Application of soft computing techniques in the optimization of 3D-printed piezoresistive sensors

Fused Deposition Modeling (FDM) stands out as one of the most accessible additive manufacturing methods, offering a wide variety of compatible materials and adaptable structural configurations for sensor fabrication. Printing parameters are critical especially to control anisotropy and printing qualities, which significantly influences the electromechanical behavior of sensors. Therefore, it is essential to identify the optimal set of parameters to achieve desired properties. This study employs a hybrid soft computing approach, combining artificial neural networks and genetic algorithms, to model and optimize the gauge factor of the 3D-printed piezoresistive sensors fabricated using conductive thermoplastic polyurethane. Experiments were conducted using response surface methodology, and key control variables, such as layer height, printing speed, shell count, infill angle, and overlap percentage, were systematically varied. By using soft computing techniques, it was revealed that a specific set of printing parameters, i.e., a layer height of 0.4 mm, a printing speed of 40 mm/s, five shells, an infill angle of 90 o , and overlap of 15 %, resulted in sensors with a maximum gauge factor of 12.5. The potential application of the optimized piezoresistive sensor for monitoring shoulder loads in educational backpacks is also highlighted. • Developed an ANN model to predict the gauge factor of 3D-printed piezoresistive sensors based on a polymer nanocomposite. • Optimized 3D printing parameters to maximize the gauge factor of piezoresistive sensors via genetic algorithm. • Successful application demonstration of the sensor for detecting shoulder load in various walking states.