Simplified Workflow for Plastic Testing and Processing Measured Data for Effective Analysis

Thermoplastics are being used in a variety of applications that demand exceptional structural performance in service conditions. Challenging service conditions including static as well as dynamic loads, may lead to the failure of thermoplastic parts. Mechanisms of thermoplastic failure include short-term stress overload, long term effects such as fatigue and creep, degradation due to temperature and chemical exposure, and stress cracking induced by environment. Part designers and analysts often use numerical analysis tools such as FEA to effectively predict part performance under short and long-term loading conditions. These results are then used to judge the possible failure modes in the design and perform possible corrective action. Accuracy of the results obtained from FEA depends heavily on the material properties and constitutive models utilized in the simulation. As such, the most important step in plastic analysis is to obtain the appropriate material properties and transfer them accurately to constitutive models used in FEA.

Characterization of thermoplastics may turn out to be an inefficient and expensive step for designers/analysts because of the several factors. First of all, thermoplastics may have significant variation in their material properties with respect to the deformation/loading modes, i.e. tensile, compression, flexural and so on. Also, thermoplastics undergo creep/relaxation even at low temperatures. This behavior must be characterized when service loads are long-term. In addition, for fiber filled thermoplastics, the characterization process is further complicated as the material properties are function of the angle between loading and fiber directions. Considering these factors, we discuss in this article the work flow to determine the necessary tests, and how to organize and perform those tests so that the characterization step is efficient. Also, in the work flow we present the process of accurately converting test data into constitutive models and validating those before running FE simulations.

The first step in the workflow is to study the part requirements and asses the loading and environmental conditions on the part for its life cycle. As far as the various mechanical loads on a particular part are concerned, it is always helpful to draw a free body diagram and understand how the various loads interact with the part and what type of deformation mode will they experience. Preloads acting on the part should also be included in this exercise for accurate estimation of the total loads. Extracting the deformation this way greatly reduces the number of material characterization tests. For example, to run a FE simulation on a part subjected to a dominant compression loading would require only compressive material properties. Testing for other modes can be skipped for this example. In certain cases, parts may be subjected to the loading conditions that induce multiple deformation modes. To get accurate predictions, it is important to test the material in all such modes and use appropriate properties in simulation.

Once the deformation modes are determined, the time-span of loading should be assessed. If the loads are short-term in nature, quasi-static tests are sufficient to characterize the material behavior. Long-term loadings require a detailed study of the viscoelastic properties of the thermoplastic material. Viscoelastic properties of thermoplastics can be measured using two techniques, namely Dynamic Mechanical Analysis (DMA) and standard creep test as per ASTM D 2990. In DMA tests, materials are subjected to oscillatory strains and corresponding stresses are measured over a range of frequencies. This is used to calculate storage and loss moduli curves of the material with respect to the frequency domain and eventually time domain by conversion. Tests are conducted using the same specimen over the range of temperatures. Using the well-known property of the viscoelastic materials called as Time-Temperature-Superposition (TTS), test data for moduli is then converted into a master curve, which gives the long-term relaxation data for the given thermoplastic material. DMA tests can be completed within hours and are very useful in judging the long-term response of the material. The second approach follows the ASTM D 2990 where the thermoplastic material coupon is subjected to a constant stress or strain in a desired deformation mode for a time period, which resembles the part lifetime. Typically, the tests are run for 1000 hours and this data is used to extrapolate the behavior for longer times if desired. Typically, these tests are expensive and time consuming. A decision between the two approaches can be taken by evaluating the short-term strains in the material. For smaller strains, the DMA approach works well and can be followed for predicting creep. For larger strains, non-linear viscoelastic effects are significant where the second approach is more accurate. However, early in the design cycle, where there is a desire for conservatism, the DMA approach can be followed to get quick answers. If desired, later in the design cycle, 1000 hour tests can be performed. If the loads are long-term and oscillating, then fatigue tests need to be performed to determine the possible long-term failure.

Finally, operating environmental conditions should be examined to determine possible exposure to elevated temperatures. For a typical thermoplastic material, its mechanical properties degrade with increase in temperature. As such, if there is significant thermal loading for the part in its life cycle, material characterization tests and FE simulation must account for the same. Based on the product requirements, significant temperatures can be extracted and the aforementioned quasi-static and long-term creep tests can be performed at those temperatures.  Quasi-static tests can be easily repeated at multiple temperatures with a relatively low cost. DMA tests for creep are designed to include the temperature effect by default. ASTM D 2990 based long-term creep tests need to be repeated for each significant temperature. For efficiency, these tests should be planned carefully. Typically a representative temperature can be selected and relaxation for other temperatures can be extrapolated for faster results.

