Compensation-based unsupported rudder wing additive manufacturing method and device

Abstract

The invention provides an unsupported rudder wing additive manufacturing method and device based on compensation, and belongs to the field of metal additive manufacturing. The method comprises the following steps: performing selective laser melting (SLM) simulation on a to-be-manufactured rudder wing by using a simulator to obtain simulation deformation data; based on a penalty function, adding and fusing the forming error and the geometric constraint to obtain a fusion objective function; the simulation deformation data, the node coordinates and the predicted node coordinates corresponding to the discrete node set of the rudder wing to be manufactured are substituted into the fusion target function, and target reversible deformation data are obtained through iterative optimization; determining a discrete node set based on a rudder wing three-dimensional manufacturing model; and performing reverse deformation compensation on the rudder wing three-dimensional manufacturing model by using the target reverse deformation data until a target rudder wing manufactured by the compensated rudder wing three-dimensional manufacturing model meets a preset condition, and outputting the target rudder wing. In conclusion, the method and the device can get rid of dependence on subjective experience, improve compensation precision and adapt to various application scenes.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the field of metal additive manufacturing, and in particular to a compensation-based support-free rudder wing additive manufacturing method and device. BACKGROUND

[0002] At present, in order to ensure the strict requirements of lightweight and high performance, the existing additive manufacturing of large-size rudder wings mostly adopts a support-free structure. However, the support-free structure method may cause residual stress deformation, thermal deformation, and molten pool collapse. In order to solve the above problems, most of the current methods rely on artificial experience for anti-deformation compensation. However, the accuracy of the anti-deformation compensation is difficult to control, it is difficult to accurately match the deformation law of the process, and it is difficult to adapt to various application scenarios. SUMMARY

[0003] The present disclosure provides a compensation-based support-free rudder wing additive manufacturing method and device to solve the problem that most of the current methods rely on artificial experience for anti-deformation compensation, the accuracy of the anti-deformation compensation is difficult to control, it is difficult to accurately match the deformation law of the process, and it is difficult to adapt to various application scenarios.

[0004] According to one aspect of the present disclosure, a compensation-based support-free rudder wing additive manufacturing method is provided, the method comprising: performing selective laser melting (SLM) simulation on a rudder wing to be manufactured by using a simulator to obtain simulation deformation data; adding and fusing a forming error and a geometric constraint based on a penalty function to obtain a fused target function; substituting each simulation deformation data, node coordinates, and predicted node coordinates corresponding to a discrete node set of the rudder wing to be manufactured into the fused target function to obtain target anti-deformation data through iterative optimization; the discrete node set is determined based on a three-dimensional manufacturing model of the rudder wing; and performing reverse deformation compensation on the three-dimensional manufacturing model of the rudder wing by using the target anti-deformation data until a target rudder wing manufactured by the compensated three-dimensional manufacturing model of the rudder wing meets a preset condition, and outputting the target rudder wing.

[0005] In addition, according to the method of one aspect of the present disclosure, the simulation deformation data includes at least one of the following: deformation direction, deformation amount, high stress area, and overhang angle distribution; the simulation deformation data is obtained by performing laser selective laser melting (SLM) simulation on the rudder wing to be manufactured by using a simulator, including: inputting SLM parameters and material properties of the rudder wing to be manufactured into the simulator; the SLM parameters include at least one of the following: laser power, scanning speed, and spot diameter; the material properties include at least one of the following: thermal conductivity, linear expansion coefficient, and elastic modulus; performing thermal-mechanical coupling simulation by using the simulator based on the SLM parameters and the material properties, extracting the deformation direction, the deformation amount, the high stress area, and the overhang angle distribution, and determining as the simulation deformation data.

[0006] Further, according to the method of one aspect of the present disclosure, the forming error is a forming sum-of-squares error of the node coordinates and the predicted node coordinates; and the geometric constraint includes a deformation constraint and a dihedral angle constraint.

[0007] Further, according to the method of one aspect of the present disclosure, the forming error and the geometric constraint are added and fused based on a penalty function to obtain a fused target function, including: accumulating and summing the difference between the node coordinates and the predicted node coordinates for all nodes of the discrete node set to obtain the forming error; when the deformation constraint and / or the dihedral angle constraint exist, determining the corresponding penalty coefficient respectively, and determining the geometric constraint based on the penalty coefficient and the penalty function; and linearly adding the forming error and the geometric constraint to obtain the fused target function.

[0008] Further, according to the method of one aspect of the present disclosure, the simulation deformation data corresponding to the discrete node set of the rudder to be manufactured, the node coordinates and the predicted node coordinates are substituted into the fused target function to obtain the target inverse deformation data through iterative optimization, including: determining the predicted node coordinates by using an iterative algorithm based on the node coordinates and the simulation deformation data; the iterative algorithm includes at least one of a genetic algorithm and a gradient descent algorithm; substituting all the simulation deformation data, the node coordinates and the predicted node coordinates into the fused target function to determine a candidate function value; when the candidate function value meets a convergence threshold, obtaining the target inverse deformation data based on the candidate function value; and when the candidate function value does not meet the convergence threshold, iteratively updating the predicted node coordinates until the updated candidate function value meets the convergence threshold to obtain the target inverse deformation data.

[0009] Further, according to the method of one aspect of the present disclosure, the preset conditions include at least one of density, accuracy, surface roughness, mechanical properties and internal defects meeting a preset threshold.

