Process and product analysis in multidisciplinary, heterogeneous engineering environments

Abstract

In heterogeneous engineering environments, such as power plant engineering, different tools and systems are used independently by engineers from various disciplines, like mechanical, electrical or software engineering. Collaboration between engineers, tools and system domains can be facilitated by designing a system of systems that implements all required use-cases. Current approaches have the disadvantage that they require stakeholders to perform critical human tasks outside of the system and are therefore time consuming, prone to errors and offer room for improvement. Another limitation of current solutions is that processes store events that do not allow a connection to the associated engineering data. The lack of such a modeled connection means that it is not possible to create queries spanning both process and product data to measure metrics and key performance indicators. As a result, the stakeholders are not able to analyze where problems in their engineering processes occur and which process factors contributed to them. Therefore, such a system also has to provide monitoring capabilities that deliver ad hoc insights into ongoing and past processes and allow to derive meaningful data based on defined metrics for project progress and collaboration analysis. This is particularly important for project and business process managers who may not only be interested in factual process data, but also in the collaborative behavior of the involved stakeholders across different domains. With such a monitoring system, they obtain more information for decision making, prediction of future trends about development activities, as well as indicators of product quality. This thesis addresses these problems theoretically and practically and offers a solution that allows improved insights into multidisciplinary engineering environments. The approach of this thesis will be the modeling and execution of collaboration related processes to collect event data from, which can be used to derive characteristics of system behavior. This is achieved by not only storing historical progress of common concept instances or signals (which represent a model of the created product) in the engineering database, but also by linking process event data with this historical progress by adding references. The implementation has been evaluated both against test scenarios with real-world test data provided by an industry partner in a case study. The results of the evaluation showed that the benefits of the solution are improved collaboration processes, identification of standard system behavior and detection of anomalies, as well as reducing the overall time effort of the involved stakeholders.