Expert System for the Machine Learning Pipeline in Manufacturing

IFAC-PapersOnLine, vol. 54, pp. 128–133

Abstract

Many success stories already exist with regard to the implementation of Machine Learning (ML) and Artificial Intelligence (AI) in manufacturing. However, companies with traditional focus on production technologies face challenges in conducting AI-projects successfully and lack knowledge of which obstacles may occur and how to decide in the implementation phase. In this paper, we develop an approach that focuses on the methodological necessary steps for the successful application of ML and AI in manufacturing. Optimization potentials and decisions to be made are outlined in every step. A main focus is put on optimizing hyperparameters of ML-models as one promising approach for improving overall ML-model performance. An expert system is presented that enables the selection of suitable hyperparameter optimization techniques. The concept is validated based on manufacturing of compressor components of a turbofan engine.

Authors 3

1. Maik Frye Aachen Fraunhofer Institute for Production Technology IPT

   Fraunhofer Institute for Production Technology IPT

   Affiliation as printed

   Fraunhofer Institute for Production Technology IPT, Steinbachstraße 17, Aachen 52074, Germany

2. Jonathan Krauß Aachen Fraunhofer Institute for Production Technology IPT

   Fraunhofer Institute for Production Technology IPT

   Affiliation as printed

   Fraunhofer Institute for Production Technology IPT, Steinbachstraße 17, Aachen 52074, Germany

3. Robert Heinrich Schmitt Aachen Fraunhofer Institute for Production Technology IPT Laboratory for Machine Tools WZL

   RWTH Aachen University · Fraunhofer Institute for Production Technology IPT

   Affiliation as printed

   Fraunhofer Institute for Production Technology IPT, Steinbachstraße 17, Aachen 52074, Germany

   Laboratory for Machine Tools WZL RWTH Aachen University, Cluster Production Engineering 3A 540, Aachen 52074, Germany