Skip to content

2.1 The software tools of the trade

In this chapter we introduce the tools used in DFM engineering and the vocabulary of AI that the rest of this study relies on. The aim is to map the scope of AI in this field — what would count as AI at all, and which parts of the work it could touch. We are not yet asking how far any company has actually taken it; that comes in Chapter 3, once there are real companies to look at. Chapter 1 described what a moulded part is, how it is made, and the rules it must obey. This chapter is about the software that does that work, and how to think clearly about where AI could enter it.

We start with the tools, because you cannot ask "where could AI fit?" until you know what software already sits in the workflow. The aim here is only to name the categories and what each is for — not to compare products or vendors.

2.1.1 The five software categories

Every moulded part passes through roughly the same set of software, each category attached to a stage of the development process described in Chapter 1:

  • CAD — computer-aided design. The tool that authors the part's geometry: the walls, ribs, bosses and rounded corners themselves. This is where the design decisions that Chapter 1's design rules govern are actually made. (Examples: CATIA, Siemens NX, Creo, SOLIDWORKS.)
  • DFM checkers. Software that reads a finished CAD model and flags where it breaks the design rules — a wall with no draft, a rib that is too thick, an undercut that will need a side action. This is the direct software embodiment of §1.3. (Examples: DFMPro, DFMXpress.)
  • Mould-flow simulation (CAE). The physics engines that predict how the melt will fill, pack, cool and warp before the mould is cut — the tools §1.2 introduced as the reason simulation exists. (Examples: Autodesk Moldflow, Moldex3D, Cadmould, SIGMASOFT.)
  • Optimization and "generative design". Tools that automatically search a space of shapes or process settings for a better one — lighter, stiffer, faster to fill. They automate the trying of many options. (Examples: the topology-optimization and DoE modules built into the CAD and CAE suites above.)
  • PLM — product lifecycle management. The system of record that stores every version of the design, tracks changes, and carries a part from release to production. It is the filing cabinet the other four tools write into. (Examples: Teamcenter, Windchill, 3DEXPERIENCE.)

2.1.2 What to notice about this toolset

Two things are worth holding on to as we go into the mechanisms.

First, the toolset is mature and genuinely capable. None of it is new, and none of it is broken. A modern mould-flow solver is a remarkable piece of engineering; a good DFM checker catches real mistakes cheaply. Whatever AI adds, it is added on top of a stack that already works — not to a vacuum.

Second, almost all of it computes rather than learns. A DFM checker applies rules a human wrote. A mould-flow solver applies the equations of physics. An optimizer searches, guided by those same rules and equations. None of them gets better at the next part because of the last one; none of them carries knowledge between jobs. That single property — computes, does not learn — is the hinge this whole chapter turns on, and the next section makes it precise.