OpenServ4P | Open, intelligent services for production

Sensor-based electronic systems play a key role for intelligent planning and resource efficiency applications in the context of Industry 4.0. Machines and plants in production with a high proportion of networked sensor systems are called Cyber-Physical Systems (CPS). Their use in production systems leads to Cyber-Physical Production Systems (CPPS). CPPS are application-specific system solutions whose processes are monitored, controlled and coordinated by integrated computing and communication units. The ability of CPS for ad-hoc networking and the availability of sensor information enable the development of novel data-driven, intelligent applications, the Smart Services. This can increase flexibility and productivity in production.

With regard to the configuration and operation of CPPS, conventional solution approaches reach their limits. Today's solutions do not sufficiently cover the vertical integration in the automation pyramid and the available transmission speeds for complex services are not sufficient. As a result, potentials for increasing efficiency and effectiveness remain unused.

The overall goal of the research project OpenServ4P is the development of a platform and intelligent services in the environment of CPPS. The basic idea of this approach is to securely connect information of a component as well as a machine to a platform and to use the information to identify new knowledge and to optimize processes. In addition to the desire for the most direct possible data access down to the sensor level, higher-level aspects such as plant safety, data security and know-how protection must also be taken into account.

• Software and hardware architecture: Building a platform for a high-performance and robust information transfer of smart services through information modeling, semantic data integration and algorithms for (pre-)processing the data volume depending on the resources in the applications.
• Predictive maintenance and intelligent quality inspection: Development of services for the predictive maintenance of production facilities and for intelligent quality inspection.
• Adaptive Risk Management: Development of a service for adaptive risk management to assess the risk situation from sensor-induced events of the CPS in CPPS and to enable targeted reactions.
• Event-based production planning and control: Development of a service for event-based production planning and control, which dynamically schedules machines and products based on a plug and play principle.
• Business models: Design of business models for Smart Services, which provide the conceptual framework for technological innovations and new potential benefits in value-added structures.

At the Fraunhofer IGCV a demonstrator for the service in the context of intelligent quality inspection was developed. The use case originates from the production of a research partner and is used for quality testing of injection molded products. The injection molded products are inspected by manual visual inspection in multi-stage controls. The intelligent service is based on an artificial intelligence approach. With the help of a trained Convolutional Neural Network (CNN), various defects such as scratches, cracks, streaks and black dots are detected and thus good components are identified. The demonstrator comprises a robot cell, an adaptive camera, an intelligent computing unit and various positioning aids. Advantages of the cloud-based, intelligent service compared to manual visual inspection are the constant inspection results as well as the increased inspection accuracy, since the evaluation is not performed by different inspectors. In addition, the approach enables the detection of quality defects that are difficult or impossible for the human eye to detect.

The project OpenServ4P was successfully completed in spring 2019 after a three-year project duration. The developed services were successfully prototyped at the research partner BSH Hausgeräte GmbH. Furthermore, the results and findings of the project are used and further developed in current research projects of the Fraunhofer IGCV (e.g. REIF, Mittelstand 4.0-Kompetenzzentrum Augsburg).

Further information:

Schreiber et al. 2019: Automated quality assurance as an intelligent cloud service using machine learning (DOI: 10.1016/j.procir.2020.01.034).