I created a 6-axis load cell test fixture to make it easier for students to use MSOE’s wind tunnel while significantly improving on the experimental data collected (2023-24). Previous fixtures relied on spring scales or coupled load cells to obtain one or two axis force data, but sensor data from the 6-axis load cell allows for measurement of forces in and moments around X,Y and Z axes. The 6-axis load cell uses an arrangement of 6 inexpensive load cells, connected to a low-cost microcontroller, which are offset by a 3D printed Stewart platform and support a laser-cut mounting fixture.

Initially, I used calibration data to fit a linear coefficient matrix, however, students were able to create designs beyond expected ranges and into non-linear response regimes. To remedy this problem, I collected additional calibration data and used it to train a shallow neural network for inference of applied forces/moments and integrated it into the MATLAB script supplied to students for data collection. I designed and created all the models, circuits, and software. I also operated all needed machines to make all of the custom components.

Software used: Solidworks, PrusaSlicer, LightBurn, MATLAB
Languages used: C++, MATLAB