A Digital Twin for Automated Root-Cause Search of Production Alarms Based on KPIs Aggregated from IoT

Abstract

A dashboard application is proposed and developed to act as a Digital Twin that would indicate the Measured Value to be held accountable for any future failures. The current study describes a method for the exploitation of historical data that are related to production performance and aggregated from IoT, to eliciting the future behavior of the production, while indicating the measured values that are responsible for negative production performance, without training. The dashboard is implemented in the Java programming language, while information is stored into a Database that is aggregated by an Online Analytical Processing (OLAP) server. This achieves easy Key Performance Indicators (KPIs) visualization through the dashboard. Finally, indicative cases of a simulated transfer line are presented and numerical examples are given for validation and demonstration purposes. The need for human intervention is pointed out.

A Framework for CO2 Emission Reduction in Manufacturing Industries: A Steel Industry Case

“…These can all be used under the prism of interconnected subsystems implied above and the cascaded Key Performance Indicators (KPIs) [24]. Furthermore, the use of local changes renders the optimization feasible and implementable; the theoretical background behind this could be considered to be the concept of the optimal substructure [25,26] and the application of Bellman's principle of optimality [27]. Finally, it is worth mentioning that the Pareto-optimality is applied at its best in similar cases [28] as the feasibility of a solution can be used to justify the integration of the "invasiveness" of an alternative scenario.…”

Implementation of IoT Platform’s Dashboards for the Visualisation of Dynamic KPIs: Insights from a Case Study

“…IoT platforms also enable to filter massive amounts of data and to parse out information and KPIs according to the different interests of the final users [6], [15]. Moreover, KPIs can be used to investigate the root causes of a problem and eliciting the future behaviour of the production thanks to the adoption of predictive algorithms [16]. An effective way of monitoring and evaluating performance is the selection of appropriate KPIs that strongly depends on the company's strategic intentions and its competitive environment [17], [18].…”

“…This family of indicators provide a synthetic measure of the expected time between two successive failures of an item, distinguishing between the expected time spent for repairing it and its expected functioning [19]. Although automated process monitoring tools are accessible and already established performance measurements are available, there is a lack of proper processes to guide the transformation of raw data into applied knowledge [16], [22] and a lack of management tools to guide the process of capturing KPIs [23].…”

A Framework for an Indoor Safety Management System Based on Digital Twin

“…DT can be realised through various technologies, such as the IoT [ 14 ], data/control models [ 15 ], and machine learning [ 16 ]. The role of the DT is to provide feedback for real-world physical systems to improve the performance of the systems.…”