Dynamic Multilevel Workflow Management Concept for Industrial IoT Systems

Abstract

Workflow management is implemented in manufacturing at many levels. The nature of processes varies at each level, hindering the use of a standard modeling or implementation solution. The creation of a flexible workflow management framework that overarches the heterogeneous business process levels is challenging. Still, one of the promises of the Industry 4.0 initiative is precisely this: to provide easy-to-use models and solutions that enable efficient execution of enterprise targets. By addressing this challenge, this article proposes a workflow execution model that integrates information and control flows of these levels while keeping their hierarchy. The overall model builds on the Business Process Model and Notation (BPMN) for modeling at the enterprise level and recipe modeling based on colored Petri net (CPN) at the production level. Models produced with both alternatives are implemented and executed in a framework supported by an Enterprise Service Bus (ESB). Loosely coupled, late-bound system elements are connected through the Arrowhead framework, which is built upon the Service-Oriented Architecture (SOA) concept. To prove its feasibility, this article presents the practical application of the model via an automotive production scenario. Note to Practitioners - The methodology detailed in this article can serve as a basis for experts who are dealing with industrial workflows. Reacting to the requirements of Industry 4.0, i.e., the virtualization, decentralization, modularity, real-time capability, and service orientation, this article provides a concept that can answer all the defined criteria. First, it adopts a new two-level approach to workflow management, which makes the understanding and control of workflows easier, enhancing transparency. Furthermore, it demonstrates how - even completely different - applications and modeling languages can be integrated into a Service-Oriented Architecture (SOA). The presented composition and the used tools are all tried and tested. Behind the solution described in this article, there is a genuine, working code wherewith the presented end-to-end workflow management can be achieved. Following the methodology detailed in this article, the readers can construct their workflow management composition. In order to report on the performance of the created solution, this article presents different measurement compositions that allow the investigation of the essential components separately, demonstrating the scalability and temporal parameters.