Use Case: The redesign of a metal industrial flow manifold

Article Highlights

Additive Manufacturing (AM)is often portrait as a “magic” technology that can print and should print everything. However, the reality is obviously different when considering not only the technical challenges, but also the economical realities of manufacturing parts and products using in part 3D printing.

Therefore, it is important to learn for the design engineers how to reach the maximum benefits by knowing the AM design rules, and keeping in mind some of the fundamental economic drivers associated with AM productions.

In this article, we will outline the design process and strategy in order to maximize the benefits offered by metal 3D printing (in this case, Stainless Steel 316).

FIG 1.: BEFORE and AFTER design optimization using Additive Manufacturing technologies (metal DMLS)

1.Problem Statement & Objectives

This product is a manifold (distribution device) used in a cleaning installation. The part is currently manufactured using conventional machining and welding technologies, and is an assembly of 7 different parts assembled by welding. The function of the manifold is to divide an inflow over 2 different channels plus adding in the meantime a solvent to the medium as uniformly as possible. Note that the interfaces (couplings) for the inflow and outflow are flexible tubing, and can therefore be modified (location).

FIG 2: Illustration of the manifold function

The key driver of this re-design project is two folds:

a. Improve the flow performance (minimize flow resistance and maximize solvent mixing efficiency)

b. Reduce manufacturing cost (if possible)

2. Why using 3D Printing?

At this stage, we need to ask the question: WHY use 3D printing? One of the common mistake is to ignore the economics and advantages of conventional manufacturing and use 3D Printing “because we can”, up until the point when the quote comes in, which often does not even include other manufacturing cost consideration.

In this case, the main goal is to improve the performance of the flow and mixing efficiency of the solvent, and, as a side effect, ensure we either match the finished part cost, or reduce it.

FIG 3.: LEFT: Current manifold part. RIGHT: After optimization. Notice the size difference to achieve similar to better performance, and enable an economical 3D printed solution (keep budget/part roughly the same or less).

3. Design strategy

The AM design (CAD design optimized to take advantage and consider the limitations of the 3D printing process) considered the boundary conditions which, in this case, were the IN/OUT flow rates, some of the hose diameters for the interfaces, as well as the uniformity of the solvent concentration in both the outflow channels.

The choice of material was driven by the resistance to corrosion, as well as the strength requirements (chock and durability for abrasion).

a) Functionality: The emphasis was given to the functionality of the mixing process efficiency, and splitting of the IN into 2 OUT channels without constricting the initial design by considering the ways of manufacturing.

FIG 4. LEFT: Current manifold part internal structure, which is based on existing sub-components available for welding/assembly. RIGHT (not to scale): Basic functional design considering a one part build (similar to casting strategy)

b) Printability & Post-processing: The functional design can then be modified to incorporate a few manufacturing considerations/restrictions, such as:

  • Direction of printing (determined by a few factors)
  • Post-processing steps required (trying to minimize, e.g. machining or surface finishing, or support removal)
  • Post-processing accessibility (can the surface/part features be accessed to finish it?)

Notice that the “grid” mixer feature added to the outflow channel are designed at a specific angle which considers self-supporting criteria of the metal DMLS process for 316 SS, as well as the direction of print.

FIG 5: Details of the implemented solvent mixing features as part of the outflow channel design to boost up the mixing efficiency in much smaller footprint. Essentially, the “grid” will act as the mixing chamber in each out channel.

4. Tools

One of the most common question coming from a design engineer, when “flashing” an AM-optimized part, is a practical one: what CAD tool did you use to get to the final design. Today, there are a few CAD modules and new stand alone SW which aim at helping the optimization process. However, this part was optimized the “good old way” using SolidWorks CAD tool.

Note that we believe that using a topology optimization software would have led to even more optimized part.
FIG 6: Capture of the CAD design using an STL viewer 123D Make. CAD design done using SolidWorks software.

As seen in section 5, one important tool required to ensure a printable part is the support generation software. We use in this case Magics (from Materialise)

5. The realities of printing the part

Choosing the correct orientation at the beginning of the design will allow a better and/or easier support structure strategy. Supports are a very important aspect of metal 3D printing, for thermal and physical stability of the part during the build process. Supports generally means:

a) More expensive part

b) Longer print time (linked to cost)

c) More complex post-processing/finishing

FIG 7: Manifold as printed, including the support structures supporting the printed part. Directly attached to the metal baseplate, it requires to be “cut off” the plate using saw blade or wire EDM techniques.
Note also that the design should consider a way to evacuate any lose powder left inside the part cavities once removed from the printer and cut off the baseplate.

6. Getting a finished part

To get to a finished part, there are several steps to consider. Specifically for this part, we had to:

a) Remove the printed part from the printing baseplate (using band saw and/or breaking using hand tools)

b) Remove the printed part from the printing baseplate (using band saw and/or breaking supports with pliers). Note that the baseplate will need re-surfacing before reusing.

c) Remove supports from the part, by breaking joints using pliers.

d) To render smooth surfaces, process the surface where the supports were in contact with the part, and left surface peaks.

e) Machine the critical interface areas (fig. 8, marked with yellow arrows) to ensure correct mating with interface hoses and avoid leaks. The key is to post-process ONLY the critical surfaces.

FIG 8: Areas requiring post-processing to ensure correct interface tolerances (surface finish, flatness, dimensions, etc.)

f) Bead blasting of the printed metal surface, to smoothen the surface finish by removing semi-sintered powder particles. This leads to the part show in Fig.9.

FIG 9: Surface finish obtained after bead blasting using Alumina beads.

7. And then, the testing!

Any part development is not complete until it undergoes a functional test. This is not unique to 3D printed part. In this case, the functional test involved the assembly of all hoses and flow apparatus, and then test for leaks, pressures, solvent concentration in each outflow channels, as well as form and fit checks.

Note that in several critical application use case (not required in this case), a CT Scan (an X-ray image made using computerized axial tomography) is required to identify possible defects built-in the part during the printing process, which could lead to catastrophic failure.

Overall Assessment

In this case, the objective was not necessarily to produce a cheaper part, but a more performing one. By combining a complex mixing structure, the solvent concentration is more uniform, while a more compact structure.

As a result, although the cost of the part is relatively the same, the assembly steps and associated labor is significantly reduced, which translate into a part lead time reduction. In this case, the labor cost reduction is counterbalanced by a slightly increased cost of the components (printed part Vs shelf components).

But the most important remains: the client is happy!