[0010] Further, according to the method of one aspect of the present disclosure, the target inverse deformation data is used to perform inverse deformation compensation on the rudder three-dimensional manufacturing model until the target rudder manufactured by the compensated rudder three-dimensional manufacturing model meets the preset conditions, and the target rudder is output, including: determining an inverse deformation offset amount by using the target inverse deformation data, and performing inverse deformation offset adjustment on the rudder three-dimensional manufacturing model to determine a rudder compensation model; when a candidate rudder manufactured by using the rudder compensation model meets the preset conditions, the candidate rudder is determined as the target rudder; and when the candidate rudder manufactured by using the rudder compensation model does not meet the preset conditions, the target inverse deformation data is updated until the updated candidate rudder meets the preset conditions to obtain the target rudder.

[0011] Further, according to the method of one aspect of the present disclosure, the method further includes monitoring the target rudder by using a sensor and feeding back.

[0012] Further, according to the method of one aspect of the present disclosure, the sensor comprises at least one of: a coaxial molten pool monitoring sensor and a laser ranging sensor.

[0013] According to another aspect of the present disclosure, a compensation-based support-free rudder vane additive manufacturing device is provided, the device comprising: an emulation unit configured to perform selective laser melting (SLM) emulation on a to-be-manufactured rudder vane by using an emulator to obtain simulation deformation data; a fusion unit configured to add and fuse a forming error and a geometric constraint based on a penalty function to obtain a fusion target function; a reverse deformation unit configured to substitute each simulation deformation data corresponding to a discrete node set of the to-be-manufactured rudder vane, a node coordinate, and a predicted node coordinate into the fusion target function, and iteratively optimize to obtain target reverse deformation data; the discrete node set is determined based on a three-dimensional manufacturing model of the rudder vane; and a manufacturing unit configured to perform reverse deformation compensation on the three-dimensional manufacturing model of the rudder vane by using the target reverse deformation data, until a target rudder vane manufactured by the compensated three-dimensional manufacturing model of the rudder vane meets a preset condition, and output the target rudder vane.

[0014] The present disclosure provides a compensation-based support-free rudder vane additive manufacturing method and device. The present disclosure performs selective laser melting (SLM) emulation on a to-be-manufactured rudder vane by using an emulator to obtain simulation deformation data; adds and fuses a forming error and a geometric constraint based on a penalty function to obtain a fusion target function; substitutes each simulation deformation data corresponding to a discrete node set of the to-be-manufactured rudder vane, a node coordinate, and a predicted node coordinate into the fusion target function, and iteratively optimizes to obtain target reverse deformation data; the discrete node set is determined based on a three-dimensional manufacturing model of the rudder vane; and performs reverse deformation compensation on the three-dimensional manufacturing model of the rudder vane by using the target reverse deformation data, until a target rudder vane manufactured by the compensated three-dimensional manufacturing model of the rudder vane meets a preset condition, and output the target rudder vane. In this way, compared with the existing manual experience compensation method, the present disclosure is free from the dependence on subjective experience, accurately captures the deformation law under the support-free manufacturing scenario through SLM emulation, and simultaneously realizes the collaborative optimization of forming accuracy and geometric constraint by combining the fusion target function constructed by the penalty function to output the target reverse deformation data that is adapted to the process characteristics. The present disclosure can significantly improve the accuracy and stability of reverse deformation compensation, effectively suppresses problems such as molten pool collapse and excessive deformation, and simultaneously adapts to the manufacturing requirements of rudder vanes of different sizes and structures and various application scenarios. In summary, the technical solution provided by the present disclosure can improve the compensation accuracy, accurately match the process deformation law, and can adapt to various application scenarios.

[0015] It is to be understood that both the foregoing general description and the following detailed description are exemplary, and are intended to provide further explanation of the subject technology. BRIEF DESCRIPTION OF DRAWINGS

[0016] The above and other objects, features and advantages of the present disclosure will become more apparent from the following detailed description when taken in conjunction with the accompanying drawings. The drawings provided below are for illustrative purposes and are included to further convey the principles of the present disclosure and are not intended to limit the scope of the present disclosure. The same reference numbers in different drawings represent the same components or steps.

[0017] Figure 1 A flowchart of a compensation-based support-free rudder wing additive manufacturing method provided by an embodiment of the present disclosure;

[0018] Figure 2 A schematic diagram of a beam-rib sub-point array type provided by an embodiment of the present disclosure;

[0019] Figure 3 A printing placement diagram and a finished product schematic diagram provided by an embodiment of the present disclosure;

[0020] Figure 4 A structural block diagram of a compensation-based support-free rudder wing additive manufacturing device provided by an embodiment of the present disclosure. DETAILED DESCRIPTION

[0021] In order to make the objects, technical solutions and advantages of the present disclosure more obvious, the following will describe the example embodiments according to the present disclosure in detail with reference to the drawings. Obviously, the described embodiments are only part of the embodiments of the present disclosure, not all embodiments of the present disclosure, and it should be understood that the present disclosure is not limited by the example embodiments described here.

[0022] At present, for large-size rudder wings, in order to ensure the strict requirements of lightweight and high performance, most of the existing additive manufacturing methods use support-free structure. However, the use of support-free structure may cause residual stress deformation, thermal deformation, and molten pool collapse. To solve the above problems, most of the current methods rely on manual experience for anti-deformation compensation. However, the accuracy of the anti-deformation compensation is difficult to control, it is difficult to accurately match the deformation law of the process, and it is difficult to adapt to various application scenarios.

