Simple Fatigue Strength Assessment of Fiber-Reinforced Plastic Components

Created by Vahid Mortazavian, Dr. Marcus Stojek, Sascha Pazour |

The use of plastic components in technical applications is growing. At the same time, the number of available material types is increasing due to optimized properties, adapted to special requirements. The measurement of characteristic values for the fatigue strength (SN-curves) is expensive and time consuming, not least because of the relatively low permissible frequencies for plastics under vibrating loads due to their self-heating. An anisotropic fatigue assessment increases the necessary amount for measurement data considerably due to the directional dependency, making anisotropic SN-curves a rare resource. Therefore, estimating or simplifying methods are very helpful for fatigue assessment and are often the only option for a simulation driven design.

Research Partner Motivation

Ascend Performance Materials is a leading global producer of specialty materials and chemicals, with a particular focus on nylon. The company was founded in 2009 and is headquartered in Houston, Texas, USA.

The transition from traditional combustion engines to electric vehicles introduces a distinct challenge for car manufacturers: the emergence of higher-frequency noise generated by electric motors, gearboxes, and associated components. While conventional methods effectively dampen noise, vibrations, and harshness up to approximately 300 Hz, electric vehicles operate at frequencies ten times higher. To address this issue, a potential solution lies in the adoption of plastic materials over metal. Fiber-reinforced plastics, which are recognized for their strength and stiffness due to the high glass contents, offer promising alternatives. However, the mechanical performance of these materials is intricately tied to fiber orientation within the components, presenting a significant challenge, especially in durability assessment. This paper explores various methodologies for conducting fatigue strength assessments, illustrated through the evaluation of a newly developed material and test platform by Ascend Performance Materials. Glass filled Vydyne AVS material portfolio Anti-Vibration System Technology | Ascend Performance Materials (ascendmaterials.com) which is specifically designed to mitigate noise, vibration, and harshness, while ensuring the requisite stiffness and strength for electric vehicle structures serves as a compelling case study in this context.

Approach

Ascend has developed a group of fiber reinforced PA66 grades with particularly good damping properties. While improving the components performance, higher damping in general reduces the allowable testing frequency even more.

A special test component has been tested in static and cyclic loading conditions. (although investigated, NVH and acoustics are not covered here). Figure 1 shows the component used in the test setup for cyclic loading.

Two different approaches are presented below. The starting point for the assessment is a FE simulation using an isotropic material model and the more complex variant with consideration of the local fiber orientations by an anisotropic material model.

Isotropic Material Model

The only material data available here are the results from a quasi-static tensile test. The corresponding material model is created as an elastic-plastic model from the stress/strain curve. The method according to Oberbach [1] is used to estimate the cyclic strength for 107 load cycles under purely alternating load. Figure 2 shows an example of the cyclic strengths estimated and measured in this way for two materials from a previous research. The procedure is described in [2] (Section 6.3) and also here: Estimation of Fatigue Strength Limits for Plastics.

Figure 3 shows the Wöhler curves estimated in this way for purely alternating and pulsating loads in comparison with the corresponding measured values.

The mean stresses determined from the FEM simulation can then be used to determine the allowed stress amplitudes.

The cyclic load factor aBK is calculated as the ratio of the occurring to the allowed stress amplitude. Depending on the toughness of the material and the local mulitaxiality of the stress tensor, either a comparative degree of utilization is determined (similar to von-Mises), or the maximum value of the main stress components is evaluated.

The determination of the mean stresses and stress amplitudes at each node, the consideration of multi-axiality and any other strength-reducing factors are of course automated in the S-Life Plastics software and are not discussed in detail here.

Figure 4 shows the result of such a cyclic strength verification for the test specimen under consideration. Different load amplitudes were applied in the simulation. The verification was carried out as an example for three different load heights at 105 load cycles. Figure 4 shows 75%, 100% and 115% of the failure load determined in the test.

