LAUSR

Maintenance services of geographically dispersed industrial applications, such as oil transfer systems via pipelines and wastewater treatment plants, are affected by high logistics costs and risks of permanent downtimes. The increasing availability of smart technologies and devices has led to the introduction of advanced prognostic and diagnostic systems to support maintenance activities. In this context, artificial immune systems support the development of industrial applications, where machines and equipment are capable of self-repairing, healing and learning due to their ability to learn from experience. However, the applicability of artificial immune systems has a limited set of contexts along with a low incidence of real-word implementations in the literature, and thus, additional explorative studies are necessary. This article describes a proposed hybrid system conceived by integrating a multi-agent system–based architecture with the main features of artificial immune systems and evaluates its potential applications in two different industrial settings. The flexibility of the behaviour of artificial immune systems methodologies allows for the implementation of a reliable diagnostic and prognostic system, while the choice of multi-agent system architecture enables a mix of autonomy and distributed processing that overcomes the strong limitations of a reduced training dataset.