department of Helicopter Technology (HT). The project will be completed by the end of 2024 and we are currently completing the pressure vessel. a design that is more of an aerospace and automotive approach. After this initial concept stage, the next step is to perform detailed structural modeling and predict the barrier performance of the wall structure.”
“The whole idea is to develop an exploratory drone with a hybrid fuel cell and battery propulsion system,” he continued. It will use the battery during high power loads (i.e. takeoff and landing) and then switch to the fuel cell during light load cruising. “The HT team already had a research drone and redesigned the powertrain to use both batteries and fuel cells,” Yeager said. “They also purchased a CGH2 tank to test this transmission.”
“My team was tasked with building a pressure tank prototype that would fit, but not because of the packaging issues that a cylindrical tank would create,” he explains. “A flatter tank doesn’t offer as much wind resistance. So you get better flight performance.” Tank dimensions approx. 830 x 350 x 173 mm.
Fully thermoplastic AFP compliant tank. For the HyDDen project, the LCC team at TUM initially explored a similar approach to that used by Glace (above), but then moved to an approach using a combination of several structural modules, which were then overused using AFP (below). Image credit: Technical University of Munich LCC.
“One idea is similar to Elisabeth [Gleiss's] approach,” Yager says, “to apply tension braces to the vessel wall to compensate for the high bending forces. However, instead of using a winding process to make the tank, we use AFP. Therefore, we thought about creating a separate section of the pressure vessel, in which the racks are already integrated. This approach allowed me to combine several of these integrated modules and then apply an end cap to seal everything before the final AFP winding.”
“We are trying to finalize such a concept,” he continued, “and also start testing the selection of materials, which is very important to ensure the necessary resistance to H2 gas penetration. For this, we mainly use thermoplastic materials and are working on various how the material will affect this permeation behavior and processing in the AFP machine. It is important to understand if the treatment will have an effect and if any post-processing is required. We also want to know if different stacks will affect hydrogen permeation through the pressure vessel.”
The tank will be entirely made of thermoplastic and the strips will be supplied by Teijin Carbon Europe GmbH (Wuppertal, Germany). “We will be using their PPS [polyphenylene sulfide], PEEK [polyether ketone] and LM PAEK [low melting polyaryl ketone] materials,” Yager said. “Comparisons are then made to see which one is best for penetration protection and producing parts with better performance.” He hopes to complete testing, structural and process modeling and first demonstrations within the next year.
The research work was carried out within the COMET module “Polymers4Hydrogen” (ID 21647053) within the COMET program of the Federal Ministry for Climate Change, the Environment, Energy, Mobility, Innovation and Technology and the Federal Ministry for Digital Technology and Economics. . The authors thank the participating partners Polymer Competence Center Leoben GmbH (PCCL, Austria), Montanuniversitaet Leoben (Faculty of Polymer Engineering and Science, Department of Chemistry of Polymer Materials, Department of Materials Science and Polymer Testing), University of Tampere (Faculty of Engineering Materials). ) Science), Peak Technology and Faurecia contributed to this research work. COMET-Modul is funded by the government of Austria and the government of the state of Styria.
Pre-reinforced sheets for load-bearing structures contain continuous fibers – not only from glass, but also from carbon and aramid.
There are many ways to make composite parts. Therefore, the choice of method for a particular part will depend on the material, the design of the part, and the end use or application. Here is a selection guide.
Shocker Composites and R&M International are developing a recycled carbon fiber supply chain that provides zero slaughter, lower cost than virgin fiber and will eventually offer lengths that approach continuous fiber in structural properties.

Post time: Mar-15-2023