In this way, by understanding the part requirements, designers/analysts can determine the type of tests needed to characterize the thermoplastic material. Figure 1 shows a schematic of this process, which starts by identifying the deformation modes of the part under operating loading conditions and incrementally adds the complexity of the loading and environmental conditions to come up with required number of tests.

Once the tests for thermoplastic characterization are determined, the actual testing process starts with injection molding of the necessary coupons. Depending on the test type, standard ASTM specimens should be molded, assuming the production process includes injection molding. Mechanical properties of the plastics depend on the quality of injection molding. As such, the supplier’s guidelines in terms of processing parameters should be strictly followed while molding the coupons. If the actual part has thicker areas than the recommended material thickness, coupons should reflect those thicknesses in order to capture process induced effects on the mechanical properties. Molded coupons should be conditioned for ~ 40 hours to the actual testing environment (temperature and relative humidity) before the actual testing. Since the thermoplastics are sensitive to the environmental conditions, it is recommended to perform the tests at a specialized lab. This ensures that accurate data can be extracted from the tests.

Basic quasi-static tests are straight forward where the goal is to obtain the material stiffness and strength properties in desired deformation mode. For FE simulation, stress-strain data from these tests should be extracted. If the application/part is subjected to high strain rate loading, these tests can be performed at corresponding strain rates to capture the rate dependence on the material properties. For DMA creep tests, the focus is on obtaining the storage and loss moduli at multiple temperatures so that the master curve constructed from the data can be used in FE simulation. In long-term creep testing, creep strains are measured over the test period for a given stress value and temperature. This data is then used to fit a constitutive model. Typical thermoplastics do not show endurance type behavior as seen in metals. In addition, due to their viscoelastic behavior there is significant ‘ratcheting’ effect under cyclic loading. To measure and capture these behaviors in a constitutive model is a challenging task and needs very high-fidelity user defined models in FE software. Having said that, a simplified S-N curve approach has been shown to work for many thermoplastics under typical loading conditions. As such, the goal of fatigue tests should be to obtain S-N curve that spans the required number of life cycles. As mentioned earlier, if the temperature and environmental conditions are critical, tests should be performed at the corresponding temperatures to obtain the accurate material properties.

Sample test data obtained from the quasi-static compressive test, compression creep tests at various temperatures and a master curve from a DMA test is shown in Figure 2.

Stress-strain data obtained from the aforementioned tests has to be accurately fitted to the constitutive models for running FE simulations. Depending on the mechanical behavior of a given thermoplastic material, elastic, elastic-plastic, hyperelastic, viscoelastic or viscoplastic constitutive models can be chosen to fit the data. For short-term loading, time-zero stress-strain data obtained from quasi-static tests is typically used to fit elastic, elastic-plastic or hyperelastic models. For long-term loadings or where the strain-rate dependence has to be considered, viscoelastic or viscoplastic models are used. Translating test data accurately to a constitutive model is a two-step process. In the first step, stress-strain data is mathematically fitted to a chosen constitutive model to obtain the model parameters. This fitting can be performed in FE software using in-built algorithms. For simple models, this can be also performed outside the tool. The second step is more of a validation exercise, where the fitted material parameters in the first step are used to a run a simulation on a test specimen under the test loading conditions.

Figure 3 shows a sample FE model of a compression test specimen used to characterize a thermoplastic material. The goal of this simulation is to validate the constitutive model used in FEA. In the first step, test data is used to obtain the material parameters for the elastic-plastic constitutive model. Compressive test is then simulated to obtain the stress-strain response of a specimen using the fitted material model. Comparison of the simulation results and test data is shown in Figure 3 where a very close agreement is observed between the two datasets. The same process should be followed for each test performed for material characterization. This approach of validation ensures that accurate properties are represented in the solver and accurate predictions can then be made for actual part performance.

In this way, by following the simplified workflow outlined in this article, designers/analysts can form a concise test plan, characterize the materials accordingly, build and validate the constitutive models to eventually predict part performance accurately and efficeintly.


John Cogger, President, Innova Engineering

Sagar Bhamare, PhD, Senior FEA Analyst, Innova Engineering

Read 508 times

Rate this item
(4 votes)

Copyright © 2017 Prototype Today ®. All rights reserved.

|   Privacy Policy |   Terms & Conditions |   Contact Us |

All trademarks and registered trademarks are the property of their respective owners.

Additive Manufacturing Today