Back to overview

Lecture

WEB Process analysis of thermoplastic-metal composite structures in 3D-Hybrid technology - A combined approach to quality assurance for robust manufacturing processes

Tuesday (28.04.2020)
14:20 - 14:40 Room 2

Thermoplastic composites with metallic structures in a combined thermoforming and injection moulding process offers new possibilities for lightweight structures. Intrinsic joining in a one shot manufacturing process opens up new possibilities using purposefully the specific properties of the different materials in an integral lightweight design. The potential of the 3D Hybrid technology is demonstrated in several preliminary projects. One result is a significant reduction in mass with the same performance of the structural components.

For an industrial application, an intrinsic process brings a lot of challenges regarding the quality assurance of the parts. One important aspect is that the combination of several process steps in a one-shot process increases the complexity through the multitude of adjustable and interacting process parameters. The correlations and interactions of the process and disturbance parameters on the component properties are partly unknown for complex hybrid structures. This leads to time consuming inter-operation loops in running-in processes and to incomplete quality assurance of the key parameters for new products.

 

In this study, a generic structure made of thermoplastic fibre composites combined with injection moulding ribs is investigated. In a first step, a pre-analysis describes the robustness of the entire process chain related to the target parameters. To identify the process, disturbance parameters and their correlations a combination of standard methods from data mining as well as from process and product optimization were applied. For an efficient data acquisition and storage, a separate procedure was developed to describe all relevant attributes of the process chain in a flowchart. Moreover, in the pre-processing of the data the reduction of multidimensional process parameters to one-dimensional characteristic values played a significant role. A parallel analysis using classical methods from the statistical design of experiments and approaches to machine learning provides statistically significant effects of the process parameters on the mechanical properties of the tested hybrid structures. The procedure presented here leads to the identification of the key parameters and supports the design of process windows. With this knowledge, new quality assurance steps can be implemented.

Speaker:
Additional Authors:
  • Moritz von Unold
    Technische Universität Dresden
  • Dr. Michael Krahl
    Technische Universität Dresden
  • Prof. Dr. Mike Gude
    Technische Universität Dresden

Dateien

Category Short file description File description File Size
Paper Version 2 2 MB Download