Online real-time dynamic slicing method for laser additive manufacturing

Abstract

The invention discloses an online real-time dynamic slicing method for laser additive manufacturing (LAM), relates to the technical field of laser additive manufacturing, and relates to the technical field of laser additive manufacturing. According to a preset processing path filling mode, layer thickness and other parameters, slicing software in the server starts to carry out partial slicing operation on the workpiece model, and a slicing and scanning path planning result is obtained; the slicing result of the part of the workpiece model is sent to an industrial personal computer of LAM equipment in an Ethernet communication mode to enable the industrial personal computer to start working; and in the LAM process, the working mode of the LAM equipment is that forming is executed while the slicing result is received, and the slicing software eliminates the formed part through Boolean operation of the three-dimensional model. According to the method, slicing while forming is supported, adjustment of important machining parameters such as the slice layer thickness, the scanning path and the scanning speed is also supported, and online regulation and control of the forming quality and online repair of forming defects are achieved. In addition, the efficient geometric algorithm is adopted, so that the slicing efficiency when the method is used for processing the complex geometric component can be effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of laser additive manufacturing, and in particular to an online real-time dynamic slicing method for laser additive manufacturing. Background Art

[0002] The basic process for laser additive manufacturing is to build a corresponding 3D model according to the part to be formed, and perform layered slicing on the model to obtain the model contour of each slice layer and plan the scanning path. The galvanometer of the LAM equipment guides the laser eye scanning path layer by layer according to the scanning path, and finally obtains a product with a morphology and structure corresponding to the actual model.

[0003] After searching the publication (announcement) number: CN118478018A, a multi-laser scanning method for additive manufacturing is disclosed, which relates to the field of additive manufacturing technology, in order to solve the problem that the projections of multiple scanning overlap lines on all slice layers on the same slice layer completely overlap and are all straight lines, resulting in concentrated heat input to produce thermal stress deformation and overburning of the molten pool, affecting the quality of parts. The multi-laser scanning method for additive manufacturing includes: slicing the part to obtain N slice layers, N≥2; all slice layers are divided into regions so that each slice layer obtains at least two scanning regions; at least two laser scanning devices are used to scan the corresponding scanning area in each slice layer, and the two laser scanning devices form a scanning overlap line for common scanning at the junction of the corresponding two adjacent scanning areas; the projection of the scanning overlap line on at least one slice layer on the adjacent slice layer is staggered with the scanning overlap line on the adjacent slice layer; the scanning overlap line is non-straight.

[0004] Although the above scheme can effectively obtain slicing results, the slicing time of the entire model is too long due to the rich characteristics and complex structure of the model cell structure in laser additive manufacturing, which greatly reduces the work efficiency; and the slicing file is too large to import and export inconveniently; it is also difficult to adjust the layer thickness and scanning strategy required for different forming accuracies in different areas; it also cannot meet the requirements of real-time optimization of parameters such as layer thickness, scanning area, and scanning method during process control. Summary of the invention

[0005] The main purpose of the present invention is to provide an online real-time dynamic slicing method for laser additive manufacturing to solve the limitations of the aforementioned existing methods such as low working efficiency and inability to achieve process parameter optimization function.

[0006] To achieve the above object, the present invention provides the following technical solutions: The working process of the entire LAM system, which includes an online real-time dynamic slicing method for laser additive manufacturing, consists of the following main steps: Step 1: Before LAM forming begins, the slicing software in the server starts to perform partial slicing operations on the workpiece model from bottom to top according to the preset processing path filling method, layer thickness and other parameters and obtains the initial slicing results; Step 2: Transmit the slicing result of this part of the workpiece model to the industrial computer of the LAM equipment through the network to start working; Step 3: During the LAM process, the LAM device receives the slicing results while executing the forming work mode, and the slicing software eliminates the completed forming parts through Boolean operations of the 3D model, thus completing the real-time slicing work.

