Digital Twin Modeling of Hand Assembled Random Wire Harnesses

Wire harnesses serve as fundamental components of modern industrial equipment. Therefore, development of the digital counterpart of a random wire harness is of great interest. In this article, a digital twin (DT) modeling method is proposed for random wire harnesses with wires of distinct colors and natural bending, with the aim to accurately predict their electromagnetic (EM) behavior. To this end, first, the basic DT framework and mathematical constraints for modeling wires inside a bundle are introduced. Within such a framework, the initial representation of the wires inside a random bundle, in terms of vertical and horizontal coordinates, is extracted by an image recognition workflow based on color photos in two perspectives, i.e., side view and top view. Then, the obtained discrete coordinates are combined with an ad hoc algorithm to generate a DT of the involved wires, reproducing the physical characteristics of the original wires. Model effectiveness is verified by visual comparison of the structure of the real bundle under analysis with the one reconstructed by the proposed method and rendered by a 3-D EM solver, and by comparing the measured scattering parameters in a broadband frequency interval of the bundle with the predicted ones.