Research Update

Defect Detection in AM Process (LPBF) (2)

Powder-based additive manufacturing-a review of types of defects, generation mechanisms, detection, property evaluation and metrology [1]

Figure 1. Schematic of balling phenomenon

Taheri et al. summarized the defects found in powder-based additive manufacturing, the mechanism of their creation, and the detection and evaluation methods. The authors state that defects are typically generated by several factors (e.g., laser power, scan speed, layer thickness, scan width, powder size, and feed rate). Balling is a defect that is mainly focused on in our study, and sub-critical energy is the leading cause of this defect. In addition, the high scanning speed causes unstable melt pools due to the capillary effect, and the formation of oxide layers will change the wetting process of printing materials. Optical inspection methods are commonly used to identify defects, which are advantages of low price and easy sensor installation. In addition, optical sensors can be installed in a printing chamber due to their relatively small volume, thus opening up the possibility of on-machine defect inspection.

CenterNet-based defect detection for additive manufacturing [2]

Figure 2. Illustration of demonstration for the proposed method

Wang and Cheung introduced the CenterNet-based defect detection model to identify defect features, types, locations, and counts from microscopic surface images captured by scanning electron microscope. The proposed model has identified the 4 different types of defects (i.e., unmelted powder, porosity, collapse (hole), and crack) on the additively manufactured part made of 316L stainless steel. Open datasets are used o evaluate the performance of this proposed model. Even though this proposed model performs better than other detection models, the detection results for defect detection in additive manufacturing showed poor detection ability. Here is the research question: Does this model able to mathematically simplify the description of the anomalies in the surface images?

References:

[1] Taheri, H., Shoaib, M.R.B.M., Koester, L.W., Bigelow, T.A., Collins, P.C. and Bond, L.J., 2017. Powder-based additive manufacturing-a review of types of defects, generation mechanisms, detection, property evaluation and metrology. International Journal of Additive and Subtractive Materials Manufacturing, 1(2), pp.172-209.

[2] Wang, R. and Cheung, C.F., 2022. CenterNet-based defect detection for additive manufacturing. Expert Systems with Applications, 188, p.116000.