[0023] Therefore, in view of the above problems, the present disclosure provides a compensation-based support-free rudder wing additive manufacturing method, which can eliminate the dependence on subjective experience, accurately capture the deformation law under the support-free manufacturing scenario through SLM simulation, and simultaneously realize the collaborative optimization of forming accuracy and geometric constraints by combining the fusion objective function constructed by the penalty function to output the target anti-deformation data adapted to the process characteristics. The present disclosure can significantly improve the accuracy and stability of anti-deformation compensation, effectively suppress the problems of molten pool collapse and excessive deformation, and simultaneously adapt to the manufacturing requirements of rudders of different sizes and structures and various application scenarios.

[0024] First, the present disclosure provides a rudder wing additive manufacturing method. Please refer to Figure 1 , Figure 1 A flowchart of a compensation-based support-free rudder wing additive manufacturing method is provided for embodiments of the present disclosure. As shown in Figure 1 , the method comprises:

[0025] In step S101, a simulation of selective laser melting (SLM) is performed on the rudder wing to be manufactured using a simulator to obtain simulation deformation data;

[0026] In step S102, the forming error and geometric constraints are added and fused based on a penalty function to obtain a fused objective function;

[0027] In step S103, the simulation deformation data, node coordinates, and predicted node coordinates corresponding to the discrete node set of the rudder wing to be manufactured are substituted into the fused objective function, and iterative optimization is performed to obtain target counter-deformation data; the discrete node set is determined based on the three-dimensional manufacturing model of the rudder wing;

[0028] In step S104, the three-dimensional manufacturing model of the rudder wing is counter-deformed using the target counter-deformation data until the target rudder wing manufactured from the compensated three-dimensional manufacturing model of the rudder wing meets the preset conditions, and the target rudder wing is output.

[0029] In the present disclosure, the simulator can be understood as a professional simulation tool or system with the ability to simulate process multi-physics field coupling calculation, and its core is to construct a virtual simulation scene equivalent to the actual manufacturing environment through numerical simulation algorithms (such as finite element method, discrete element method, etc.), and output real simulation results reflecting the deformation and stress characteristics of the rudder wing during the manufacturing process.

[0030] In the present disclosure, selective laser melting (SLM) can be understood as an additive manufacturing technology based on metal powder, and its principle is to selectively irradiate a metal powder bed according to a predetermined three-dimensional model slicing path using a high-energy laser beam, so that the powder particles are rapidly melted and fused under the action of the laser, and are gradually solidified and accumulated along the laser scanning path, and finally a dense three-dimensional solid part is formed.

[0031] In this disclosure, simulation deformation data can be understood as a set of quantitative data output by an SLM simulator after simulating the rudder manufacturing process, which comprehensively reflects the deformation and stress state of the rudder. The simulation deformation data of this disclosure may include: deformation direction, deformation amount, high-stress areas, and overhang angle distribution. Specifically, deformation direction refers to the displacement direction (e.g., warping, contraction, bending) of various characteristic regions of the rudder (such as wingtip, wing root, web, etc.) under thermal and residual stress; deformation amount refers to the magnitude of displacement of each characteristic region along the deformation direction, which is the quantitative basis for anti-deformation compensation; high-stress areas refer to stress concentration areas detected in the simulation, which are prone to plastic deformation or cracking and require careful adaptation in anti-deformation design; and overhang angle distribution refers to the distribution of areas in the rudder structure that exceed the critical angle for self-support, which are prone to molten pool collapse, and its distribution data can help optimize the geometric rationality of the anti-deformation scheme.

[0032] In this disclosure, the penalty function can be understood as a mathematical tool to transform a constrained optimization problem into an unconstrained optimization problem. Its core logic is: when the optimization scheme satisfies geometric constraints, the penalty function term takes the value of zero or a minimum, without affecting the core objective (minimizing the forming error); when the optimization scheme violates geometric constraints, the penalty function term generates a corresponding penalty value based on the degree of violation, with larger penalty values ​​for more severe violations, thus forcing the optimization process to converge in the direction that satisfies the constraints. The penalty function integrates the two major objectives of forming error and geometric constraint compliance into a single optimization objective, simplifying the solution process.

[0033] In this disclosure, forming error can be understood as a quantitative index of the deviation between the actual geometric shape of the rudder after forming and the original design shape, and it serves as the basis for measuring the effect of anti-deformation compensation. The forming error in this disclosure is the sum of squares of the forming error between the nodal coordinates and the predicted nodal coordinates, that is, the sum of the deviations in three-dimensional space between the original nodal coordinates and the predicted nodal coordinates after applying anti-deformation for each node in the discrete nodal set, obtained by squaring the values. This calculation method can amplify the influence weight of significant deviations, while avoiding the mutual cancellation of positive and negative deviations, and more accurately reflects the overall forming accuracy.

[0034] In this disclosure, geometric constraints can be understood as restrictive conditions imposed on the anti-deformation scheme of the rudder based on the characteristics of SLM technology, material properties, and equipment processing capabilities, and these constraints can be flexibly adjusted. The geometric constraints of this disclosure include deformation constraints and overhang angle constraints. Specifically, deformation constraints mean that the maximum anti-deformation amount of each feature node must not exceed the material's plastic deformation limit and the SLM equipment processing capability threshold; overhang angle constraints mean that the overhang angle of all surface elements must be controlled within the threshold of the self-supporting critical angle.

