Utilization of acoustic signals with generative Gaussian and autoencoder modeling for condition-based maintenance of injection moulds

Rønsch, Georg Ørnskov; López-Espejo, Iván; Michelsanti, Daniel; Xie, Y.; Popovski, Petar

doi:10.1080/0951192x.2022.2128218

Cited by 1 publication

(1 citation statement)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Other mold monitoring sensors previously suggested for predictive maintenance include strain gauge sensors [ 4 ], 3D accelerometers [ 8 ], and custom pressure sensors [ 8 ]. Furthermore, less conventional approaches have been studied, and show potential for being used for this purpose in the future, as is the case of acoustic signals collected from microphones installed near the mold and machine [ 9 ]. Data from these sensors and/or other sensors installed in the mold and machine, as well as contextual data can be collected, processed, and used to develop predictive maintenance algorithms, as seen in the work of Nunes et al [ 10 ], through the application of generalized fault trees and abnormality detection.…”

Section: Introductionmentioning

confidence: 99%

Development of an Open-Source Injection Mold Monitoring System

Gomes

Cadete

Ferreira

et al. 2023

Sensors

View full text Add to dashboard Cite

In the highly competitive injection molding industry, the ability to effectively collect information from various sensors installed in molds and machines is of the utmost relevance, enabling the development of data-based Industry 4.0 algorithms. In this work, an alternative to commercially available monitoring systems used in the industry was developed and tested in the scope of the TOOLING 4G project. The novelty of this system is its affordability, simplicity, real-time data acquisition and display in an intuitive Graphical User Interface (GUI), while being open-source firmware and software-based. These characteristics, and their combinations have been present in previous works, but, to the authors’ knowledge, not all of them simultaneously. The system used an Arduino microcontroller-based data acquisition module that can be connected to any computer via a USB port. Software was developed, including a GUI, prepared to receive data from both the Arduino module and a second module. In the current state of development, data corresponding to a maximum of six sensors can be visualized, at a rate of 10 Hz, and recorded for later usage. These capabilities were verified under real-world conditions for monitoring an injection mold with the objective of creating the basis of a platform to deploy predictive maintenance. Mold temperature, cavity pressure, 3-axis acceleration, and extraction force data showed the system can successfully monitor the mold and allowed the clear distinction between normal and abnormal operating patterns.

show abstract