Method of performing powder bed fusion process

A method of performing powder bed fusion process, comprising:

• providing a powder bed, wherein the powder bed includes a substrate and a powder layer, the powder layer contains a plurality of powders, and the powder layer is deposited on the substrate;
• obtaining a group of information of the powder bed, wherein the group of information of the powder bed includes a powder size distribution, material properties of the powders, and shape of the powders;
• performing a powder bed simulation to obtain a packing density corresponding to different thickness of the powder layer by using the group of information of the powder bed according to a first algorithm, wherein the packing density of the powder bed is higher than a predetermined packing density value;
• obtaining a group of parameters of a laser, wherein the group of parameters of the laser includes a laser type, and values of a laser power, a scanning speed, and a laser spot size;
• performing a Ray Tracing simulation for the powder layer to obtain an absorptivity profile along a depth of the powder layer by using material properties of the powders, the group of parameters of the laser, and the thickness of the powder bed according to a second algorithm;
• performing a heat transfer simulation to obtain a peak temperature and a dimension of a melt pool by using the group of parameters of the laser, a temperature of the substrate, the thickness of the powder layer, and the absorptivity profile according to a third algorithm, wherein the dimension of the melt pool includes a length, a contact width, and a depth of the melt pool;
• constructing a first surrogate model to obtain a plurality of first processing maps for different thickness of the powder layer by inputting the peak temperature and the dimension of the melt pool in to artificial neural networks, wherein the first processing maps includes a number of points, and each point is corresponding to a combination of a scanning speed value and a laser power value;
• calculating a depth of the melt pool of each point in the first processing maps by inputting the scanning speed value and the laser power value into the first surrogate model;
• comparing the depth of the melt pool of each point with a predetermined depth value,
• when the depths of the melt pool of the points are smaller than the predetermined depth value, the points in the first processing maps are not applicable;
• when the depths of the melt pool of the points are greater than the predetermined depth value, comparing the depth of the melt pool with a laser beam radius,
• when the depths of the melt pool of the points are greater than the laser beam radius, the points in the first processing maps are not applicable;
• when the depths of the melt pool of the points are smaller than the laser beam radius, performing a parameter setting operation by setting the points in the first processing maps as a first group of parameter values;
• performing a plurality of cycles of selective laser melting operation on a workpiece, wherein each cycle of selective laser melting operation including:
• performing the parameter setting operation by a controller using the first group of parameter values;
• performing a laser melting operation on the powder layer of the workpiece;
• and
• measuring a temperature distribution on a top surface of a solidified layer by using an infrared thermal camera, wherein the solidified layer is the powder layer melted by the laser and solidified.