Enhanced Planning Of Production Plants: A Case-Based Reasoning Driven Approach

Abstract

Advanced industrial developments lead to increasingly customized products and shortened product life cycles. In the context of production systems, this necessitates faster engineering of suitable production resources and layouts to cope with the increasing product variety. Current engineering processes rely mainly on adapting already realized solutions and leveraging past experiences to address new project challenges. However, the knowledge about efficient planning processes is often tied to individual employees. These experiences cannot be utilized consistently, particularly in employee departure or absence due to illness. To counteract this problem, companies attempt to digitize and store knowledge in various ways. Nevertheless, the company-wide and person-independent retrieval of crucial information is still difficult or impossible. Influencing factors are, among others, non-standardized information models and forms of description. In response to these challenges, this paper introduces an approach for the standardized modeling and crosscompany provision of experiences in production plant planning. Based on the paradigms of case-based reasoning and vendor-neutral data modeling using AutomationML, a system for selecting production resources and planning related layouts is demonstrated. By determining the similarity of new product structures, whose production facilities have yet to be engineered, with products whose production facilities are already realized, suitable existing solutions regarding production resources and their placement can be submitted explicitly to a planning expert. The approach is exemplified by a scenario of engineering an assembly system for electrolysis stacks. For this purpose, the similarity determination is performed using the Hamming similarity. Thus, it can be shown that case-based reasoning, which is already successfully used in other domains, has a significant potential to accelerate the subprocesses of production plant planning.