Search

Galling wear in sheet metal stamping processes can degrade the product quality and adversely affect mass production. Studies have shown that acoustic emission (AE) sensors can be used to measure galling. In the literature, attempts have been made to correlate the AE features and galling wear in the sheet metal stamping process. However, there is very little attempt made to implement machine learning (ML) techniques to detect AE features that can classify non-galling and galling wear as well as provide additional wear-state information in the form of strong visualisations. In the first part of the paper, time domain and frequency domain analysis is used to determine the AE features that can be used for unsupervised classification. Due to galling wear progression on the stamping tools, the behaviour of AE waveform changes from stationary to a non-stationary state. The initial change in AE waveform behaviour due to galling wear initiation is very difficult to observe due to the ratio of change against the large data size of the waveform. Therefore, a time-frequency technique "Hilbert Huang transform" is applied to the AE waveform as that is sensitive to change of wear state and is used for the classification of "non-galling" and the "transition of galling." Also, the unsupervised learning algorithm fuzzy clustering is used as comparison against the supervised learning techniques. Despite not knowing a priori the wear-state labels, fuzzy clustering is able to define three relatively accurate distinct classes: "unworn", "transition to galling" and "severe galling". In the second part of the paper, the AE features are used as an input to the supervised ML algorithms to classify AE features related to non-galling and galling wear. An accuracy of 96% was observed for the prediction of non-galling and galling wear using classification, regression tree (CART) and neural network techniques. In the last part, a reduced short time Fourier transform of top 10 absolute maximum component AE feature sets that correlates to wear measurement data "profile depth" is used to train and test supervised neural network and CART algorithms. The algorithms predicted the profile depth of 530 unseen parts (530 unseen cases), which did not have any associated labelled depth data. This shows the power of using ML techniques that can use a small data training set to provide additional predicted wear state on a much larger data set. Furthermore, the ML techniques presented in this paper can be used further to develop a real-time measurement system to detect the transition of galling wear from measured AE features. [ABSTRACT FROM AUTHOR]

In the sheet metal stamping process, during sliding contact between the tool and sheet, it is expected that severe events such as tool wear or fracture on the sheet generate acoustic emission (AE) burst waveforms. Attempts have been made in the literature to correlate the AE burst waveform with the wear mechanisms. However, there is a need for additional studies to understand the frequency characteristics of the AE burst waveform due to the severity and progression of the galling wear. This paper will determine the AE frequency characteristics that can be used to monitor galling wear, independent of the experimental process examined. The AE burst waveforms generated during the stamping and scratch tests are analysed in this paper to understand the change in the AE frequency characteristics with the galling severity. These AE burst waveforms were investigated using the Hilbert Huang Transform (HHT) time-frequency technique, band power, and mean-frequency. Subsequently, these AE frequency features are correlated with the wear behaviour observed via high-resolution profilometer images of the stamped parts and scratch surfaces. Initially, the HHT technique is applied to the AE burst waveform to understand the influence of wear severity in the power distribution over the wide AE frequency range. Later, the AE bandpower feature is used to quantitatively analyse the power in each frequency interval during the unworn and worn tool conditions. Finally, the mean-frequency of AE signal is identified to be able to determine the onset of galling wear. The new knowledge defined in this paper is the AE frequency features and wear measurement feature that can be used to indicate the onset of galling wear, irrespective of the processes examined. [ABSTRACT FROM AUTHOR]

In the stamping process, tools are prone to an adhesive wear mode called galling. This galling wear mode on the stamping tool results in an abrasive wear modes like ploughing and cutting on the workpiece. To study the adhesive and abrasive wear modes relevant in sheet metal stamping processes, scratch tests were performed under controlled conditions where galling, ploughing and cutting can be observed. Two sets of scratch tests were performed to study the galling behaviour using force and acoustic emission sensors. In the first test set, scratch tests were performed at a different depth of penetration to segregate the non-galling and galling conditions. In the second test set, scratch tests were performed at a different sliding distances to understand the influence of galling on the abrasive wear modes. To study the galling behaviour, acoustic emission and force features from both of the test sets were correlated with profilometry wear measurement features like profile depth, wear index, attack angle and volume measurement of galling. From the quantitative measurement of galling on the indenter, a minimal lump was observed on the indenter when the cutting at the edges of scratch was observed. A much larger lump was observed on the indenter for conditions when fracture was observed on the workpiece at the centre of the scratch. The acoustic emission burst signal and unstable force behaviour was mainly observed when the signficant lump was observed on the indenter. The methodology adopted to investigate galling wear in this study lays the strong foundation to develop real-time monitoring systems to observe the transition from non-galling to galling conditions. • AE peak feature can be used to study initiation of galling. • Tangential force can be used to indicate prior indication of galling. • Indenter attack angle plays an important role on the abrasive wear transition. • Profile depth is quantitative wear measurement to study abrasive wear transition. [ABSTRACT FROM AUTHOR]