Emergent diagnosis for Evolvable Production Systems

Abstract

A wide range of paradigms have been proposed that, taking advantage of the emerging IT technologies, increasingly support autonomous and evolvable shop-floors which are able to react to disturbances and opportunities. Among the emerging paradigmatic proposals the concept of Evolvable Production System (EPS) is emerging as it proposes a holistic approach with clear prospects for enhancing business and environmental competitiveness. EPS systems are knowledge driven and are dynamic compositions of intelligent and interacting modules. While most of the research in such distributed industrial systems is focused in the study and establishment of control structures, the problem of diagnosis has been left relatively unattended. There are however significant open challenges in the diagnosis of such modular systems including: understanding fault propagation and ensuring scalability and co-evolution. A diagnostic architecture is proposed that addresses these issues.

An agent-based interaction-oriented shop floor to support emergent diagnosis

“…It operates by adjusting the agents diagnostic state in respect to a given interaction when its direct neighbors denote a change in their states for the very same interaction. From a network perspective clusters of agents tend to adopt, after some time, a converging opinion on the origin of the fault and the propagation path [21]. There is not a global model of the system, it emerges out of module interaction, therefore the diagnostic performance is not significantly affected by the addition and removal of modules nor by their number in the system.…”

Diagnosis in Networks of Mechatronic Agents: Validation of a Fault Propagation Model and Performance Assessment

“…The system for testing hypothesis H.1 has been previously documented in [30][31][32] where some preliminary informal tests have been carried out to detail the significance of the nature of knowledge representation [30] and the supporting IT infrastructure [31,32]. The system makes use of a Hidden Markov Model (HMM) [33] that from, a set of 18 possible observations, attempts to infer each agent state.…”

Characterization of Costs and Strategies for Automation in Evolvable Production Systems

“…In recent years, the active EPS research community has extensively studied different aspects of such domain which, in turn has spread in different directions through a large amount of heterogeneous contributions: research requirements [10], design support tools and models [11,12], business processes [5,13,14], diagnostic [15,16], system deployment and description [17,18], system configuration [19], self-learning mechanisms [20], self-organization theory [21] and management of complexity [22] among the others. Most of the research efforts that shaped the EPS paradigm at the early stage have been carried out within the large collaborative European project EUPASS [23] while the more recent developments have been attained within the following European project IDEAS [24,25].…”