Smart Resilience Services for Industrial Production

Sabine Janzen; Nurten Öksüz-Köster; Jan Sporkmann; Martin Schlappa; Jens Gerhard; Lucia Ortjohann; Philipp Becker

In: 22. VDI-Kongress AUTOMATION 2021 - Navigating towards resilient production. VDI Automatisierungskongress (AUTOMATION-2021), June 29-30, Virtual, Germany, VDI, 2021.


Nowadays in our global and interconnected economy, disruptions in industrial production systems are one of the leading business risks. In general, disruptions and their implications on the production system can be manifold: damaged production facilities resulting from natural catastrophes, lockdown of economies during COVID19 pandemic, trade restrictions due to political embargos and many more. Hence, managing industrial production systems should be more and more centered around questions of resilience. Companies capable to permanently adapt to internal and external changes and disruptions are denoted as “resilient”. Tracking the overall performance of the production system, anticipating potential disruptive events and deploying effective measures in advance are among the most crucial factors. Accompanied by an increasing complexity in production (e.g., due to Industrie 4.0), technical support in form of Smart Resilience Services (SRS) is needed for successful resilience management. Objective of the SPAICER project is the development of a data-driven ecosystem based on lifelong, collaborative and low-threshold SRS using leading AI technologies and Industry 4.0 standards. Building on that, SPAICER aims to anticipate disruptions and optimally adapt production planning accordingly at any time. Therefore, generic and specific AI modules, so-called building blocks, are aggregated individually driven by requirements to form SRS. The modules can be divided into three categories: Perceive (sensors), Think (logic), and Communicate (actuators). Generic AI building blocks provide basic functionalities for SRS, such as the connection of internal or external data sources, pre-processing of sensor data, calculation of similarity measures, sovereign and secured data exchange etc. Specific AI building blocks correspond to the needs of the respective domain and satisfy special requirements of an SRS version for certain production sectors, e.g., (disruption) pattern recognition, planning of optimized action sequences, natural language processing for representation of recommendations (NLP), automated machine learning (AutoML), (semantic) representation of data, assets, services, knowledge. Adopting the concept of smart services (cf. Smart Service Welten), SRS are individually configured packages of products and services with clear value proposition that need to be specified by applying well-founded methods. But, due to non-linear, network-like character of the intended data-driven ecosystem with concurrent processes as well as multiple entry and exit points, the application of well-known linear approaches from requirements and service engineering is not appropriate. In this contribution, we present a method for specifying SRS relevant for industrial production for establishing a resilience business continuity. Therefore, we combine approaches from Design Science (e.g., Design Thinking), Requirements Engineering (e.g., Service Use Cases) and Service Engineering (e.g., Situation-Service-Fit). The application of the method will be exemplified by the specification of SRS within the SPAICER project giving an overview of as-is and to-be situations in resilience management in industrial production. Insights are substantiated by results of an empirical quantitative study with representatives of diverse production companies evaluating specified SRS regarding their relevance, situation-service-fit, perceived usefulness, intention to use and intention to pay for the service.


Deutsches Forschungszentrum für Künstliche Intelligenz
German Research Center for Artificial Intelligence