[0035] In this disclosure, the fusion objective function can be understood as a multi-objective optimization mathematical model that balances the accuracy of forming and anti-deformation compensation with the geometric feasibility of the rudder. Its core is to avoid compensation deviations or structural failures caused by single-objective optimization by integrating the two major objectives of forming error and geometric constraint compliance. Specific determination methods are detailed below and will not be elaborated here.

[0036] In this disclosure, the discrete node set can be understood as the set of discretized points extracted from the surface and key internal regions of the 3D manufacturing model of the rudder to be manufactured after meshing, used to characterize the geometry and deformation features of the rudder. The selection of nodes can follow the principle of densification in key regions and simplification in non-key regions, and there are no specific restrictions.

[0037] In this disclosure, node coordinates can be understood as the original geometric coordinates of each node in the discrete node set in a preset three-dimensional coordinate system (usually with the center of the SLM equipment worktable as the origin, the X-axis along the length of the worktable, the Y-axis along the width of the worktable, and the Z-axis along the machining height). Their values ​​are directly derived from the three-dimensional manufacturing model of the rudder to be manufactured and are the reference coordinates for calculating node deformation and inverse deformation.

[0038] In this disclosure, the predicted node coordinates can be understood as the expected three-dimensional coordinates of each node in the corrected three-dimensional manufacturing model of the rudder after applying inverse deformation to each node.

[0039] In this disclosure, iterative optimization can be understood as an optimization method that takes minimizing the value of the fusion objective function as the core objective and gradually approaches the optimal solution through a cyclical process of repeatedly adjusting the coordinates of the prediction nodes and recalculating the objective function value.

[0040] In this disclosure, the target anti-deformation data can be understood as a set of quantified data used to reverse-correct the original rudder model after multi-objective fine-tuning by a fusion objective function. Because it corresponds to and compensates in reverse with the simulation deformation data, the deformation direction of the target anti-deformation data is opposite to that of the simulation deformation data. The deformation amount also needs to be dynamically adjusted based on the optimization results of the fusion objective function. At the same time, it needs to adapt to the stress release requirements of high-stress areas and the self-support requirements of overhang angle areas to ensure that the deformation generated by the corrected model during manufacturing process exactly cancels the original simulation deformation, and the final formed part meets the design requirements.

[0041] In this disclosure, the three-dimensional manufacturing model of the rudder can be understood as a three-dimensional digital model of the rudder used for processing by SLM equipment.

[0042] In this disclosure, the preset conditions can be understood as a set of quantitative evaluation criteria for judging whether the target rudder meets the manufacturing requirements. The preset conditions of this disclosure include: at least one of density, accuracy, surface roughness, mechanical properties, and internal defects meeting a preset threshold. Among them, density can be understood as the ratio of the actual dense volume to the theoretical geometric volume after the rudder is formed (usually expressed as a percentage); accuracy can be understood as the degree of conformity between the actual geometric parameters (such as size, shape, and position) of the target rudder and the corresponding parameters of the original design model; surface roughness can be understood as the degree of microscopic unevenness on the surface of the rudder, which can reflect the surface processing quality; mechanical properties can be understood as the physical characteristics exhibited by the rudder under stress, which is a core indicator for measuring the load-bearing capacity and service reliability of the part; internal defects can be understood as structural incompleteness existing inside the rudder that exceeds the allowable range, including pores, cracks, lack of fusion, inclusions, etc.

[0043] In this disclosure, the target rudder can be understood as a rudder product manufactured by SLM process using a 3D manufacturing model of the rudder corrected by the target anti-deformation data, and whose various performance indicators (density, accuracy, surface roughness, mechanical properties, internal defects) meet the preset conditions. It can take into account both lightweight and high performance requirements, and has no obvious forming defects such as deformation, cracking, or molten pool collapse.

[0044] Specifically, the following steps can be performed when additively manufacturing control wings: Step 1: Start the simulator to perform a full-process SLM simulation (covering powder melting, layer-by-layer deposition, solidification and cooling, and stress evolution). After the simulation, extract the simulation deformation data (including deformation direction, deformation amount, high-stress area distribution, and overhang angle distribution of each feature region). The main deformation parts, maximum deformation amount, and risk areas (such as high-stress cracking risk areas and large overhang angle collapse risk areas) of the control wing can be analyzed using data visualization tools (such as cloud maps and line graphs). Step 2: Construct a fusion objective function, where the first objective term is forming error, and the second objective term is geometric constraint compliance (penalizing anti-deformation schemes that exceed the self-supporting critical angle and stress concentration threshold through a penalty function). Substitute the simulation deformation data corresponding to the discrete node set, the original node coordinates, and the initial predicted node coordinates into the fusion objective function for iterative fine-tuning, dynamically optimizing the predicted coordinates of each node, and finally outputting target anti-deformation data that balances error control and geometric rationality. Step 3: Based on the target inverse deformation data, perform inverse deformation correction on the original 3D design model (e.g., for wingtip upturn deformation, apply the corresponding inverse deformation amount in the opposite direction), generate the corrected rudder 3D manufacturing model, and determine whether the corrected target rudder meets the preset conditions. If not, return to step 2 to readjust the target inverse deformation data until the verification is successful.

