Transforming Engineering

The list of industries experiencing huge disruption fueled by “technology” is growing by the day. Insurance agents and therapists are replaced by apps. Tesla breaks the dealership model. Commercial real estate fears a spare bedroom and Zoom account. My career of mechanical design consulting is less of an industry than a tool, but it too is changing rapidly. I have tried to stay up on the trends so that my team and I don’t go the way of the insurance agent. Here are a few of the new tools and methods we are tracking, learning and (when appropriate) applying in our client work. I really do not know which ones will rise steeply in the next year or so and which ones require a lot more time to take hold. These are organized into evolutionary (capable of speeding up and extending the processes we already use, like breeding Clydesdales to pull a heavier wagon) and revolutionary (capable of changing our processes or letting us make knew things with new processes, like inventing an internal combustion engine, eliminating horse drawn wagons).

Evolutionary Tools and Methods

Web based CAD

If product design is a nail, solid modeling is our hammer. I have never needed to draft on a board or design in 2D, but I’ve seen SolidWorks eat ProEngineer’s lunch by taking solid modeling off of the workstation and onto the Windows platform. OnShape and AutoDesk are primed to out-SolidWorks SolidWorks and move solid modeling from the desktop to the cloud. This change is inevitable, but also painfully slow since CAD platforms are almost as sticky as accounting systems. While this shift will empower a few interesting new workflows (like design branches and collaborative design) this move won’t rock any of our worlds or welcome in a new hardware revolution. And as a design consultancy, we will be one of the later adopters as we must use the platforms of our clients. Startups and new grads will learn on OnShape and Fusion (my 6th grader knows it better than I do). SolidWorks users will be in demand to maintain legacy designs as the company goes the way of Friendster if they don’t watch out.

Model based definition

Solid modeling is fantastic at creating form and integrating with prototyping methods like 3D printing. But high-volume manufacturing will always require additional documentation to communicate color, material, finish, fit and other critical to quality characteristics. Making stuff will always have variability so nominal design is just the start. For all of my career I’ve heard that drawings would disappear, and while “prints” are less common, PDFs of 2D drawings with traditional borders, tolerance blocks, datums, GD&T and the tolerance stacks that drive them are just as critical as ever. Model based definition, if adopted, should eliminate these PDFs and all their drawbacks (association with solid models, revision control, formatting, legibility, etc.) while not losing the details captured. The same meta data that the parts and assemblies need can live with the solid model. New standards and interoperability that the PDF has perfected over time is required and Adobe is poised to provide that with their 3D PDF standard. A simple export to a 3D PDF from your CAD system can include data about material, finish, color, tolerance and even assembly process. It will also be easier to interpret by less skilled manufacturing labor and quicker to review by product development teams. This step is one of those friction reducers that seems minor but provides jet fuel to a process.

Extreme analysis including digital twins and machine learning

Faster computers, cheaper sensors and AI has moved high-end analysis and simulation out of the rocket and car design world into the hands of the rest of us (in terms of cost and time). Gaining a greater understanding of the design margins and unexpected behaviors of your products before you build the first can greatly reduce the time to market and increase the confidence that what you deliver will be of high quality. For those engineers and on those projects that haven’t taken advantage of FEA, CFD, and math modeling this is a step change and perhaps this belongs in the next category. But fundamentally better analysis tools still require a mature design before they can be applied so they do not fundamentally upend the standard design-prototype-test loop even if the prototype is digital instead of physical.

Revolutionary Tools and Methods

Generative design

This design-prototype-test loop IS upended by generative design, a truly mind-blowing technology. We have experimented with a couple of the leading software offerings – ParetoWorks and Fusion. ParetoWorks requires a primitive form and one or more load sets and gives you a minimal form optimized for strength or stiffness. Fusion similarly requires a primitive form and set of loads but offers up many alternative forms that all can carry the loads (though may not be as lightweight). The outputs of generative design are not finished products. They don’t really take manufacturing process into account and are not parametric models. They serve as feasible “underlays” for industrial designers and in some cases act as inspiration to those same designers. This flips the process, pre-qualifying designs before they reach the engineer allowing confident form creation with minimal risk of tear-ups after an aesthetic has been chosen by the team.

Mixed reality, VR, AR design evaluation

In recent years, no technology has shown greater potential to speed up and cheapen the decision-making process than XR (VR and AR). One tenant of the design thinking process is creating alternatives and testing those alternatives before committing to a single path. This evaluation has often required the creation of multiple prototypes and due to budget limitations, these prototypes must be simplified (dumbed down or scaled down) in the early stages. With VR, several high-fidelity models can be rendered in VR, at scale and placed into different use environments or states allowing users to experience and compare them. We have been experimenting with overlaying virtual models over very low fidelity models equipped with position sensors allowing manipulation of the virtual model using real affordances. We are just scratching the surface of what XR can do. Tools like Gravity Sketch allows high quality solid modeling inside the VR environment allowing real time evaluation and iteration.

Design for additive manufacturing

3D Printing has followed a decades long hype cycle which created cynics of many of us. Sure it’s invaluable to create early prototypes, but the surface finish, strength, material characteristics, and slow build speed have kept it as a proto-only option. In the background, brilliant folks have been solving these problems and closed in on some early production-at-scale applications. Getting a completely new, ever-changing manufacturing option is amazing and a little intimidating. It is hard to know if we are recommending something better or not quite ready for prime time. Design for additive can be easier than designing for injection molding but it does have a new skill set. You no longer need to worry about draft, action, texture limitations, and cavitation, but now your design choices can dramatically affect nesting, build time and cost. You also must stay on top of the machine capabilities, material choices and extras like texture mapping to take real advantage of the process. Additive unlocks some crazy opportunities, shortening time to market, allowing part consolidation, extreme customization, and distributed manufacturing.

The maker movement and the democratization of engineering

Maker spaces are everywhere including elementary and middle schools. Mark Rober posts viral videos about engineering principals used to prank porch thieves. Onshape is taught in middle schools around the world. 3D STL files are available free online to print on $100 home 3D printers. Kids and adults are realizing the thrill of creating something from their own ideas or at least from scratch using how to videos to learn skills, concepts, and physics. This movement is creating a new crop of engineers – drawn to the field not for its potential for earning or because they were good at math and physics in high school, but because they are great problem solvers and are excited to make things. I admit to some fear that they won’t all make the leap from making something to engineering something that is desirable, feasible and viable. But with mentorship and training it means we are starting with curious and creative folks from all walks of life. We need to seek out these curious folks and create a more diverse engineering future. This generation will create the hardware of the future that can solve the biggest problems we face today – sustainable energy and global warming, income inequality, affordable housing and urban design, responsible tech.

I’m optimistic about the digital transformation of hardware engineering. I know that software scales better than hardware and is creating wealth at an incredible clip. It is also solving some wicked problems especially with the maturation of AI/machine learning. But some of the biggest challenges require hardware – converting the sun into energy, storing that energy, getting stuff to people’s front door. Us mechanical engineers gave up our slide rules, drafting boards and pocket protectors. Let’s keep changing for the better.