You are here

Development of a composite pressure vessel

News International-French

23 Mar 2011

This paper sums up five years of research and development focused on producing composite pressure vessel prototypes for the transportation of substances in powder form. Twenty years ago, Spitzer Eurovrac, a European leader in bulk powder transportation, was one of the first to consider replacing stainless steel with aluminium for this kind of vessel. They successfully reduced the weight of the vessel and then, a few years later, the weight of the trailer chassis. Since 2002, the company has been preparing the transfer from metal to composite with the support of Cetim-Cermat for design, calculation and process development.

(Published on October - November 2007 – JEC Magazine #36)





Designing and producing composite pressure vessels is not new and many different solutions have been developed over the past twenty years. These solutions proved to be particularly efficient when steel or stainless steel was directly replaced with composite materials. For example, composite tanks are increasingly used for the storage of gaseous, solid and liquid materials. While gas storage tanks are concerned by pressure vessel regulations, solid and liquid storage tanks are more often used as static tanks without any pressure, except the hydrostatic pressure of the fluid. The novelty of the project discussed in this paper is due to the fact that it combines many different constraints in response to the regulations applicable to pressure vessels and to road and goods transportation. The objective was to reduce the weight of the tank by over 20% in comparison with aluminium.


The Spitzer group has a world-wide reputation as a specialist in the field of vehicles equipped to discharge powders under air pressure. The group offers two types of products: horizontal tanks, which discharge the product from the bottom; and tipper tanks, which discharge via a single outlet at the back of the tank. These two types of tanks may be mounted on a trailer or semi-trailer chassis, on a supporting frame, or be used as a road container. The trailer and semitrailer chassis are manufactured in Germany, France and Hungary.


Cetim-Cermat has participated actively in the project from the very beginning. Cetim-Cermat is a French laboratory with expertise in metallic, polymer and composite materials, as well as in measurement and testing. The company is also a close partner of the Cetim, the Technical Centre for Mechanical Industries.


Composite tipper tanks

Spitzer decided to shift from metal to composite in 2002. Only the tipper tanks (see figure 1) were concerned by this development, the trailer chassis was not modified.



The project initially aimed at optimising transportation costs (reducing travelling costs, making the loaded and return trips profitable) and allowing quick and easy cleaning of the tank’s inner walls so as to avoid contamination.


Composite materials fulfilled the project requirements since the internal gelcoat demonstrated a perfectly smooth, nonporous surface that can be washed easily. Moreover, its greater heat insulation coefficient reduces condensation risks and hence the formation of crusts. The choice of available resins also considerably extends the range of products that can be transported, while minimising corrosion risks. Depending on the fibres used, the weight of the tank can be cut by up to 40%. A quick calculation shows that, at equal strength, carbon fibre makes it possible to save about 800 kg compared to a total weight of 2,000 kg for an aluminium tank.


The main steps of the project were as follows:


  • researching the applicable standards (transportation, composite materials, etc.),
  • drawing up specifications,
  • analysing the various components of the project,
  • researching and defining the characteristics of the various materials to be used,
  • analysing and calculating the project using the finite element method (FEM),
  • researching and optimizing the project as a function of the initial calculation results,
  • verifying the optimized study using the finite element method,
  • drawing up the final detailed plans in compliance with the 97/23/EEC directive,
  • having the plans and calculations verified by a certified organisation,
  • researching and producing the necessary tools,
  • launching the prototype.



This paper only describes the calculations and processes used to mould the prototypes. As part of the finite element calculations, an IDEAS thin-shell laminate was used for the general approach and volume modelling was used to study local behaviour. Figure 2 shows the shell finite element model of the larger, SK 65 type tank (64 m3). The main parameters considered were the following:


  • internal proof load: 3 bar pressure, tank filled with 64 tonnes of water,
  • internal service pressure: 2 bar,
  • hydrostatic pressure of the tank filled with water: 64 tonnes of water,
  • stresses due to the full tank elevation at 45°,
  • consideration of the frame holders,
  • stresses due to deceleration (2g forward), full tank,
  • possible external pressure,
  • design taking into account processes, parts, gathering, safety coefficient,
  • weight reduction compared to an aluminium solution,
  • European directive 97/23/EEC.



Figure 3 shows the results obtained with the main load parameters (3 bar of internal pressure, tank filled with 64 tonnes of water) and figure 4 shows the behaviour of the tank under external pressure, calculating the buckling modes.



During this step, our main objective was to create a tank while keeping in mind the process feasibility in industrial conditions. Thus, the tank was designed taking into account cost requirements while trying to optimize the material and process solutions. For example, a hybrid solution combining glass and carbon fibres was chosen for the raw material despite the weight saving generated by glass.


The final results were in line with the specifications, the cost target, and the industrial process requirements:


  • Tsaï-Hill criterion: 0.3 (maximum value) calculated from the proof load (3 b-64 tonnes full),
  • Transverse shear approach included,
  • Location of critical areas, contact hoop/central cradle,
  • Definition of processes for each tank component,
  • Weight reduction over 700 kg,
  • Composite layers made of carbon and glass fibres, locally reinforced with a core material,
  • Calculations validated by Bureau Veritas.


Vacuum-moulding technologies were probably developed to reduce the quantities of noxious elements escaping to the working environment during resin processing. In most cases, the infusion experiments produce a significant increase in mechanical properties compared to hand lay-up. Production Time Analysis shows a decrease in the total working hours.


The next step was to select a partner for the manufacturing step. Rousseau, a composite manufacturer located in the Deux-Sèvres department (France) with a national reputation in the field of filament-wound fibre-reinforced polyester tanks, was selected to manufacture the prototypes.


Manufacturing processes

The cylindrical shell was manufactured using filament winding technology and the following material combination (from the inside to the outside): gelcoat / glass fibres / carbon fibres / foam & glass core / carbon fibres.


The moulded heads and manholes were produced by reaction injection moulding (RTM), using glass fibres, carbon fibres and a gelcoat. They were joined to the cylindrical shell by adhesive bonding (structural adhesive), followed by interior and exterior hand lay-up. The tank was then painted. The first prototypes were moulded with vinylester resins but epoxy resin (or even thermoplastic resins in the future) could also be used to meet VOC regulations. The prototypes were then tested and qualified. Only the most significant test – proof load testing – is described in this paper.


Proof load testing was the main test apart from dimensional and visual checks. It consisted in filling the tank, placed on the front and central hoops, with 64 tonnes of water and then increasing the internal pressure from 0 to 3 bar. After the test, the tank was expected to show no visible leaks or



breaks. To compare the calculated and actual values, the first prototype was instrumented with strain gauges during the test. Figure 5 compares the actual values (cyan) and the calculated values (yellow) in strategic areas of the tank. It shows that the theoretical values were close to the measured values.



Finally, several prototypes were moulded (see figures 6 and 7). The prototypes were subjected to driving tests for two years, and then the industrialization stage was started. The project meets the economical and technical requirements, although the manufacturing process is still being optimized.