[0045] The following will explain in detail how to obtain simulation deformation data, including:

[0046] Input the SLM parameters and material properties of the rudder to be manufactured into the simulator; the SLM parameters include at least one of the following: laser power, scanning speed, and spot diameter; the material properties include at least one of the following: thermal conductivity, coefficient of linear expansion, and elastic modulus;

[0047] Based on SLM parameters and material properties, a thermo-mechanical coupling simulation is performed using a simulator to extract deformation direction, deformation amount, high stress area and overhang angle distribution, which are then used as simulation deformation data.

[0048] In this disclosure, SLM parameters can be understood as a set of process parameters that regulate the SLM additive manufacturing process and affect the forming quality and efficiency. Their values ​​directly determine the interaction effect between the laser and the metal powder (such as melting sufficiency and molten pool stability), thereby affecting the forming accuracy, density, and internal stress state of the rudder. The SLM parameters of this disclosure may include, but are not limited to, at least one of the following: laser power, scanning speed, and spot diameter. Among them, laser power can be understood as the energy intensity of the laser beam output in the SLM equipment, which is the core energy input parameter that determines whether the metal powder can be melted and fused quickly; scanning speed can be understood as the speed at which the laser beam moves along a preset path on the surface of the metal powder bed; spot diameter can be understood as the effective size of the spot formed on the powder bed surface after the laser beam is focused, and its size determines the concentration of laser energy and the width of a single cladding pass.

[0049] In this disclosure, material properties can be understood as the inherent physical and mechanical characteristics of the metal material used for the rudder, which affect heat conduction, phase transformation, and stress evolution during the SLM forming process. These properties are the fundamental inputs for the simulator to accurately simulate deformation and stress states. The material properties disclosed in this disclosure may include, but are not limited to, at least one of the following: thermal conductivity, coefficient of linear expansion, and elastic modulus. Thermal conductivity can be understood as the material's ability to conduct heat, reflecting the rate at which heat from the molten pool is transferred to the surrounding powder and substrate during the SLM process. It directly affects the cooling rate of the molten pool and the temperature field distribution, thus affecting residual stress and deformation magnitude. The coefficient of linear expansion can be understood as the elongation or contraction rate per unit length of the material when the temperature changes. The elastic modulus can be understood as the material's ability to resist elastic deformation, characterizing the material's stiffness under stress, and is a key parameter for the simulator to calculate stress and deformation.

[0050] In this disclosure, thermo-mechanical coupling simulation can be understood as a multi-physics coupled numerical simulation method that combines heat conduction analysis with stress-strain analysis based on the physical essence of the SLM process. Its core is to first calculate the temperature field evolution of the rudder during the SLM process (laser melting, layer-by-layer deposition, solidification and cooling) (such as melt pool temperature, cooling rate, and temperature gradient), then input the temperature field results as loads into the stress analysis module. Considering the changes in material thermophysical properties with temperature, solidification phase transformation effects, and other factors, the module calculates the thermal stress, residual stress, and corresponding deformation response caused by temperature changes, ultimately achieving accurate simulation of the thermal and mechanical interaction during the rudder manufacturing process.

[0051] Specifically, the following steps can be performed when obtaining simulation deformation data: Step 1: Organize the SLM process parameters, determining the specific values ​​for laser power, scanning speed, and spot diameter, as well as the complete material property parameters used for the rudder, including thermal conductivity, coefficient of linear expansion, elastic modulus, and other parameters in different temperature ranges, ensuring that the correlation between parameters and temperature matches the actual working conditions. Step 2: Configure the boundary thermal conditions, mechanical conditions, and process boundaries in the simulator consistent with actual SLM manufacturing. Start the simulator and perform step-by-step simulation according to the SLM manufacturing sequence to realize the iterative process of powder laying, scanning, temperature field calculation, and stress deformation calculation until the layer-by-layer stacking simulation of the entire rudder is completed. Step 3: After the simulation, extract core data from the simulation results: Extract the deformation direction by analyzing the displacement vectors of each node to determine the main deformation direction of multiple feature regions; extract the deformation amount by calculating the displacement amplitude of each feature region node; identify high-stress regions by extracting regions where the stress value exceeds the material yield strength threshold, and determine the location, range, and peak stress of stress concentration; determine the overhang angle distribution by calculating the angle between each element and the horizontal plane, filtering out regions exceeding the self-supporting critical angle, and forming an overhang angle distribution cloud map; finally, perform noise reduction processing on the extracted data, and integrate it into the final simulation deformation data set.

[0052] The following section will explain in detail how to determine the fusion objective function, including:

[0053] For all nodes in the discrete node set, the differences between the node coordinates and the predicted node coordinates are accumulated and summed to obtain the shaping error.

[0054] When deformation constraints and / or overhang angle constraints exist, the corresponding penalty coefficients are determined respectively, and the geometric constraints are determined based on the penalty coefficients and penalty functions;

[0055] The forming error and geometric constraints are linearly added together to obtain the fusion objective function.

[0056] In this disclosure, the penalty coefficient can be understood as a weighting parameter used to adjust the importance of geometric constraints in the fusion objective function. Its core function is to balance the optimization priority between forming accuracy and process feasibility. The value of the penalty coefficient can be dynamically adjusted according to the actual needs of rudder manufacturing.

[0057] In this disclosure, linear addition can be understood as a mathematical processing method that numerically superimposes the forming error term and the geometric constraint penalty term according to a fixed ratio. Its core feature is maintaining the independence and traceability of the two optimization objectives. This superposition method is not a simple numerical summation, but rather integrates geometric compliance while clearly prioritizing accuracy through preset weights (the penalty coefficient already contains weight attributes). The advantage of linear addition is that during the optimization process, the change in a single objective function value can simultaneously reflect the reduction of forming error and the improvement of constraint compliance, simplifying the logic for judging iterative convergence.

