Analysis of coding methods in quantum neural networks for failure diagnosis in rolling elements

In several production systems, the monitoring of equipment degradation is important, where the identification of the health status of the machinery provides useful information for decision making. Among the various components of the machines, the rolling elements are key parts for the proper functioning of the system, requiring a correct maintenance policy. In this sense, for an accurate identification of the health status of these components, machine learning methods are used. Specifically, quantum computing models can be applied through quantum machine learning (QML). Therefore, this work seeks to investigate how different quantum coding methods, in QML models, perform in the diagnosis of failure in rolling elements. Methods such as angle coding, amplitude coding, ZZFeaturemap and Pauli Z, X and Y feature maps will be analyzed. For validation, two bearing databases will be used.