| Abstract Scope |
Cooperative 3D printing (C3DP) is a primitive form of swarm manufacturing (SM) that utilizes multiple 3D printing robots to overcome the scalability issues present in traditional FDM 3D printing. Our prior work has established various aspects of C3DP’s fundamental operational framework, but full system autonomy, reliability, and adaptability cannot be achieved without advanced process monitoring and prediction frameworks. To this end, we present a methodology designed for C3DP that utilizes computer vision to assess part quality in-situ and a machine learning model to identify trends and make predictions based on collected datasets. To collect data, a full-factorial DoE was conducted by sweeping parameter combinations of bed temperature, nozzle temperature, print speed, and extrusion multiplier on standardized test coupons, while quantifying warping and top-down quality as response variables. Parameter interactions and physics-informed features were also incorporated, and the final model demonstrated advanced & improved prediction capability during empirical validation studies. |