Plug and Produce for Industry 4.0 using Software-defined Networking and OPC UA

Abstract

The manufacturers are in quest for flexible and agile production facilities capable of accommodating changes to product specification. The need for flexible production facilities is stemming from the desire for customized products and fluctuating market trends. Industry 4.0 impels for adaptable manufacturing plants by employing intelligent devices and advanced communication technologies. The complexity of the configuration process determines the adaptability of production facilities to accommodate changes to the production process.We propose a systematic integration process and multi-level production system using Software-defined Networking (SDN) and OPC Unified Architecture (OPC UA) to reduce the configuration complexity to a Plug and Produce level. OPC UA, as a serviceoriented middleware, provides the tool-set for semantic modeling and automatic device discovery. However, due to the multicast nature of the OPC UA discovery mechanism, the existing approaches require intelligence at the device level to select the desired device to connect to. In contrast, our proposed solution shifts the intelligence to a centralized SDN controller to route multicast traffic to facilitate device discovery. The combination of an SDN controller and OPC UA discovery enables the integration of new devices by adding more intelligence to the device discovery.

A Generic Plug & Produce System Composed of Semantic OPC UA Skills

“…Nestability of Industry 4.0 components, as defined in RAMI 4.0 [40], can be achieved by using multiple subnets, e.g., on the workcell level, to encompass other components in logical terms and to abstract away the underlying components on a higher level using skill composition as described in the following sections of this paper. An alternative to hierarchical grouping is the use of software-defined networking (SDN) as shown in [43].…”

“…In this case, a higher-level component needs to intelligently select the correct endpoint based on additional information from the task or world model, e.g., the robot's position and reachable robot's working area. Another solution could be to use Software-Defined Networking (SDN) to create different network segments and thereby shift the intelligence from the lower level to a central instance [43].…”

Learning Factories on Demand

“…It is important here that these interfaces are valid across all platforms. An example of this can be found in Madiwalar et al [26], who defines a "plug and produce" environment for hardware components via OPC UA.…”

Towards a Modular Architecture for Industrial HMIs

“…• Flexible Communication Network: This requirement is also coming from "Plug-and-Produce" architecture. It basically means that the communication network used to exchange the information between HMI client applications and modules must be flexible to accommodate the restructuring of modular HMI and machines without the need of reconfiguring the network devices manually [15]. • Independent Deployment: Independent deployment is a key feature of a micro-service architecture, it reduces the scope of deployment of an HMI module to a specific automation component.…”

“…• Element Registration: The screens in the HMI client applications are created by integrating and organizing operating screen elements. The element registration is similar to the "skill as a service" concept presented in [15], where a device offers skills as services to be used by other devices. In our case, the skills are operating screen elements that are offered by an HMI module as services to be used by other HMI client applications.…”