[0007] Step 4: During the LAM process, the process control system of the LAM equipment will adjust the forming layer thickness and scanning path filling mode according to the actual forming results fed back by the process detection system. The slicing software will perform subsequent slicing processing according to these parameters, thereby realizing the dynamic slicing function.

[0008] Preferably, the slicing result in the step one refers to the slicing result being processed by a workstation or server and then transmitted to the forming CNC software in the industrial computer of the LAM equipment through a high-speed network communication method; the slicing referred to in the step one means first dividing a part of the three-dimensional model of the workpiece into several layers according to a predetermined layer thickness, and then filling the laser scanning path for each layer of the divided model according to a predetermined strip scanning method, and the slicing file is a file containing one or several layers of processing area data and processing path data and stored in a certain format.

[0009] Preferably, the specific operation of the slicing software in step three to eliminate the completed formed part through Boolean operation of the three-dimensional model is: using the edge of the depth map as the outline of a layer of slicing model, using the stretching modeling method to establish a 3D model of the slice based on the outline, and then performing Boolean difference operation on the two models to obtain the model of the remaining part.

[0010] Preferably, the specific steps of the real-time dynamic slicing are as follows: Step 1: Read the workpiece CAD model file through the slicing software in the server, and set the initial LAM process parameters and strip filling mode.

[0011] Step 2: Resample the model surface to obtain a point cloud reconstructed according to the surface morphology and the expected forming accuracy. Set the initial layer thickness in the process parameters as h0, and calculate the point cloud of the Z axis 0 height and its small neighborhood as the first layer of point cloud through the window (ROI) setting. Convert the point cloud into a depth map, and obtain the boundary of the path planning through the image edge.

[0012] Step 3: Fill the laser scanning path in the cross-sectional area according to the strip and contour filling direction. At the same time, the starting point of each scanning path in the strip should be adjusted according to whether it is inside the area. If several layers are planned at a time, after completing the path planning of the current layer, adjust the path filling direction according to the increment preset in the process parameters and refill the next layer.

[0013] Step 4: After the initial part of the slicing file that has completed slicing and path planning is transmitted to the industrial computer of the LAM forming equipment via high-speed Ethernet, the workpiece preparation process can be started.

[0014] Step 5: After completing the forming of several layers in a stage, the formed body surface morphology scanning device detects the formed morphology, and the point cloud of the morphology detection can be reconstructed into the actual forming result according to the steps of point cloud-depth map-stretching modeling, and the remaining model to be further formed is obtained through Boolean difference operation.

[0015] Step 6: When the process parameter optimizer adjusts the layer thickness to h and the scan filling direction angle to θ based on the analysis of the forming process status and results, the dynamic slicing method will re-slice the model according to the new layer thickness and fill the scan path in the updated slice layer according to the new scan filling direction angle θ; or when the optimizer decides to insert a layer with a thickness of h fix When a repair layer is created, the dynamic slicing method calculates the scan area of ​​the repair layer based on the point cloud depth map method. The so-called process parameter optimizer is an intelligent agent built based on the adversarial generative imitation learning method. The process parameter optimizer not only gives the optimized layer thickness, but also gives the optimized path filling direction, based on which the processing path can be filled into the newly planned slice.

[0016] Step 7: Repeat steps 5 and 6 until the entire model is completed.

[0017] Preferably, the point cloud-depth map conversion process based on the projection method is: Step 1: Establish the image coordinate system and divide the coordinate system according to the forming accuracy; Step 2: Take the point with the smallest x and y values ​​in a layer of point cloud with a height of h as the origin of the image coordinate plane; Step 3: Set a window according to the forming accuracy, and the window traverses from the origin in the positive direction of the x-axis and the positive direction of the y-axis; Step 4: The principal value of the z-axis coordinate of each point in the window is used as the pixel value of the corresponding pixel point in the image plane; Step 5: If the window has not traversed all the point clouds, the window is offset by a window width and step 4 is repeated; Step 6: If the current point cloud layer is traversed, the process ends.

