Overcoming the Barriers in Diagnostics and Prognostics of the Circular Industrial System by Hidden Markov Model

In this paper, a novel method has been proposed to overcome the difficulty of handling technical heterogeneity and varying operating condition, while designing Prognostics and Health Management (PHM) systems. The motivation towards this work is to facilitate better decision making on re-purposing, in the context of Circular economy. In the case of re-purposing, there are multiple second life applications of a product. It is utterly important to know the state of health in prior for sorting the product efficiently into the above-said applications. However, estimating the health state of products with complex degradation mechanism, particularly electronic devices, is difficult, due to technical heterogeneity and varying operating conditions. The method proposed, is based on the black box modeling of the system, relying on the real-time data acquired through the sensors installed to monitor the component. A data-driven stochastic method based on hidden Markov model is proposed. A region based model database has been created to provide versatility to the overall system.