Many fields have witnessed the emergence of a revolutionary vision that not only represented a sharp break from the past but also reshaped their development for a period of many years. In the field of mechanical design software, where new products with hundreds of innovative features are introduced every year, the development of parametric, feature-based, fully associative, solid modelling technology has played such a role. Twenty-years ago, engineers were thinking through the formulas for geometric shapes as they positioned 2D or 3D entities, with full knowledge that unforeseen elements in the design would be likely to force them to start over from scratch many times before they were done.

Then Samuel P. Geisberg, a mathematics professor who left Russia to come to the United States, had a vision of a better way to do mechanical design. His goal was to develop a modelling system using features and parameters, a method of linking dimensions and variables to geometry in such a way that when the parameter values change, the geometry updates accordingly. With this innovation, many design concepts could be explored and changes could be made remarkably quickly compared with the redrawing required by traditional CAD. It’s not hard to make the case that the development of mechanical design software since that time has merely represented the fleshing out the original parametric vision. This article will explore the emergence of parametric, feature-based, fully associative, solid modelling and its continued development with new features and capabilities that have kept it at the leading edge of mechanical design to this very day.

Before parametric modelling

The first commercial computer aided design (CAD) tools, which were introduced in the 1970s, were primarily a replacement for the drawing board. So, while they improved productivity and accuracy of drawing creation, they did not have a major impact on the mechanical design process. A considerable amount of geometrical and trigonometric calculations were required to create even relatively simple components. The surface and solid modelling tools that followed from the late 1970s to the late 1980s extended the drafting paradigm by allowing engineers and designers to draw lines in 3D space. The result was wireframe models that could later be patched with surfaces. The extension of the drafting paradigm to 3D space had the effect of increasing the complexity of the required calculations to the point that considerable background in mathematics and surface geometry was often required to effectively utilise the tools.

Another basic problem with the earlier generation of design software was that all the geometry was explicitly created in reference to a spatial coordinate system. This meant that changes to the design often required that major sections of the design, or in many cases the entire design, be re-created from scratch. For example, consider the case of a 1/4 inch thick plastic moulding with hundreds of holes, bosses, and ribs. Let’s suppose that at some point in the design process the decision is made to increase the thickness to 3/8 inch. Now the through-holes in the part have suddenly become blind, and bosses and ribs have become partially buried. Substantial and time-consuming changes needed to be made merely to reproduce the original design intent. This was the situation faced by every computer-aided design software user up to the point when Geisberg entered the scene.

Turning vision into a product

Arriving in the United States in 1974, Geisberg worked for two of the pioneering computer aided design firms, Computervision and Applicon. He became frustrated with the limitations of traditional CAD systems and spoke to his employers about developing a system that was less rigid and more flexible. His employers were focused on making incremental improvements to their existing products so Geisberg obtained venture capital backing from Charles River Ventures and other investors and founded Parametric Technology Corporation, now PTC, in 1985 to pursue his vision. Charles River brought Steven Walske in as Chief Executive Officer in 1986 and the Pro/Engineer product was shipped in 1988, marking the beginning of a new era in mechanical design.

One key difference between Pro/Engineer and all other CAD tools up to that point is the way geometry is modelled. The software contains solid primitives called features consisting of common engineering shapes such as holes, slots, bosses, fillets, chamfers, protrusions, shells, etc. These features know how to behave relative to each other and are defined by a set of parameters. Rather than drawing the geometry line by line and arc by arc, the software prompts the engineer for specific constraints. For example, Pro/Engineer prompts an engineer for the surface to begin a through-hole and the diameter, which is a parameter, of the through-hole. It doesn’t ask for a depth because the software knows that a through-hole always goes through the full depth of the material. This means that when the depth of the material changes, also a parameter, all the through holes referencing that material will update automatically. Pro/Engineer is parametric in that objects are positioned using parameters and mating constraints rather than the coordinate system. This means engineers can change dimensions or move objects and the connected geometry will automatically move or change itself to stay in sync.

Streamlining the design process

This invention dramatically streamlined the mechanical design process, making it practical for engineers to create many more design variations at a higher level of integrity in a fraction of the time. Going back to the earlier example, when the thickness of the plastic part is increased, the through-holes will automatically adjust to go all the way through the thicker part, the bosses and ribs will move so that they remain affixed to the surface of the part, and so on. The practical impact of the parametric modelling concept is that engineers can now change designs and modify dimensions with a few keystrokes or mouse motions. The underlying geometry of the part automatically adjusts itself to accommodate the change.

