A study of micromanufacturing process fingerprints in micro-injection moulding for machine learning and Industry 4.0 applications

Abstract

This paper discusses micromanufacturing process quality proxies called “process fingerprints” in micro-injection moulding for establishing in-line quality assurance and machine learning models for Industry 4.0 applications. Process fingerprints that we present in this study are purely physical proxies of the product quality and need tangible rationale regarding their selection criteria such as sensitivity, cost-effectiveness, and robustness. Proposed methods and selection reasons for process fingerprints are also justified by analysing the temporally collected data with respect to the microreplication efficiency. Extracted process fingerprints were also used in a multiple linear regression scenario where they bring actionable insights for creating traceable and cost-effective supervised machine learning models in challenging micro-injection moulding environments. Multiple linear regression model demonstrated %84 accuracy in predicting the quality of the process, which is significant as far as the extreme process conditions and product features are concerned.

Rapid tooling: investigation of soft-tooled micro-injection moulding process characteristics using in-line measurements and surface metrology

“…For investigating if the S/O point is affected by the deformation occurred in the cavity size, analysis of the unfiltered pressure data near the pressure-controlled S/O points was performed. The S/O point is of great importance for filling control in micromoulding, because of the extreme injection speeds, flow rates and associated momentum of the injection piston and the unit (Whiteside et al , 2005; Gülçür and Whiteside, 2021). Because of significant gate cross-section decrease at 28%, an overshoot in injection pressure can be expected.…”

“…Each pair of the microfeatures were placed progressively away from the gate with approximately 3.5 mm distance for the evaluation of micro-replication and possible deformation on the soft tool. A single, circular recess with 600 μm depth and 800 μm diameter was also placed at the far end of the cavity where the replication would be most sensitive to the process variations and pressure drops during the filling process (Griffiths et al , 2014; Gülçür and Whiteside, 2021; Gülçür et al , 2021). The disc-shaped main cavity had a diameter and thickness of 13 and 1.25 mm, respectively.…”

“…Although these reports prove to be invaluable when characterising a soft-tooled μ-IM process and make excellent contributions to the literature, there is still a need to implement further in-line measurement methods for processes in the μ-IM regimes. The in-line monitoring aspect of conventional μ-IM is a prominently researched area because of challenging process environments and miniature products (Whiteside et al , 2004; Whiteside et al , 2005; Gülçür and Whiteside, 2021; Santoso et al , 2022; Zhang et al , 2021; Gulcur, 2019). A majority of the similarities in the process can be exploited to characterise soft-tooled processes.…”

“…The advancements and developments towards miniaturisation in AM also overlap with the trends in the highly capable and established manufacturing technique known as micro-injection moulding (μ-IM) (Bibber, 2014; Bibber, 2012; Whiteside et al , 2003; Gülçür and Whiteside, 2021). The technique has been the state-of-the-art for high-volume and economical manufacture of miniature medical devices, optoelectronics and micromechanical components (Gülçür et al , 2021; Romano et al , 2019; Baruffi, Calaon and Tosello, 2018; Zhang et al , 2021).…”

A Multimodal Deep Learning-Based Fault Detection Model for a Plastic Injection Molding Process

“…Especially in the plastic injection molding domain, there have been several machine learning-based fault detection approaches. Support vector machine (SVM) and multiple linear regression (MLR) are employed to predict the quality of the manufacturing processes [29], [30]. Ventura and Berjaga make a comparison of statistical discriminant analysis techniques, SVMs, and partial least squares [31].…”