Composite pressure vessels and why it’s worth researching

Forfatter: Publisert: Oppdatert:

PhD student Eivind Hugaas gives us an exclusive glimpse into his world of lightweight hydrogen pressure vessels in this blog post

My name is Eivind and my work makes up the larger part of Research Area 2.5 in MoZEES, which focuses on hydrogen pressure vessels made out of carbon fiber reinforced polymer (CFRP). Currently I’m just finished with the compulsory subjects and ready to work with research full time while also producing some hopefully exciting content for MoZEES!

Complex Composites

To achieve sufficient strength and avoid hydrogen cracking issues encountered when using metals in combination with hydrogen, CFRP is a good alternative to steel for hydrogen pressure vessels. However, CFRP and composite materials in general are complex compared to steel. Additionally, composites have seen less research than steel, partly because steel has been around for much longer than composites, in a modern sense.

Figure 1: Filament winding in the lab at NTNU.

The limited knowledge and the complexity of the materials results in costly and inefficient testing procedures for composite pressure vessels when used to store such a highly sensitive medium as hydrogen. The testing is a huge cost driver and high safety factors are employed in design compared to steel structures to counter any uncertainty in the material behavior, prohibiting use of the material’s full potential on par with steel. As a result, there lies great promise in increasing the material knowledge through research so that analytical approaches can replace some of the testing and more of the materials potential can be employed in the design.

The testing is a huge cost driver and high safety factors are employed in design compared to steel structures to counter any uncertainty in the material behavior

Production of fiber reinforced polymers can roughly be divided into two different production methods; either fiber mats are stacked as mostly flat panels or strands are woven onto a structure. How the polymer/epoxy is introduced into the structure varies. The former method is the preferred method for producing pressure vessels and is called filament winding.

Figure 1 shows a pressure vessel being produced on NTNU’s filament winding machine and Figure 2 shows a flat specimen being produced. The material behavior of the two production methods differ, even though the same constituent materials are used. Most of the current research has been carried out on specimens made out of mats and less on filament wound ones. This is largely due to that flat specimens are more convenient to work with and simpler to produce compared to filament wound specimens.

The core idea of this PhD is to allow for use of simpler testing procedures and higher confidence in analytical estimations of fatigue in filament wound pressure vessels.

Figure 2: Production of a flat glass fiber reinforced polymer sheet at NTNU with mats stacked on top of each other using vacuum to infuse the epoxy into the layup.

One of the most promising alternative test methods for filament wound pressure vessels is the use of split disk testing on pressure vessel cut outs.

Figure 3: Split disk test rig with a pressure vessel cut out installed. The two cut outs with the black ring inside has not got the liner removed yet. The liner can be seen as the winding mandrel in Figure 1.
Figure 4: Optical fibers glued onto a flat specimen.

Figure 3 shows the split disk test fixture along with pressure vessel cut outs from the vessel shown in production in Figure 1. During the summer and autumn the plan is to establish better knowledge on how pressure vessel cut outs behave when using this test method, logging strain throughout static and cyclic loading using optical fibers glued to the samples. Figure 4 shows optical fiber glued onto a flat GFRP specimen and Figure 5 shows how the strain fields from the fiber can be visualized. Optical fiber allows for measurement of near continuous strain fields making for convenient comparison with numerical models. Employing correlation between numerical models and experiments using key parameters such as strain allows for better simulation of full scale designs.

Composite materials are fickle to work with and can often surprise even the most experienced

There lies a great potential increase in confidence of analytical estimations if the already existing research on flat specimens can be applied to pressure vessel geometries. Establishing a link between material properties of flat and filament wound specimens holds great value. Therefore, along with testing using the split disk method, testing on flat specimens with the same constituent materials is also an important part of this PhD to compare.

Due to glass fiber reinforced polymer (GFRP) being a much more convenient material to work with as it is nearly translucent, this is used instead of CFRP for the larger part of the work. The material mechanics are the same for the interest of this work.

Figure 5: Post processing of optical fiber strain readings during a tensile test can give a fancy visual representation of the strain distribution.

The day to day work is very much based on logging practical experience in the lab. Composite materials are fickle to work with and can often surprise even the most experienced; this makes for a dynamic and interesting workflow demanding flexibility in both mindset and planning.