A predicting system (5) for automatic prediction of porosity appearance generated during Laser Powder Bed Fusion (L-PBF), performed by an additive manufacturing (AM) system (1) from at least one material (2), comprising modules being configured for training a neural network:
- at least one module configured to read a kinematic laser file describing laser exposure over a layer of a part slice, the kinematic laser file comprising at least an X coordinate and an Y coordinate of each of a plurality of points in a laser path, the plurality of points being situated on a 2D mesh representing the layer, a point of said plurality of points being associated with an exposure time representing a time at which a laser passes through the point and an exposure type indicating whether a point of said plurality of points belongs to a part contour or to an inner laser path, and to convert the kinematic laser file into a dataset comprising an array of pixels discretizing the layer, a pixel of said array being associated with local coordinates and with an exposure time representing a time at which the laser passes through in the local coordinates of the pixel;
- a generating module (51) configured to generate labels of pore representing presence or absence of pores in each pixel of the dataset using a porosity simulator (50), wherein the porosity simulator is image-based and the generated label of pores for a pixel of the dataset is in the form of 1 or 0, representing respectively presence or absence of pores with respect to the pixel;
- a pre-training module (52) comprising:
- a first pre-training sub-module (521) configured to create a first data-set by, for each pixel, processing input features processing pore label features based on the labels of pores generated by the porosity simulator, said input features and pore label features constituting feature vectors of the second dataset associated with said pixel;
- and
- a second pre-training sub-module (522) configured to use the feature vectors of the first data set to generate a pre-trained model (200);
- and
- a training module (53) comprising:
- a first training sub-module (531) configured to create a second data-set by, for each pixel, processing input features and pore label features based on tomography of manufactured coupons (20) said input features and pore label features constituting feature vectors of the second dataset associated with said pixel;
- and
- a second training sub-module (532) configured to use the feature vectors of the second data set to train the pre-trained model (200) to generate a trained model (300);
- one or more preprocessing modules configured to obtain input features comprising Correlated Density Spectrum (CDS), the input features obtained by:
- generating a distance matrix based on kinematic laser data via kinematic processing (SA), by calculating Euclidean distances of each pixel of the dataset belonging to an inner laser path, referred to as inner pixel with all the pixels of the dataset through which a laser passes;
- calculating (SB) a distance correlation of pixels comprising covariance matrix based on the distance matrix;
- generating (SC) CDS based on the covariance matrix;
- one or more additional modules configured to:
- predict porosity appearance generated during Laser Powder Bed Fusion (L-PBF), performed by an additive manufacturing system (1) from at least one material (2), the predicting module configured to apply (S5) the trained model on new data continuously provided by the AM system (1) in order to generate a predicted porosity appearance;
- and
- post-process the predicted porosity appearance so as to generate a quality report.