The initial parametric vision included associativity, where changes ripple through the design and all related deliverables. Up to that point, CAD software was similar to the drawing board in that each drawing or model was independent of the others. Drafters had to wait until the design was done before they could make 2D manufacturing drawings. If one view of a three view set of manufacturing drawings changed, the others had to be manually changed as well. The advent of associative CAD meant that downstream tasks such as drafting could be started in parallel with upstream activities such as design, thereby shortening development time. This was the birth of concurrent engineering. Whenever the design changed, the manufacturing drawings and other downstream activities referencing that change would automatically update.

Changing the ground rules

If that was it, the concept of parametric modelling would be a milestone in the development of mechanical design software, but not something that we would be talking about today. But the parametric concept so changed the ground rules in mechanical design that it has ushered in a stream of major developments that are continuing to drive software development to this day. Each of these developments flows from the original concept of embedding intelligence into the features that define a part thus automating the tedious aspects of mechanical design and allowing the engineer to focus on the creative process.

The expansion of the original parametric modelling concept was driven by customer demand. One of the first improvements that customers asked for was assembly modelling. Assembly modelling was a natural extension of the parametric modelling concept that essentially provided the same advantages to complex assemblies that the initial concept had provided to component parts. Soon after parametric modelling reached users, versions became available that provided hierarchically linked assembly layouts designed to simplify the conceptual design of complex assemblies and relate all base components using global dimensions, relations, and common datums. Users can use either graphics or spreadsheets to adjust parameters for the entire assembly to play out “what-if” scenarios for different conceptual designs and perform packaging studies to represent assembly components.

Moving downstream

The next major logical extension was downstream into the manufacturing arena. Pro/Engineer modules were introduced that utilise the original product design definition to delineate the process steps and operations required to build the design. The principle of associativity was extended to allow interactive changes to the design model to propagate through the development of the manufacturing tool paths, which reference the part model, and update manufacturing instructions and documentation. The final step was typically the creation of code that was used to drive computerised numerical control (CNC) machinery such as turning centres and machining centres. Later enhancements extended this concept to other machining operations such as sheet metal forming and blanking.

As parametric modelling tools became the standard for mechanical design, their use expanded into industries with specialised requirements. For example, designers of automobile bodies and consumer electronics packages, among other products, need to define the flowing surfaces that often distinguish cutting-edge industrial design. In the past, designers of these products typically used sketchpads, modelling clay and foam, and specialised surface modelling software to define the product geometry. Mechanical engineers usually wound up with the challenge of mathematically defining the geometry to a level of precision that enabled it to be economically manufactured. These requirements were addressed with the development of parametric modelling tools with powerful surfacing capabilities needed to turn out the complex surfaces that characterise cutting edge consumer products. The parametric surfaces produced by these tools are designed in relation to the rest of the model so as the design changes, so do the surfaces, and vice-versa.

Integrating analysis

By enabling engineers to generate a vastly higher number of design alternatives in a relatively short period of time, parametric modelling raised the question of how engineers would determine which designs were better than others. Clearly the time and money did not exist to physically build and test all of these designs. Computer aided engineering tools that could analyse the performance of a design from a structural, thermal, fluid flow, and other standpoints predated the original parametric modelling concept. The critical innovation that appeared less than a decade after the original parametric concept was linking the parametric modeller to the analysis tool so that engineering could optimise performance, reduce manufacturing costs, improve quality. In this way, engineering analysis becomes an integral part of the design process.

As competition increased, manufacturers began developing products in a globally distributed environment, creating the need to collaborate across time zones and between different companies. Product Data Management (PDM) software used to store, control and provide access to design models and other engineering information was developed to address this need. But its usefulness was limited by the fact that it was developed separately from design software. An important milestone in the development of mechanical design was the integration of PDM capabilities with the core parametric modelling software so they both work as a single integral system. Using this system, companies can create highly accurate digital products, collaborate digitally throughout their extended value chain, control all associated product information and processes, and communicate via dynamic technical publications that reference the right version of the model. This product development system becomes expandable, linking legacy applications and heterogeneous CAD systems while protecting past IT investments. These newer capabilities have become collectively known as Product Lifecycle Management (PLM) solutions.

Each of these developments, and many others that space does not exist to include here, flowed logically from the original concept of simplifying the mechanical design process by providing a mathematical framework that maintains logical consistency between the different elements of the design. Parametric, feature-based, fully associative, solid modelling, which began by revolutionising the way that engineers define the geometry of mechanical designs, has been integrally linked with data management and collaboration to dramatically improve the entire product development process. The result is that engineers have the power to model any product design from simple to complex, define products to a higher level of precision, create more design alternatives, predict the performance of those alternatives without the need for a physical prototype, and manage the entire product development lifecycle. For these reasons, parametric modelling remains in exactly the same position that it was in 20 years ago, at the cutting edge of mechanical design.