[0018] Compared with the prior art, the present invention has the following beneficial effects: 1. In the present invention, not only can the slicing be performed while the preparation is being performed during the forming process, but it can also be combined with a process control system, which can not only save the time required for executing the slicing process, but also realize a fully automatic slicing layer thickness adjustment and scanning strategy.

[0019] 2. The geometric algorithm adopted by the present invention has high computational efficiency: the three-dimensional grid model is converted into a point cloud and then into a depth map, and the complex three-dimensional relative position calculation is converted into two-dimensional image processing, which can effectively improve the geometric processing efficiency.

[0020] 3. The slice layer thickness in the present invention is adjustable, and the forming layer thickness can be adjusted according to the actual forming result, thereby improving the shape and property control capabilities of the forming process.

[0021] Fourth, the present invention can perform online repair: the repair process can be planned according to actual needs, the geometric forming errors in the forming process can be repaired, and the geometric accuracy of the formed body can be ensured. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 This is a flow chart of the real-time dynamic slicing method of the present invention; Figure 2 This is a flow chart of the point cloud-depth map conversion method in the present invention; Figure 3 This is a schematic diagram of the result of point cloud-depth map conversion; Figure 4 A schematic diagram of a laser scanning path filled in a cross-sectional area; Figure 5 A schematic diagram of a laser scan path filling another area. DETAILED DESCRIPTION

[0023] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further explained below in conjunction with specific implementation methods.

[0024] The dynamic slicing function of the present invention will be described with the following examples.

[0025] Preferably, the specific steps of real-time dynamic slicing (such as Figure 1 As shown) are as follows: Step 1: Read the workpiece CAD model file through the slicing software in the server, and set the initial LAM process parameters and strip filling mode.

[0026] Step 2: Resample the model surface to obtain a point cloud reconstructed according to the surface morphology and the expected forming accuracy. Set the initial layer thickness in the process parameters as h0, and calculate the point cloud of the Z axis 0 height and its small neighborhood as the first layer of point cloud through the window (ROI) setting. Convert the point cloud to a depth map (the conversion steps are as follows Figure 2 As shown, the results are Figure 3As shown in the figure), the boundary of the path planning is obtained by finding the edge of the image.

[0027] Step 3: Fill the laser scan path in the cross-sectional area according to the strip and contour filling direction (such as Figure 4 At the same time, the starting point of each scanning path in the strip should be adjusted according to whether it is inside the area. If several layers are planned at a time, after completing the path planning of the current layer, the path filling direction is adjusted according to the increment preset in the process parameters and the next layer is refilled.

[0028] Step 4: After the initial part of the slicing file that has completed slicing and path planning is transmitted to the industrial computer of the LAM forming equipment via high-speed Ethernet, the workpiece preparation process can be started.

[0029] Step 5: After completing the forming of several layers in a stage, the formed body surface morphology scanning device detects the formed morphology, and the point cloud of the morphology detection can be reconstructed into the actual forming result according to the steps of point cloud-depth map-stretching modeling, and the remaining model to be further formed is obtained through Boolean difference operation.

[0030] Step 6: The process parameter optimizer adjusts the layer thickness to h and the scanning filling direction angle to θ based on the analysis of the forming process status and results.

[0031] Step 7: The dynamic slicing method will repeat steps 5 and 6 according to the modified layer thickness h and direction angle θ until the entire model is prepared.

[0032] Step 1: Read the workpiece CAD model file through the slicing software in the server, and set the initial LAM process parameters and strip filling mode.

[0033] Step 2: Resample the model surface to obtain a point cloud reconstructed according to the surface morphology and the expected forming accuracy. Set the initial layer thickness in the process parameters as h0, and calculate the point cloud of the Z axis 0 height and its small neighborhood as the first layer of point cloud through the window (ROI) setting. Convert the point cloud to a depth map (the conversion steps are as follows Figure 2 As shown in the figure), the boundary of the path planning is obtained by finding the edge of the image.

