Sponsored By

Analyzing plastics with FEA: Part 5

April 1, 1997

7 Min Read
Plastics Today logo in a gray background | Plastics Today

Process simulation technology using a finite-element model has definitely evolved in the last two decades. Its current ability to predict resin behavior under a variety of processing and design conditions allows part designs, tooling, and processing parameters to be developed and optimized more quickly and economically. To find out the benefits and limitations of today's IM simulation packages, IMM spoke with Stephen De Fosse, engineering team leader at Lexmark's Plastics Technology Center (Lexington, KY).

Cooling analysis of the housing for a string trimmer. Shaded image shows cooling lines and the mold surface temperature of the cavity at a certain point in the cycle. Cooling uniformity was critical to holding straightness tolerance.

De Fosse shared some insights derived from his "Hitchhiker's Guide to Injection Molding Simulation." He believes you can gain pertinent information from computerized simulations if they are conducted properly. "Using a finite-element mesh on a centerline wireframe geometry model, part designers can run a series of process iterations to find out the best processing scheme and its effect on ultimate part characteristics.

"The goal is to optimize each phase of the molding process, employing a systems approach to examine the relationship between materials, tooling, and pro- cessing," he notes. "The benefits include less material usage, faster cycle times, and improved part performance."

Key Elements

Values for many of the design, tooling, material, and processing variables that affect an injection molded part can be built into the meshed model or can be user specified for each analysis, according to De Fosse. These include inlet melt temperature, coolant temperature, injection time, gate and runner dimensions, and others. Some of these can be derived from material characterization data.

Analysis results will then generate key dependent variables such as injection pressure and pack/hold times. To obtain full value from the simulation process, he cautions, you must interpret the analysis output and report it in practical terms to all engineering functions.

  • Material characterization. Different phases of the analysis require different sets of data. For example, flow phase simulation requires a series of shear rate vs. melt viscosity curves generated at three or four temperatures spanning the recommended processing range. Also, you need to input items such as heating and cooling rates, Tg, thermal conductivity, CTE, specific heat, and ejection temperature.

    For pack and hold phase analysis, you'll need the PVT (pressure, specific volume, temperature) relationship of the resin being molded. Cooling analysis starts with a model of the cooling channels, then requires input on mold material and coolant. This is then combined with the design model plus data already generated from flow and pack analyses. Essentials for warp and shrinkage simulation include bulk modulus, in-plane shear modulus, and Poisson's ratio in two directions.

    Now for the $64,000 question - where do designers find all of this data? De Fosse admits that few raw material suppliers will be able to come up with all the types of information needed. Private testing labs or qualified universities can fill in the gaps, but if a complete characterization is required, the price can run more than $5000 for one resin.

    "Still, existing databases of leading simulation programs have thousands of resin grades characterized at least for flow analysis. There are fewer resins characterized for pack, shrink, and warpage analyses, but their numbers grow regularly," he says. If dimensional tolerances are standard and minor inaccuracy can be tolerated, designers can sometimes substitute a generic set of data to avoid material testing fees and the one to three weeks it requires. "Remember, however, that you run the risk of error in the analysis," De Fosse notes.

  • Model building. This step begins with whatever CAD information is available. It's best to work from a solid model generated by a design system fully compatible with yours. Barring that, you can import IGES wireframe or surface data. As a last resort, you must build the model from scratch based on 2-D CAD data or dimensioned drawings. This last option can double the time it takes to complete an analysis; on the other hand, cleaning up a bad mid-surface can take just as long.

    Two factors are critical to accurate model building - the level of model detail selected and proper mesh density. How much detail you need depends on the part design and your goals for simulation. Also, remember that more detail equals more time and cost for the analysis.

    Shrinkage and warpage analysis of a narrow housing. The four shaded images show the predicted distortion in the x direction (top left), y direction (bottom left), z direction (top right), and the total distortion (bottom right). Correlation to actual molded parts was good on direction but underpredicted on displacement.

    In addition to relying on experience, De Fosse recommends adding features near probable gate locations, identifying potential backfill situations, and modeling draft angles on ribs and bosses as well as any feature that projects more than the wall thickness with a height-to-thickness ratio of 2:1.

    Mesh density rules also apply here. For example, tunnel gates need a minimum of five elements to represent the taper. Meshes surrounding gate elements should be denser to capture the rapid change in melt velocity at the start of cavity fill. At times, higher density is also helpful at the end of fill to reflect melt front velocity acceleration as cavity volume approaches zero.

  • Flow or filling phase. The main goal for this phase is to control melt front advancement so cavity filling is balanced; that is, the entire perimeter of the melt front from each gate reaches the end of flow at the same time. This ensures that pressure is evenly distributed to help avoid warpage.

    Analyses at this stage can determine several parameters, including gate and runner location, size, and geometry needed for a balanced fill. It's important to change over from filling to packing phase simulation at between 95 and 99 percent of cavity fill. This avoids a pressure spike when injection speed returns almost instantly to zero at the point where the cavity volume is filled.

    Filling analysis of a wide-format plotter cartridge carrier. The shaded image shows a three-plate mold system and predicted melt front advancement plot at the end of the injection phase. Locating and sizing gates for cavity filling balance and verification of filling small thin details was the objective of this analysis.

    Other processing parameters that can be controlled using flow simulation include material temperature, injection pressure, shear stress, and shear rate. Temperature limits for most polymers are available from material suppliers. Injection pressure guidelines depend on the material, application, and equipment, but a safe rule of thumb is to minimize this parameter. Shear stress requirements are more nebulous. Different polymers can tolerate different amounts, and uneven shear stress distribution can indicate weak spots or potential distortion.

  • Shear rate is a bit clearer. Shear imparts frictional heating, thus lowering melt viscosity. Too much shear heating means degradation. So controlling the amount of heating by limiting shear rate is essential.

    Filling phase simulation can also indicate where vents should be located by illustrating areas of potential gas trapping. Knit line formation can also be spotted this way, and controlled to some degree by changing process conditions and gate configurations.

  • Packing and holding phase. The primary reason for running pack phase analysis is to look for temperature and pressure gradients across a part's primary walls. Gradients can result in density shifts, which lead to anisotropic shrinkage and thus warpage, sinks, voids, molded-in stresses, and other defects.

  • Cooling phase. This analysis lets you look at the way mold surface temperatures change as a function of time during the molding cycle. The difference between core and cavity wall temperatures contributes to how much warpage can be expected. Cooling analysis also helps to determine the ideal cooling time for a molded part, and helps evaluate potential mold designs.

  • Shrinkage and warpage. Tools to analyze shrink and warp are fairly recent developments, according to De Fosse, and certain features are still being refined. In his experience, the accuracy of results from these software packages can vary from high to low.

Sign up for PlasticsToday newsletter

You May Also Like