Additive Manufacturing, Topology Optimization and 3-D Printing

Additive manufacturing (AM), topology optimization and 3-D printing have produced some remarkable changes in the manufacturing sector, enabling companies to make parts whose geometries would have been all but impossible using traditional techniques. Still, being a relatively young technology, AM faces some challenges before it can enjoy more widespread use.

One of our largest customers has categorized these challenges, in order of importance to them:

  1. Build volume (the size of the part that they are able to make): While tiny and complex is interesting, large and complex is where the money is.
  2. Build speed: Even with a new facility with 100 DMLS (Direct Metal Laser Sintering) machines to print one part, 24 hours a day, every day, this customer can’t manufacture parts fast enough to keep up with demand.
  3. Build failures: These are caused by thermal distortions, build plate connection failures, interference with the powder spreading mechanism because of distortion, etc.

We’ll discuss the obstacles and potential solutions for the first two of these items today. In a future blog, we’ll take up the issue of build failures and one other industry-wide concern:

  1. Confidence in the material properties being produced: How can we be sure that a part made using AM has consistent mechanical properties throughout the part?

Let’s take a closer look at the first two issues.

Additive Manufacturing and Build volume

Parts made by AM tend to be tiny. These parts have incredible resolution, with details that could not be made by any other process — not casting, and not subtractive machining. Metal AM started out as a bit of a novelty item, great for printing desktop trophies to show things like metal lattice structures balanced on the head of a dandelion gone to seed, but not really very good at printing parts that actually did anything useful. Where was the problem, why were the build volumes small?  The difficulty was in scaling up the internal mechanisms for the DMLS machines — the powder spreaders and the laser controls.  Making these mechanisms larger while maintaining the incredible resolution required to produce precision parts proved to be problematic.

topology optimization

Highly-detailed Additive Manufacturing part

But that’s changing, and relatively quickly. Smart engineers are producing ever larger control and material handling systems, while maintaining the precision for which the AM processes are known. ExOne, for instance, is producing machines that print not the part, but the mold for traditional casting, and these molds are nearly the size of a pickup truck bed. Also, DMLS machines, the predominant commercial choice of aerospace companies, are getting larger and larger, so much so that it’s possible to print multiple versions of the same part at the same time, with no loss of the precision that would sacrifice surface finish, material properties, etc. It’s obvious that engineers are quickly solving the scale-up challenge.

Additive Manufacturing and Build speed

While the precision parts, particularly in metal, that are coming out of today’s AM machines are marvels to hold, spin, bolt to your car, etc., it can be maddening to watch an AM machine at work. Many builds take up to 8 hours. Two of the biggest factors in build speed are layer thickness and laser travel velocity. Smart companies have already started tackling this problem from a number of different angles. Several companies have put multiple lasers in the same machine, with multiple control mechanisms. Productivity scales nicely when you do that. So, if one laser is good, and two lasers is better, then what about three?

This reminds me of the razor blade wars that went on for a while. For many years, there was one razor, then there were two and it was a huge breakthrough in shaving speed and reduced bloodletting potential. Then there were three blades, then four. I have no idea how many blades are in vogue now, whether they are stationary, or vibrate or what, but it’s billions of dollars in market value for something that seems pretty simple, in hindsight.  (Why didn’t I think of that?)  But, back to the point, Additive manufacturing production speeds are going up rapidly, and, no coincidence, so are the number of AM parts being put into service as final production parts.

So, additive manufacturing may have some problems, but problems with great promise attract great minds, and these minds are knocking down the barriers faster than most thought possible. Some forward thinking companies are racing into the age of additive manufacturing, while others are cautiously optimistic and taking a wait and see approach.

Additive manufacturing at the University of Pittsburgh

Clearly, there is much that remains to be done in the future. ANSYS took a step toward that future by partnering with the University of Pittsburgh to establish the ANSYS Additive Manufacturing Research Laboratory at the Swanson School of Engineering in June 2016. Researchers collaborating with ANSYS engineers in the laboratory will put the power of ANSYS simulation solutions to overcome some of these challenges.
optimized topology

Optimized topology using ANSYS Mechanical (created by CADFEM for ANSYS)

Also, the recently released ANSYS 18 contains a new topology optimization feature in ANSYS Mechanical can define the region and loads for a given scenario or multiple scenarios, and let the solver use a physics-driven approach.

For more information watch this webinar and learn about all the enhancements to ANSYS Mechanical, for additive manufacturing and for other engineering applications. And watch for our second blog in this series, in which we’ll tackle challenges 3 and 4 listed above.

5 thoughts on “Additive Manufacturing, Topology Optimization and 3-D Printing

  1. One more challenge that is important to a lot of manufacturers is the SURFACE ROUGHNESS PROBLEM.

    Not only is the surface roughness higher due to the nature of the powder metals being used and their particle size, but it also inherently means that you have to slow down your process in order to produce a smooth surface of the part because the “ridges” created on the part are inversely proportional to the speed of manufacture.

    A very large part of Additive Manufacturing challenges are in post-processing. If it’s a powder, how do you clean off the particles? If it’s direct-write with a wire feed how do you debur? A lot of R&D is actually going into this area now.

    Additive Manufacturing was first introduced in 1986 and a lot of challenges have been empirically figured out, like the powder metal composites used and the binding material, to stabilize the shape in a powder based process, however, the core technology problems that you mention are large hurdles in this area and a lot of innovation is needed to discover the solutions.

    Which is why I believe that ANSYS simulations will play a CRUCIAL PART IN THE SUCCESS OF ADDITIVE MANUFACTURING as a whole, as well as adoption of the process in general.

    • Sarah, excellent points about surface roughness as well as the post processing aspect of AM. Surface roughness is being tackled in a number of ways, most of which are proprietary to the printer manufacturers and involve novel laser scanning techniques, hence another need for multiple lasers in a single machine.

      You obviously are well versed in AM, and I’d enjoy connecting with you to discuss your experiences that allowed you to arrive at your conclusions. If you get a chance, connect with me here and drop me a note on how to best have a conversation with you:

  2. Pingback: Tips on working with 3D printing and additive manufacturing for CAD designers

  3. Pingback: Semiconductor Engineering .:. Blog Review: Feb. 8

  4. I completely agree with this write up and Sarah Emily Yack. I work in 3D printing service Company and we mostly do short-run manufacturing. This is for sure traditional manufacturing can produce much more quantity and at faster speed. Even if we produce a small complex object for few cubic centimeters it takes hours to build. But no other technology can produce a resolution like 3d printing produces.


Comments are closed.