Low Cycle Fatigue Behavior and Variable Amplitude Fatigue Life Calculations for an SRIM Polymer Matrix Composite 930405

The objective of this research was to determine the feasibility of applying strain based fatigue life calculation models, which are commonly used for metals, to smooth SRIM polymer matrix composite axial specimens subjected to variable amplitude loading. A thorough investigation of the monotonic and strain controlled constant amplitude low cycle fatigue behavior of this material was conducted, including the effects of mean strains/stresses on the fatigue life of smooth specimens. Using these results, mean stress life calculations were made on the constant amplitude tests, as well as on smooth specimens subjected to strain controlled variable amplitude loading, using the Morrow and SWT mean stress models. These results were compared to experimental data, and it was found that the correlation between experimental and calculated lives was very poor, for both the constant amplitude and variable amplitude tests. An improved strain based model for variable amplitude fatigue life predictions was proposed which had good correlation with the constant amplitude data. Life calculations were made on the variable amplitude tests using this model and compared to the same experimental data. It was found that this model improved the correlation between experimental and calculated lives, and in some cases the correlation was found to be excellent. Thus, this new model appears to present a more feasible method to the design engineer for making reliable fatigue life predictions for this type of polymer matrix composite subjected to variable amplitude loading.

The fatigue behavior of continuous fiber reinforced polymer matrix laminates has been investigated heavily in the past 15 years, due to the desire of the aerospace industry to use these materials, with their high strength to weight ratios, in cyclic loading situations. Consequently, numerous papers have been published documenting the fatigue and fracture behavior of a large number of polymer laminates with various types of reinforcement. More recently, a different class of polymer composite materials, often referred to as automotive/industrial polymer composites, has gained interest as a possible alternative to metals and the more expensive polymer laminates in many load carrying applications. This class of materials includes randomly oriented continuous fiber reinforced polymer composites, such as structural reaction injection molding (SRIM). Much like their counterpart high performance aerospace polymer composites, these materials possess the same advantages of high strength to weight and stiffness to weight ratios, although generally to a lesser extent.

As the use of this class of polymer composites grows, the need to understand the fatigue behavior likewise increases. As a result, greater attention has been placed on investigating the response of automotive/industrial polymer composites to cyclic loads [1 - 3]. However, much as in the case of polymer laminates, we are a long way from fully understanding, modeling, and predicting the fatigue behavior of this class of polymer composites. In response to this deficiency, the Society of Automotive Engineers Fatigue Design and Evaluation (SAEFDE) Committee initiated a task group to study the fatigue behavior of automotive/industrial polymer composites. The long range goal of this task group is to develop a method for modeling the fatigue behavior and predicting the fatigue life of components made from polymer composite materials under various loading conditions, much as has been accomplished for metals. As a first step, this committee chose to study a typical SRIM polymer composite. This material was selected because it has gained popularity in the automotive industry due to its ease of manufacture, low cost, and design flexibility, and it is expected to see increased use in the near future in applications where cyclic loading occurs. In addition, since the matrix is of a thermosetting type, it is not sensitive to cyclic heating and creep at room temperature, eliminating a problem that occurs with many polymer composites when they are subjected to room temperature cyclic loading.

Current research literature reveals that the most common method of modeling the fatigue behavior of many types of polymer composite materials is through the use of a continuum damage mechanics approach, in which the decrease in stiffness during cyclic loading is most often used as the fatigue damage parameter. Currently, this is impractical to the design engineer for life predictions due to the complex analysis involved, and because the measurement of stiffness at a point in a component in the field is a very difficult and unreasonable task. However, fatigue life prediction methods for metals have been developed which have proven to be fairly reliable in many circumstances. One such model is the strain based low cycle fatigue (LCF) model, into which various parameters and modifications have been incorporated to account for mean stress/strain effects, such as the Smith Watson Topper (SWT) parameter and the Morrow mean stress model. Numerous commercial computer programs are available that utilize these models. In most cases, the variable amplitude history is reduced to a series of cycles through the use of a cycle counting procedure such as rainflow counting, and a mean stress model is used in conjunction with a linear damage summation rule to predict when the component will fail. The advantage of this type of model is primarily the simple and practical way in which it lends itself to making fatigue life predictions on components in the field under complex loading conditions. Clearly, it would be desirable to have similar fatigue life prediction methods available for polymer composite materials. It is this desire that provides the motivation for this study.

The objective of this research is to investigate the feasibility of applying strain based fatigue life calculation models, which are commonly used for metals, to smooth SRIM polymer composite axial loaded specimens subjected to both constant and variable amplitude loading. In addition, it will be determined if a simple, easy to use model can be developed that will improve the fatigue life calculations. To accomplish this, the monotonic, cyclic, and strain controlled low cycle fatigue behavior of this material, including the effects of mean stresses and strains, must first be fully investigated. From the information gathered through this testing, fatigue life calculations will be made on smooth specimens subjected to variable amplitude loading using some common models that are often applied to metals, as well as any new models that may be developed in this study. The results from these analyses will then be compared to experimental data to determine their accuracy and the feasibility of their use to the design engineer.