Using the EinScan SP V2 structured light scanner, I captured the geometry of the base pump. Lighting and surface quality proved critical—glossy, dark surfaces cause noise and distortion in the scan, so I ensured soft diffuse lighting and coated the ABS casing in a matte scanning spray.

The raw scan data was processed using Shining3D software, which produced a high-resolution mesh (~0.05 mm accuracy) for use in CAD.

In Autodesk Fusion, I firstly removed the majority of the mesh to speed up processing times.

I then used mesh-based cross sections to build up the profile of a mating interface for the lid, allowing for cable routings and screw bosses for securing the part.

This provided a good base to begin iteration.

Onr of the narrow fitment tests.

Failed orientation test.

Using PLA and the Prusa MK4S, I printed 11 narrow prototypes to dial in the fit around the main body. 

I then moved on to iterating the full shell, including experimenting with design concepts,  speeds and orientations.  

Volumetric flow rate of a printed prototype.

With Prusaslicer I was able to almost remove evidence of the Z-seam by tuning a scarf-seam.

I also experimented with using an automotive filler and primer to improve surface finish, but ultimately felt the minute gains over a good print were not worth the effort in a production scenario.

Final "production" part.

Composite 3D render of the final part.

The final part included a two-tone aesthetic insert and fulfilled it's purpose as an accurate and functional replacement part. 

It has a perfect fit, full compatibility with cable and air-hose routing and a straightforward printing profile using minimal support geometry.

I recieved a final grade of 89/100.