Encoding Stability into Laser Powder Bed Fusion Monitoring Using Temporal Features and Pore Density Modelling

Abstract

In laser powder bed fusion (LPBF), melt pool instability can lead to the development of pores in printed parts, reducing the part’s structural strength. While camera-based monitoring systems have been introduced to improve melt pool stability, these systems only measure melt pool stability in limited, indirect ways. We propose that melt pool stability can be improved by explicitly encoding stability into LPBF monitoring systems through the use of temporal features and pore density modelling. We introduce the temporal features, in the form of temporal variances of common LPBF monitoring features (e.g., melt pool area, intensity), to explicitly quantify printing stability. Furthermore, we introduce a neural network model trained to link these video features directly to pore densities estimated from the CT scans of previously printed parts. This model aims to reduce the number of online printer interventions to only those that are required to avoid porosity. These contributions are then implemented in a full LPBF monitoring system and tested on prints using 316L stainless steel. Results showed that our explicit stability quantification improved the correlation between our predicted pore densities and true pore densities by up to 42%.

Combining high-throughput imaging in visible and SWIR wavelengths for in-situ porosity prediction in laser powder bed fusion

“…18 Instead we utilize a compact model which was previously demonstrated to make reliable predictions. 13 The model is a 2 layer perceptron consisting of 2 fully connected dense layers with 164 neurons in the hidden layer, a sigmoid activation function in the hidden layer, and a rectified linear unit (ReLu) activation function in the output. The model is implemented in python language using Keras neural network library.…”

“…The unprocessed images are cropped to a 40x40 pixel region centered around the melt pool's center. 13 Cropping and centering aim to eliminate extraneous data and speed up processing.…”

“…One of the most promising approaches is to use high-speed cameras that capture the thermal radiation emitted by the molten pool and the surrounding powder bed. 13 These cameras can operate in different spectral ranges, such as visible or short-wave infrared (SWIR), and provide rich information about the temperature, shape, size, and dynamics of the molten pool, as well as the spatter, porosity, and surface roughness of the solidified part. 14 However, the raw images from the cameras are often noisy, distorted, and complex.…”

A GPU optimization workflow for real-time execution of ultra-high frame rate computer vision applications

“…To evaluate our optimization workflow, we start from an existing LPBF video analysis algorithm [7]. An overview of the algorithms and the hardware-software platform utilized for this evaluation can be found in sections 3.1 and 3.2, respectively.…”

“…One such application is laser powder bed fusion (LPBF), a widely used metal additive manufacturing (AM) technique that produces high quality parts for various industries [3,25]. However, it suffers from printing defects, mainly keyhole and lack-of-fusion pores, resulting in sub-standard parts and increased scrap rates [7].…”