[0058] Specifically, the following steps can be performed when determining the fusion objective function: Step 1: Traverse all nodes in the discrete node set, calculate the three-dimensional difference between the node coordinates and the predicted node coordinates for each node, square the differences of each axis, and sum them to obtain the error contribution value of a single node. Finally, accumulate the error contribution values ​​of all nodes to form the specific value of the forming error term. Step 2: For deformation constraints, combine the plastic deformation limit of the rudder material and the maximum processing offset capability of the SLM equipment to determine the allowable deformation threshold for each node; for overhang angle constraints, mark the regions exceeding the threshold based on the powder self-supporting critical angle, etc., and construct a geometric constraint penalty. That is, if the node deformation exceeds the deformation constraint threshold, calculate the deformation constraint penalty value as (actual deformation - allowable threshold) × the corresponding penalty coefficient; if the overhang angle of the region where the node is located exceeds the critical value, calculate the overhang angle constraint penalty value as (actual overhang angle - critical angle) × the corresponding penalty coefficient. Sum the two types of penalty values ​​for all nodes to obtain the final value of the geometric constraint penalty term. Step 3: Directly add the forming error term and the geometric constraint penalty term to form the complete fusion objective function.

[0059] For example, this disclosure also provides specific embodiments for determining the fusion objective function, including:

[0060] First, the forming error target term is defined, which can satisfy the following formula:

[0061]

[0062] Among them, U comp To compensate for displacement. F obj (U comp S represents the calculated value of the fusion objective function after compensation. design,i The coordinates of the i-th node in the theoretical design model (i.e., the node coordinates of this disclosure). S comp,iTo compensate for the coordinates of the i-th node in the model (i.e., the predicted node coordinates of this disclosure). D predicted,i To compensate for the predicted reverse deformation of the i-th node in the model during the printing process.

[0063] Next, the geometric constraints are defined, which can satisfy the following formula:

[0064] g i (U comp )≤0

[0065] Among them, g i This is the geometric constraint function.

[0066] Deformation constraints can satisfy the following formula:

[0067] g1 = T min -T current ≤0

[0068] Where g1 is the deformation constraint function. min T is the minimum allowable deformation threshold. current This represents the predicted deformation under the current anti-deformation scheme.

[0069] The overhang angle constraint can satisfy the following formula:

[0070] g2=θ current -θ max ≤0.

[0071] Where g2 is the overhang angle constraint function. θ max This is the maximum permissible overhang angle threshold. θ current This represents the overhang angle under the current anti-deformation scheme.

[0072] Then, based on the penalty function, a fusion objective function for the minimum error objective term and geometric constraints is constructed, which can satisfy the following formula:

[0073] F augmented (U comp ) = F obj (U comp )+P(U comp )

[0074] Wherein, P(U comp F is the penalty function term. augmented (U comp ) is the fusion objective function.

[0075] P(U comp It can satisfy the following formula:

[0076]

[0077] Where, λ iThis is the penalty coefficient.

[0078] The following will explain in detail how to obtain the target inverse deformation data, including:

[0079] Based on node coordinates and various simulation deformation data, the predicted node coordinates are determined using an iterative algorithm; the iterative algorithm includes at least one of the following: genetic algorithm and gradient descent algorithm;

[0080] Substitute all the simulation deformation data, node coordinates, and predicted node coordinates into the fusion objective function to determine the candidate function value;

[0081] When the candidate function value meets the convergence threshold, the target inverse deformation data is obtained based on the candidate function value;

[0082] When the candidate function value does not meet the convergence threshold, the predicted node coordinates are iteratively updated until the updated candidate function value meets the convergence threshold, thus obtaining the target inverse deformation data.

[0083] In this disclosure, iterative algorithms can be understood as a class of numerical optimization methods that take minimizing the value of the fusion objective function as the core objective, and gradually approach the optimal solution by repeatedly adjusting the coordinates of the predicted nodes and recalculating the function value. Its core logic is to use the calculation results of the previous iteration to adjust the optimization direction, eliminate deviations through multiple rounds of iterative iteration, and ensure that the final output of the predicted node coordinates matches the deformation law of the rudder, providing a precise basis for anti-deformation compensation. Specifically, genetic algorithms can combine the predicted node coordinates, randomly generate an initial population in the solution space, determine the fitness of individuals based on the fusion objective function value, select high-quality individuals for crossover and mutation operations, iteratively generate a better population, and finally converge to the globally optimal combination of predicted node coordinates. Gradient descent algorithms can be understood as local optimization algorithms based on function gradient information. By calculating the gradient of the fusion objective function with respect to each predicted node coordinate (i.e., the direction with the largest rate of change of the function value), the predicted node coordinates are adjusted along the opposite direction of the gradient with a preset step size, gradually reducing the objective function value, and can be combined with genetic algorithms.

[0084] In this disclosure, the candidate function value can be understood as a specific numerical value calculated by substituting the predicted node coordinates, the original coordinates of discrete nodes, and the corresponding simulation deformation data obtained in a certain round of iteration into the fusion objective function. This value is the core indicator for evaluating the quality of the current predicted node coordinates.

