JPT logo
SPE logo

Qualification of Composite Pipe

This paper describes the qualification process for a thermoplastic composite pipe manufactured for use in high-performance riser, jumper-spool, and intervention-line applications. The pipe is manufactured from polymer, carbon-fiber, and glass-fiber materials with an automated laser-based welding process.

Product Architecture

The body of the pipe contains no metallic components, only polymer and fibers. These pipes consist of a smooth-bore inner-pipe precursor, extruded from Victrex PEEK polymer, with a thin wall thickness onto which layers of composite tape are welded by laser to form the pipe wall laminate (Fig. 1). This laminate can be any thickness and orientation of fibers, with the function of providing the strength and stiffness of the pipe. Once manufactured, the function of the inner-pipe precursor is to provide high-integrity sealing, resistance to hydrocarbon and other fluid and gas permeations and high temperatures, and a smooth bore for high fluid-flow rate.

The internal bore is typically in the range of 2 to 6 in., and the lengths can range from a few meters to kilometers for spoolable applications. At either end are steel end fittings that provide the interface to the external subsea system. The composite end fitting uses a steel collar that interfaces with a taper in the laminate to transmit bending and axial loads from the pipe body to the external system.

Manufacturing Processes

A thermoplastic composite pipe is typically manufactured from individual composite tapes on the order of 10 mm in width, 0.2 mm in thickness, and several hundred meters in length. The composite tapes consist of many thousands of glass or carbon fibers impregnated with the thermoplastic matrix. Adjacent tapes are placed with the same orientation by robot, to form plies. The plies may be oriented in different directions through the wall thickness of the pipe to tailor the mechanical properties of the structure.

A laser or other heat source is used to heat each incoming tape and the surface of the pipe to a temperature above the melting temperature of the thermoplastic polymer. The thermoplastic welding process enables high levels of control and design flexibility to vary the layup and wall thickness along the length (for instance, to build in bend stiffeners).

Long lengths of pipe require spooling during manufacture because the pipe requires multiple passes through the welding equipment to build up the wall thickness. After manufacture, spooling may also be necessary for transportation and installation.

The automated laser-welding process has some applications in the automotive and aerospace industries but has not previously been used for subsea-pipe applications. As with any composite product, the manufacturing process has an important influence on the structural quality of the finished laminate. Hence, the mechanical properties of the pipe are a function not only of the precursor fiber and polymer but also of the manufacturing processes during pipe construction.

A detailed discussion of current technology-qualification standards as well as product-qualification standards is provided in the complete paper.

Approach to Testing and Analysis

The manufacturer follows the pyramid principle shown in Fig. 2. It is considerably cheaper, safer, and quicker to test on small-scale coupons rather than on full pipe structures, particularly at extreme temperatures and when establishing chemical-resistance limits of the material in the presence of oilfield fluids or hydrogen sulfide gas.

Long-term material behavior such as fatigue testing to cycles on the order of 106 can be evaluated an order of magnitude more quickly on material coupons than on pipes. Such effects on the material do need to be verified at a pipe level higher up the pyramid. However, where possible, such testing can be minimized by use of the structural model to predict behavior of the full-scale pipe.

Small Scale. Historically, the majority of composite-material-testing methods and laboratory-test equipment has been designed for use with flat coupons. This creates a significant challenge for manufacturers of pipes, whose manufacturing equipment is specifically designed to make specimens that are tubular.

For fiber-dominated properties such as tensile or compressive strength along the axis of a ply, the values are dominated by the fibers themselves and the bond strength to the matrix. For these tests, the manufacturer has elected to use standard test methods and to adapt the laser-­welding manufacturing process to produce flat coupons.

However, for shear properties, or those properties transverse to the axis of a ply, the strengths are dominated by the thermoplastic matrix and the quality of the welding process. To obtain realistic data, the coupon-manufacturing process needs to accurately represent physics similar to that which occurs in the production of the pipe. Therefore, unidirectional tubular coupons are selected for these tests.