[0034] Step 3: Fill the laser scanning path in the cross-sectional area according to the strip and contour filling direction. At the same time, the starting point of each scanning path in the strip should be adjusted according to whether it is inside the area. If several layers are planned at a time, after completing the path planning of the current layer, adjust the path filling direction according to the increment preset in the process parameters and refill the next layer.

[0035] Step 4: After the initial part of the slicing file that has completed slicing and path planning is transmitted to the industrial computer of the LAM forming equipment via high-speed Ethernet, the workpiece preparation process can be started.

[0036] Step 5: After completing the forming of several layers in a stage, the formed body surface morphology scanning device detects the formed morphology, and the point cloud of the morphology detection can be reconstructed into the actual forming result according to the steps of point cloud-depth map-stretching modeling, and the remaining model to be further formed is obtained through Boolean difference operation.

[0037] Step 6: When the optimizer decides to insert a layer with a thickness of h fix When the repair layer is completed, the dynamic slicing method will calculate the scan area of ​​the repair layer according to the point cloud depth map method. Then, the dynamic slicing method fills the scan path in the repair area according to the scan filling direction angle θ. After the repair layer is completed, the slicing software will continue the previous printing task. For specific steps, please refer to Example 1.

[0038] Steps: End.

[0039] The method is characterized in that a three-dimensional grid model is converted into a set of uniformly distributed point clouds, and the model entity area is determined by converting the point cloud into a depth map, so that the range of the scanning path that needs to be filled can be determined. The filling of the scanning path is to perform strip filling in the filling area according to a predetermined filling mode and filling direction. The depth map can use a projection method to map the height information of the point cloud to a two-dimensional image space; A layer of point cloud is intercepted from the bottom of the model, and a height map is generated from it to determine the entity area of ​​the model. After thresholding and binarizing the height map, Laplace transform is performed to obtain the outline of the area. The outline is stretched to obtain a three-dimensional model of a forming layer. The Boolean difference operation between the model and the original model can remove the sliced ​​and path-planned parts from the original model. This cycle is repeated from bottom to top until the entire model is processed, so the geometric operation efficiency of this method is high.

[0040] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. An online real-time dynamic slicing method for laser additive manufacturing, characterized in that: The working process of the whole LAM system including the method of the present invention is the following main steps: Step 1: Before LAM forming begins, the slicing software in the server starts to partially slice the workpiece model according to the preset processing path filling method, layer thickness and other parameters and obtain the slicing result; Step 2: Send the slicing result of this part of the workpiece model to the industrial computer of the LAM device through network communication to start working; Step 3: During the LAM process, the LAM equipment can perform the forming work while receiving the slicing results, without having to cut the entire model before working, which greatly improves work efficiency.

2. The online real-time dynamic slicing method for laser additive manufacturing according to claim 1, characterized in that: The slicing result in step one refers to the vector data containing the start and end points of the scanning path obtained after geometric calculation and path planning of the model to be formed; the slicing referred to in step one means first dividing the three-dimensional model of the workpiece into thin slices according to a predetermined layer thickness, and then filling each layer of the thin slice with a laser scanning path according to a predetermined strip scanning method, and the slicing file is a file containing one or several layers of processing area data and processing path data and stored in a certain format.

