Mould flow simulation can predict filling, packing, cooling and warpage before steel is cut, if it is set up with the right data and read with experience. What it does well, where it misleads, and how to use it on real projects.
The mould flow report was beautiful: colour plots, animations, a confident summary that filling was balanced and warpage would be under 0.3 mm. The first moulded parts warped by more than a millimetre. When the analysis was re-run, three things had gone wrong: the material data was for a generic grade rather than the specific glass-filled nylon being moulded, the cooling channels had not been modelled, and the analysis assumed a uniform mould temperature that the real tool could not achieve. The simulation had been accurate about a mould that did not exist.
Mould flow analysis is one of the most valuable tools in plastic part and tool design. Like any simulation, it is only as good as its inputs and the experience of the person reading it.
What mould flow does well
- Filling: predicting whether the part will fill, the fill pattern, pressure needed and where weld lines and air traps will form.
- Gate location: comparing gate positions quickly to balance flow, control weld line locations and reduce pressure.
- Runner balancing on multi-cavity and family tools.
- Packing: predicting volumetric shrinkage and the likelihood of sink marks and voids.
- Cooling: identifying hot spots and comparing cooling layouts.
- Warpage: predicting the magnitude and direction of distortion, and separating its causes: differential cooling, differential shrinkage and fibre orientation.
- Fibre orientation in filled materials, which can be exported to structural FEA for more realistic strength predictions.
Where it misleads
- Material data. Results depend heavily on accurate viscosity, PVT and mechanical data for the exact grade. Generic data gives generic answers.
- Geometry. Small features, thin sections and sharp details may be meshed poorly or simplified out.
- Boundary conditions. If cooling, venting, hot runner behaviour or actual machine capability are not modelled, results describe an ideal world.
- Absolute values. Predictions are often better at comparing options, such as gate A versus gate B, than at forecasting exact numbers.
- Process reality. The real process will be set by people on a real machine, which may not match the simulated settings.
Choosing the right level of analysis
- Fill-only analysis: quick and useful for gate location, fill balance and weld lines early in design.
- Fill, pack and cool: needed for sink, shrinkage and cooling layout decisions.
- Full warpage with fibre orientation: essential for large, flat, precise or glass-filled parts, and before committing to tight tolerances.
Using it well on real projects
- Run it early, when design and gate changes are still cheap, then again on the final tool design.
- Use specific material data, and model the real cooling layout and runner system.
- Ask clear questions: where should the gate go, will this rib sink, which cooling layout reduces warpage?
- Compare options rather than relying on single absolute predictions.
- Validate against the first trial, and feed what you learn back into future analyses.
Reading a report critically
A useful mould flow report states the material grade and data source, the mesh type and quality, the process conditions assumed, what was and was not modelled, and clear recommendations. If a report is mostly colourful pictures with no recommendations, ask for the decisions it supports.
Used well, mould flow is a way to buy certainty cheaply, the same principle we described in prototype or simulate. Used poorly, it is an expensive way to feel confident.