[0085] In this disclosure, the convergence threshold can be understood as a threshold for determining whether the candidate function value has reached a stable state. That is, when the calculated value of the candidate function changes less than the threshold multiple times or at least once, the optimization process is considered to have converged, and the iterative calculation is stopped. The convergence threshold can be flexibly set in conjunction with the manufacturing precision requirements of the rudder, and no restrictions are imposed here.

[0086] Specifically, the following steps can be performed when determining the target inverse deformation data: Step 1: Select an iterative algorithm based on the complexity of the rudder structure. For complex rudders with multiple curved surfaces and stress concentrations, a genetic algorithm for global optimization is preferred; for relatively simple rudders, a gradient descent algorithm can be used directly, or a combination strategy of global optimization using a genetic algorithm and local fine-tuning using a gradient descent algorithm can be employed. Step 2: Correct the coordinates of each node based on the simulated deformation data to initially determine the predicted node coordinates. Step 3: Substitute the predicted node coordinates, node coordinates, and complete simulated deformation data into the fusion objective function, and obtain the candidate function value for the first iteration through numerical calculation, recording this value as the iteration benchmark. Step 4: Directly compare the current candidate function value with the preset convergence threshold; if the difference is less than the convergence threshold, or if the change in function value for several consecutive iterations is less than the convergence threshold, the optimization is considered converged; if not, adjust the predicted node coordinates according to the rules of the selected iterative algorithm, and substitute the updated predicted node coordinates into the fusion objective function to calculate new candidate function values ​​until the convergence condition is met. Step 5: When the candidate function value meets the convergence threshold, extract the final predicted node coordinates and calculate the difference between the original coordinates and the final predicted coordinates of each node, which is the target inverse deformation data.

[0087] Specifically, the following steps can be performed when determining the target inverse deformation data:

[0088] Multi-objective optimization algorithms (such as NSGA-III algorithm, genetic algorithm, etc.) are used to fuse the objective function F. augmented (U comp As the optimization objective, iterative solution makes F augmented (U comp Minimize the compensation displacement D predicted,i During the iteration process, the algorithm continuously updates the predicted inverse deformation D at each node. predicted,i And calculate the geometric constraint function g in real time. i (U comp ) and penalty function term P(U comp ), until the convergence condition is met (such as F in multiple consecutive iterations). augmented (U comp The change in displacement is less than the convergence value. The optimized compensation displacement U... comp Substitute the 3D model of the rudder wing into the model to generate the manufacturing model after anti-deformation correction; import the model into the SLM simulator for verification simulation, extract the deformation data, density, stress distribution and other indicators after forming, and verify whether the preset conditions are met until the verification is successful.

[0089] The following will explain in detail how to determine whether the preset conditions are met and how to output the target rudder, including:

[0090] Using the target inverse deformation data, the inverse deformation offset is determined, and the inverse deformation offset is adjusted on the three-dimensional manufacturing model of the rudder wing to determine the rudder wing compensation model.

[0091] When a candidate rudder wing manufactured using the rudder wing compensation model meets the preset conditions, it is determined as the target rudder wing.

[0092] When the candidate rudder generated using the rudder compensation model does not meet the preset conditions, the target anti-deformation data is updated until the updated candidate rudder meets the preset conditions, and then the target rudder is obtained.

[0093] In this disclosure, the reverse deformation offset can be understood as a displacement quantization value, opposite to the direction of the simulated deformation, applied to each discrete node based on the target reverse deformation data to counteract the simulated deformation generated during the manufacturing process of the rudder wing SLM. Its core is to convert the reverse deformation displacements (components along the X, Y, and Z axes) of each node in the target reverse deformation data into specific offset parameters that can be directly used for model adjustment, i.e., the compensation displacement mentioned above. This offset can be used to correct the node coordinates of the original rudder wing 3D manufacturing model, causing the model to undergo reverse pre-deformation, so that after actual manufacturing deformation, it precisely returns to the design dimensions and shape.

[0094] Specifically, the following steps can be performed when outputting the target rudder: First, extract the inverse deformation displacement components of each discrete node from the target inverse deformation data and directly determine them as the inverse deformation offset of each node; Second, import the rudder compensation model into the control system of the SLM equipment, match the process parameters consistent with the simulation parameters, and complete the printing of the candidate rudder; Third, perform multi-dimensional, full-dimensional judgments based on preset conditions, verifying whether each candidate rudder meets the preset conditions: High-precision industrial cameras can be used to photograph each surface of the candidate rudder, visually inspecting for defects such as molten pool collapse, surface cracks, and severe spatter; a surface roughness tester can also be used on aerodynamic surfaces and assembly datum surfaces. Detection points are selected in areas such as the surface to measure and determine whether the surface roughness meets the preset threshold. A coordinate measuring machine can be used to measure the actual size of the candidate rudder based on the detection features set in the original design model, and calculate whether the dimensional deviation meets the preset accuracy threshold. An industrial scanner can be used to perform a full-size scan of the candidate rudder, reconstruct a three-dimensional image of the internal structure, analyze the size and distribution of internal defects, and determine whether the preset thresholds for internal defects and density are met. Tensile tests (measuring tensile strength, yield strength, and elongation) can also be performed using an electronic universal testing machine, and impact tests (measuring impact toughness) can be performed using a pendulum impact testing machine to determine whether the preset thresholds for mechanical properties are met. When at least one or more preset conditions are met, the rudder compensation model is determined to have met the compensation standard, and the target rudder is output. When all preset conditions are not met, the target anti-deformation data is updated again until the rudder compensation model meets the compensation standard.

