Contatti
Telefono: 02.2399.8534
Skype: matteo.pacher@polimi.it
E-mail: matteo.pacher(at)polimi.it

Manufacturing and Production Systems

Relatori: Prof.ssa Barbara Previtali, Prof.ssa Mara Tanelli

Tutor: Prof. Roberto Corradi

Università di Provenienza: Università degli Studi di Trento - Ingegneria Meccatronica

Titolo della Tesi: Laser Cutting Monitoring and Control for Real-Time Dross Attachment Regulation

Laser Cutting Monitoring and Control for Real-Time Dross Attachment Regulation

Introduction

For many applications and particularly for cutting metal sheets and thin tubes, laser cutting has become the reference technology thanks to its flexibility and the gain of productivity if compared with other competitive technologies such as abrasive water jet cutting and electrical discharge machining. The quality of the laser cutting process is traditionally determined considering different quantities, namely, dross attachment, kerf width, surface roughness, heat affected zone and presence of burns on the cut edge [1]. However, these features do not have the same relevance, and it has been found that dross attachment is the most influential output parameter [2], [3]. Accordingly, this study focuses on dross attachment only and proposes a continuous and quantitative measurement variable that is to be estimated via monitored quantities.
The amount of dross attachment produced during the process is eventually regulated to a desired level thanks to the development of a closed-loop control system.

Objectives

The main objectives of the thesis according to the scheme of the process include:

  • the definition of a quantitative and continuous output parameter, y, related to the dross attachment formation;
  • selection of the source of information to be used for the online estimation of the system output;
  • investigation of the source of information to be used for estimating the output, and development of the estimator yielding an accurate estimate of the defined process output;
  • identification of the system dynamics and design of the closed-loop controller structure.

Results

A custom monitoring architecture was designed to collect the process emission through a coaxial industrial camera. This solution permitted to enhance the time resolution of two order of magnitudes with respect to relevant scientific and industrial literature.
The presented method for dross attachment estimation relies on process emission images. Image information is processed at different levels to produce significant geometrical features of the laser irradiated zone. These features are then mapped via an artificial neural network to produce an online estimation of dross attachment.
The resulting dynamical system has been identified and a controller is finally built up to either minimize or set dross attachment to a desired level. The performance of the proposed control system has been experimentally validated. As a final remark, the costs of the implemented solution are limited possibly favoring its introduction in the industrial field.

Conclusions

The presented results definitely open the possibility of designing a new generation of intelligent machines that can adapt the processing conditions to the specific manufacturing needs. An ad hoc trade-off between quality and productivity can be found for the specific production job and quality assurance can be provided. These features well integrate with the concepts of Industry 4.0 and Zero Defect Manufacturing and constitutes a considerable contribution for shaping the Industry of the future.

References

[1] C.L. Caristan, Laser Cutting Guide for Manufacturing, Dearborn, Michigan: Society of Manufacturing Engineers, 2004.
[2] M. Pacher, L. Monguzzi, L. Bortolotti, M. Sbetti, B. Previtali, Quantitative identification oflaser cutting quality relying on visual information, in LiM 2017 Proceedings, Munich, pp. 1–11, 2017.
[3] C. Goppold, T. Pinder, P. Herwig, Transient beam oscillation with a highlydynamic scanner for laser beam fusioncutting, Advanced Optical Technologies, vol.5, no. 1, pp. 61–70, 2016.

Acknowledgements

The presented PhD project has been founded by the BLM GROUP and has been realized in strict cooperation with Adige SPA. The authors are particularly grateful to the colleagues of Adige SPA for the professional support received during the study and the precious advices about the laser cutting technology.