Even at 75% of the critical load, local areas in the critical component notch are above a load factor of 1. The extent to which this material failure can also be interpreted as a component failure in the sense of a fracture cannot be answered without further ado. The maximum degree of utilization occurring in the left-hand image (at 75% load amplitude) is already 143%. At 100% of the failure amplitude measured in the test, the result (center image) shows a continuous overload (above 100%) in the critical area, with a maximum value of 335%. Without question, the result would be described as conservative. The failure load is underestimated. Nevertheless, the critical position in the component is correctly recognized and the permissible load height is correctly predicted in approximate size. In view of the available input data, this is certainly a good result.

Anisotropic Material Model

Taking into account the local fiber orientations from a preceding filling simulation, an anisotropic material model can be used in the structural simulation with the help of Converse and MatScape, for example. The occurring stresses are then present at every point of the component in the fiber coordinate system. Stiffness (and strength) depend locally on fiber direction and degree of orientation.

Figure 5 shows an example of the stress/strain curves (in flow) for a representative volume element (RVE) with different degrees of orientation. The phenomenological material model (Hill) used in MatScape, determines the model parameters Rij in such a way, that the respective material behavior is approximated by scaling from a measured reference curve.

Thus, after calibration of the anisotropic elastic-plastic material model, the static strengths are available for every direction and every degree of orientation. This is directly comparable with the situation in the isotropic case above.

The main difference in the anisotropic case is the different strength for each stress component (static and cyclic). It is no longer possible to evaluate the von Mises Stress or Principal Stress, for example, because the directional information is no longer available in the stress invariants. Instead, the degrees of utilization of the individual stress components must be evaluated separately against the respective strengths. Only then can the component-by-component degrees of utilization be calculated into an overall value. This is done here experimentally using the Hill method (see Figure 6). It is assumed that the R-factors Rij determined from the static curves, are also applicable to the cyclic strengths of the respective stress components.

As in the isotropic case, the failure-critical location on the component is correctly identified (see Figure 7). The cyclic load factor aBK is exceeded for the first time when the failure load measured in the test is reached (see Figure 8 center image). The course of the degree of utilization along the critical notch differs from that of the isotropic verification. This shows the influence of the local fiber orientation. Overall, the anisotropic verification is less conservative and, depending on the interpretation, is closer to the test results.

As in the isotropic case, however, the correlation of the locally calculated degrees of utilization, or their exceeding the value 1, cannot be directly equated with the component failure.

Summary and Conclusion

Even using only the material data usually available from short-term tensile tests, it is possible to make reasonable estimates of the service life of plastic components under cyclic load in the present case. The isotropic approach based on the Oberbach approach is significantly more conservative. In contrast, the anisotropic method shows no signs of exceeding the design limit until nearly 100% of the load amplitude. This highlights the greater accuracy of the anisotropic approach, which also allows for additional weight savings in design. Both approaches are still the subject of investigations. They will be incorporated into the S-Life Plastics methodology in the medium term. 


Authors:
Vahid Mortazavian, Global CAE Manager, Ascend Performance Materials, Royal Oak, MI, USA
Dr. Marcus Stojek, Managing Director, PART Engineering GmbH, Bergisch Gladbach, Germany
Sascha Pazour, Simulation and Sales Engineer, PART Engineering GmbH, Bergisch Gladbach, Germany

Literature:
[1] Karl Oberbach, „Calculation of Plastic Components, Calculation Methods and Allowable Strength Limits“, in Tagungsband Konstruieren mit Kunststoffen, 11. Konstruktions-Symposium der DECHEMA, Frankfurt/Main, 1981, Bd. 91, S. 181–196.
An annotated reprint in English language in pdf format is available here free of charge.
[2] M. Stommel, M. Stojek, und W. Korte, FEM zur Berechnung von Kunststoff- und Elastomerbauteilen, 2. Aufl. München: Carl Hanser Verlag, 2018.
Available only in German language here.

 

Back