the polymer pages

Moldflow Leverages a Decade of Mold Trials to Augment Extensive Product Validation Efforts

By Robert Newman, Moldflow Corporation

---

Knowledge becomes wisdom only after it has been put to practical use.  ~ Anonymous

---

Moldflow developers and quality assurance engineers employ a wide array of case studies and validation tests on every software version to confirm the accuracy and consistency of simulation results. These include numerous studies on many parts from a wide range of industries such as electronics, appliance, automotive and telecommunications. Recently, Moldflow has augmented these validation and testing cases by capitalizing on the value of historical material testing and mold trial studies.

In the past decade, Moldflow Plastics Labs has tested over 5,000 polymer materials for CAE analysis. These tests included mold trials and measured shrinkage on 894 non-confidential resins. This data represents a wealth of knowledge concerning polymer material shrinkage behavior. By compiling this data and automating an unprecedented number of MPI® simulation runs and data comparisons, Moldflow has put this knowledge to a new practical use: validating improvements to code and data.

For every resin studied, mold trials were performed such that process parameters affecting shrinkage were varied over a broad range of values. These parameters include: injection rate, melt temperature, mold temperature, packing pressure, packing time and part thickness. A total of 25 to 30 sets of processing conditions were studied for each material.

As a result, experimental shrinkage data in both parallel and perpendicular directions was available from more than 23,000 molding experiments. This data was compiled and tools were developed to automate the running of more than 23,000 MPI simulations in order to establish software predictions corresponding to each molding experiment. The simulation results were then compared to the experimental shrinkage data. From this, the parallel and perpendicular shrinkage prediction accuracy could be determined for each of the cases.

One valuable way to look at the results was to consider how many times the MPI solution predicted a result within a desired level of precision. The percentage of simulations that obtained a certain desired level of accuracy was easily determined from the compiled results.

By repeating this process with a new MPI software version or different material models selected, the effect of changes to software or material data models can be clearly seen. In the last year alone, Moldflow has compiled well over 200,000 simulation results in this manner to confirm MPI product improvements. For example, the table below shows accuracy improvements with each major release of MPI software when comparing predictions to experimental values of shrinkage perpendicular to the flow direction.

Table 1. Percent of 23,299 mold trial simulations that predicted perpendicular shrinkage to within 10 percent of experimental values when using the Corrected Residual In-Mold Stress (CRIMS) model* for three major releases of MPI software.

MPI 2 63.19% MPI 3 73.75% MPI 4 75.30%

Note: for more information on the CRIMS model, refer to the March 2002 issue of Flowfront, "CRIMS, the Benchmark Standard for Shrinkage Prediction."

By plotting the percent of simulation runs that obtained a given level of accuracy against the accuracy level compared to experimental results, the relative change due to software improvements can be shown graphically; see Figure 1.

Figure 1. Data comparing 55,601 simulation runs to experimental shrinkage data, showing the percent of simulations obtaining a given level of accuracy in predicting shrinkage perpendicular to the flow direction using 894 materials with the CRIMS model, for three major releases of MPI software.

The effect of different shrinkage models on prediction accuracy also can be seen in Figure 2.

Figure 2. Data comparing 75,756 simulation runs to experimental shrinkage data, showing the percent of simulations obtaining a given level of accuracy in predicting shrinkage in the flow direction using 894 materials with the CRIMS model versus the uncorrected residual stress model for two major releases of MPI software.

The data was further analyzed so that effects of material types and fillers could be distinguished. This allowed for specific verification of code improvements affecting certain types of materials or fiber orientation predictions. In all, the automated process to run simulations and the resulting data allowed for additional checking of:

• Build-to-build consistency and improvements (MPI 2.0, MPI 3.0, MPI 4.0)
• Consistency of results based on platform usage (UNIX versus PC platforms)
• The effect of proposed code improvements
• The effect of viscosity and shrinkage model usage
• Prediction accuracy versus material type
• Prediction accuracy versus filler

Because the analysis results could be sorted and compiled by material type and filler, Moldflow was able to provide this information directly to MPI users within the MPI 4.0 environment. This gives MPI 4.0 users further insight into the accuracy of the software for their material type, as well as the effect of material models on the results; see Figure 3.

Figure 3. The Material Data Method Report in MPI 4.0 software allows users to review information about how available material data may affect simulation results.

Moldflow is now extending this validation work to include similar assessments of pressure prediction accuracy by utilizing our library of over 43,000 molding trials showing cavity pressure traces for approximately 1,800 materials. As simulation technology and material data requirements evolve, we are committed to validating the impacts of changes to deliver improved products to our customers.

Material suppliers who use the MPL/Data Fitting utility can flag materials that are approved to be included in subsequent releases of the Moldflow Material Database, which streamlines the process of database update and revision. Moldflow customers can find database updates available for download from the Moldflow Community Center.

For more information about Moldflow Plastics Labs material testing services and Moldflow material databases, send e-mail to mpl@moldflow.com or visit www.moldflow.com/services/mpl.asp.