Having decided the basic merits of flat vs. tubular coupons, the choice of test method for each property still requires some careful thought. This choice of test method and tabbing material affected the value of compressive strength along the axis of the carbon fibers in early experiments on the manufacturer’s material. Compressive testing of composites is notoriously difficult because of the inherent high strength of the material and influence of the test equipment. All the test methods are legitimate, but depending on the details of the implementation, the range of mean strengths that were measured is between 400 and 1000 MPa. Thus, before the specifics of the small-scale testing are decided, it is useful to have some knowledge of how the actual laminate behaves in order to judge which coupon results are reasonable.

It is important to avoid introducing artificial scatter into the coupon-test data, either from the coupon manufacture from the base laminate or from the test methods. Most composite-design methodologies require the use of characteristic strengths rather than mean values for design purposes. A characteristic strength is a certain number of standard deviations, typically two, away from the mean value measured in the test. Hence, artificial scatter can affect the useful capacity of the material significantly; this is a key challenge in testing composite materials. Therefore, it is important that every coupon be inspected after testing to ensure that it failed from a legitimate mechanism, rather than from a laboratory artifact, because invalid failure mechanisms typically lead to high variation in results.

Local Pipe-Body Analysis

A reliable structural-analysis model of the pipe body is as important as the test data. According to the pyramid principle, the model is needed to connect the small-scale coupon data to the prediction of the full-scale product performance. A validated analysis model is needed to avoid “qualification by testing,” which implies full-scale testing of each load case of concern. A validated analysis model allows these load cases to be simulated at low cost and high speed with a computer.

The wall of a typical thermoplastic composite pipe consists of plies of composite material arranged at different orientations selected by the designer to give the required structural response for the application. The manufacturer’s analysis model generates a layer-by-layer model for each laminate tape layer, with its defined properties and orientation with respect to the pipe axis. The modeling approach is developed on the assumption that the composite tapes are transversely isotropic in their structural response, allowing the number of elastic constants to be reduced from 21 to five. This greatly simplifies the volume of material properties that must be measured and fed into the model, as well as the complexity of modeling and post-processing.

Global Analysis

For a new product, the load cases to consider are not always obvious because the pipe may behave quite differently from traditional products. A global structural model of the pipe within the proposed subsea system must be used to accurately define the load cases of greatest concern for subsequent assessment in the local pipe model.

Having established the properties of the unidirectional laminate plies at the small scale, it becomes important to test at the intermediate scale with multidirectional laminate, with ply angles that are representative of those in the commercial product. This intermediate-scale testing allows the local analysis model to be validated, and adjusted if necessary.

In general, the manufacturer uses the simpler maximum-failure-strain model for an initial assessment of new designs. This is reasonable, because the design allowable strains are typically one-half to one-third of the failure strains, and, in this region, the structural response is quite linear. The computationally intensive progressive-damage model is used for subsequent checks of specific load cases deemed to be of highest risk.

Full Scale

The top of the qualification pyramid requires testing on full-scale pipe. This is to provide the final evidence and empirical data to verify the modeling approach and the use of materials testing from smaller scales. Strain gauging and digital-image correlation techniques were used to assess the mechanical behavior against analysis-model predictions. It is also important to validate trends in long-term properties such as fatigue performance seen on small-scale coupons or intermediate-scale laminates with checks at full scale.

While this approach is applicable to loading such as pressure, bending, and tension, certain effects (such as rapid gas depressurization and impact damage) cannot be computationally simulated so easily and, therefore, are best assessed by testing.

This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper OTC 27179, “Qualification of Composite Pipe,” by Jonathan Wilkins, Magma Global, prepared for the 2016 Offshore Technology Conference, Houston, 2–5 May. The paper has not been peer reviewed. Copyright 2016 Offshore Technology Conference. Reproduced by permission.