3. The online real-time dynamic slicing method for laser additive manufacturing according to claim 2, characterized in that: In step three, the slicing software eliminates the completed formed part through algebraic and Boolean operations based on the depth map. The specific operation is: the point cloud of the model is obtained by the point cloud resampling method, and the point cloud at the bottom of the model is selected by the height value of the point cloud, and then the point cloud is converted into a depth map. The outer contour of the depth map is obtained by the image contour algorithm, and the three-dimensional basic model that needs to be cut is obtained by stretching the outer contour. At the same time, the surface morphology of the formed body scanned by structured light can be used to obtain a depth map of the actual formed surface in the same way. The difference between the two depth maps can be used to obtain a depth map of the bottom morphology of the unprocessed part; During the forming process, the slicing software can adaptively adjust the thickness of the model slices according to the actual forming results, or plan the repair layer for morphology repair as needed, and then plan the laser scanning path of the current layer according to the minimum input difference. The layer thickness that needs to be adjusted or the repair layer that needs to be supplemented is given by the server in the LAM forming process. The adaptive layer height adjustment determines the layer thickness of the next layer based on the mean of the difference, and at the same time, the depth map contour is stretched into a 3D thin layer according to the layer height, and then Boolean operations are performed with the original CAD model. The threshold method is used to determine whether the difference area in the figure meets the conditions for planning the repair layer. If the conditions are met, the repair mode is first formulated and then the path is planned according to the properties of the reserved area after thresholding.

4. The online real-time dynamic slicing method for laser additive manufacturing according to claim 3, characterized in that: The specific steps of the real-time dynamic slicing are as follows: Step 1: Read the workpiece CAD model file through the slicing software in the server, and set the initial LAM process parameters and strip filling mode; Step 2: Resample the model surface to obtain a point cloud reconstructed according to the surface morphology and the expected forming accuracy. Set the initial layer thickness in the process parameters as h0, and use the window (ROI) setting to take the point cloud at the Z axis 0 height and its small neighborhood as the first layer of point cloud. Convert the point cloud into a depth map, and obtain the boundary of the path planning through the image edge calculation; Step 3: Fill the laser scanning path in the cross-sectional area according to the strip and contour filling direction. At the same time, the starting point of each scanning path in the strip should be adjusted according to whether it is inside the area. If several layers are planned at a time, after completing the path planning of the current layer, adjust the path filling direction according to the increment preset in the process parameters and refill the next layer; Step 4: After the initial part of the slice file that has completed the slicing and path planning is transmitted to the industrial computer of the LAM forming equipment via high-speed Ethernet, the workpiece preparation process can be started; Step 5: After the forming of several layers in a stage is completed, the formed body surface morphology scanning device detects the formed morphology, and the point cloud of the morphology detection can be reconstructed into the actual forming result according to the steps of point cloud-depth map-stretching modeling, and the remaining model to be formed is obtained through Boolean difference operation; Step 6: When the process parameter optimizer adjusts the layer thickness to h and the scan filling direction angle to θ based on the analysis of the forming process status and results, the dynamic slicing method will re-slice the model according to the new layer thickness and fill the scan path in the updated slice layer according to the new scan filling direction angle θ; or when the optimizer decides to insert a layer with a thickness of h fix When the repair layer is being planned, the dynamic slicing method will calculate the scan area of ​​the repair layer based on the point cloud depth map method. The so-called process parameter optimizer is an intelligent agent built based on the adversarial generative imitation learning method. The process parameter optimizer will not only give the optimized layer thickness, but also the optimized path filling direction, based on which the processing path can be filled into the newly planned slice; Step 7: Repeat steps 5 and 6 until the entire model is completed.

5. The online real-time dynamic slicing method for laser additive manufacturing according to claim 4, characterized in that: Point cloud-depth map conversion process based on projection method: Step 1: Establish the image coordinate system and divide the coordinate system according to the forming accuracy; Step 2: Take the point with the smallest x and y values ​​in the z = 0 plane in the model coordinates as the origin of the image coordinate plane; Step 3: Set a window according to the forming accuracy, and the window traverses from the origin in the positive direction of the x-axis and the positive direction of the y-axis; Step 4: The principal value of the z-axis coordinate of each point in the window is used as the pixel value of the corresponding pixel point in the image plane; Step 5: If the window has not traversed all the point clouds, the window is offset by a window width and step 4 is repeated; Step 6: If the current point cloud layer is traversed, the process ends.

Images