A Hybrid Dimensional Error Compensator for FDM ۳D printers Using GMDH Neural Network and Taguchi Orthogonal Arrays

The increasing growth of computer-aided design has led to the popularity of additive manufacturing methods and the expansion of their applications.Therefore, eliminating its weaknesses has become of special importance. One of the most common additive manufacturing methods is FDM. This method, along with many advantages, has serious weaknesses, the most obvious of which is the lack of dimensional accuracy.In this study, a compensation model for the dimensional error is represented regardless of its source. A multi-input–single-output (MISO) data is prepared by designing of experiments using the Taguchi orthogonal arrays and obtaining the results from ۳D printed samples. Next, a GMDH neural network is applied, which uses a simple nonlinear regression formula in each neuron but can create very complex neuron combinations. So, it is possible to analyze small or even noisy data. The case study shows a decrease in the RSME for the Nominal CAD Model displaying the compensator's efficiency.