[0095] The following will elaborate on the compensation-based additive manufacturing method for unsupported rudders disclosed herein, which also includes:

[0096] Sensors are used to monitor the target rudder and provide feedback.

[0097] In this disclosure, the sensor can be understood as a detection device used to capture the forming state parameters of the wing in real time during the additive manufacturing process. The sensor disclosed in this disclosure can be at least one of the following: a coaxial molten pool monitoring sensor or a laser ranging sensor. The coaxial molten pool monitoring sensor can be arranged coaxially with the printing laser beam to collect information such as the temperature distribution, width dimension, and liquid metal flow state of the molten pool in real time, directly reflecting the local forming quality; the laser ranging sensor accurately measures the actual height, contour deviation, and overall deformation of each layer of the wing during the printing process through the non-contact laser ranging principle. The two can be used alone or in combination to provide high-fidelity, real-time detection data support for deformation feedback adjustment.

[0098] Specifically, this disclosure can also incorporate a deformation feedback model for deformation feedback. That is, by using the aforementioned sensors to monitor the width / dimension (or actual height) of the molten pool in real time, a simplified deformation feedback model can be established, which can be expressed by the formula ΔP = k*(D actual -D nominal The system calculates and automatically adjusts the laser power based on the monitoring results, changing the local energy input to pull back the deformation trend.

[0099] Where ΔP is the laser power increment that needs to be adjusted. actual This refers to the current layer height deviation or molten pool characteristic parameter measured by the sensor. (D) nominal ) represents the theoretical layer height or standard molten pool characteristic parameter, and k is an empirical coefficient.

[0100] For example, this disclosure also provides a complete compensation-based additive manufacturing process for unsupported rudder wings, including:

[0101] (1) Establishing a coupling method for inverse deformation design and finite element simulation based on the penalty function method (i.e., equivalent to the determination of the fusion objective function in this disclosure):

[0102] Construct an inverse deformation optimization framework, defining the design variable as the displacement vector U of the global nodes of the 3D model. comp ={u x ,u y ,u z The objective function is the sum of squared shape errors between the formed part and the design model: Introducing geometric and process constraint functions g i (U comp )≤0, including: deformation constraint g1=T min -T current≤0 sag angle constraint g2=θ current -θ max ≤0.

[0103] Constructing an augmented objective function using the penalty function method:

[0104] (2) Establish a coupling interface with the simulation software ANSYS Workbench to achieve the following data interaction and process automation:

[0105] 1. Optimize the framework: The compensation model S can be implemented using STL format or API interface. comp Import into ANSYS Workbench; for example, the optimization framework calls the DesignModeler or SpaceClaim interface of Workbench through a Python script, automatically adjusts the coordinates of the model nodes according to the displacement vector \(U_{\text{comp}}\) to generate a compensation model;

[0106] 2. Use the preset thermo-mechanical coupling analysis template in the simulation software to calculate the deformation field D during the printing process. predicted For example, by using the Workbench ACT plugin or WB Script, automatically perform mesh generation, thermo-mechanical coupling solution (including transient temperature field calculation and structural deformation analysis), and export the results to CSV or rst format;

[0107] 3. Extract the geometric parameters (deformation T) of the compensation model from the simulation software. current , suspension angle θ current It is used for constraint function evaluation; for example, the optimization framework reads the simulation result file of Workbench through the pyansys library, extracts the nodal deformation and geometric feature parameters, and uses them for the calculation of objective function and constraint function.

[0108] 4. Feed the deformation field and geometric parameters back to the optimization framework to drive the design variable U. comp Iterative updates. When the augmented objective function converges and the constraints are satisfied, the optimal compensation displacement is output. Generate the final inverse deformation model.

[0109] (3) Printing process parameters can also be optimized during the unsupported simulation process.

[0110] Through extensive process experiments, a laser process parameter library was established for GH4169 material with different structural characteristics. The core of this parameter library is the mapping relationship between the overhang angle (θ) and the laser energy input (E). Key process strategies: Small overhang angle region (θ≤45°): A "low linear energy density" strategy is adopted. Specifically, this involves reducing laser power (P), increasing scanning speed (V), and decreasing scanning spacing (h). For example, for a 60° overhang angle, P = 170-190W, V = 1100-1300mm / s, and h = 0.08-0.10mm can be used. Simultaneously, "contour offset scanning" is employed in this region, i.e., scanning the outer contour of the overhang surface first, then filling the interior, to provide a stable "skeleton." Core rib beam region: To ensure density and strength, a medium linear energy density is used, and a "staggered scanning" or "checkerboard partition scanning" strategy is employed to disperse thermal stress. Skin region: To improve surface quality and forming efficiency, a relatively high scanning speed combined with moderate power can be used.

[0111] (4) The entire process can also be dynamically coordinated and controlled with feedback:

[0112] Substrate preheating and constraint: Preheating the molded substrate to the range of 150℃~200℃ significantly reduces the temperature gradient between the printed layer and the solidified part, thereby reducing thermal stress from the source.

[0113] Simultaneously, the substrate clamping scheme is optimized to provide reasonable constraints. Intelligent scanning path planning: a strategy of "long and short vector hybridization" and "interlayer scanning rotation" (such as rotating each layer by 67° or other golden ratio angles) is adopted to interrupt the continuity of the laser scanning path and prevent excessive accumulation of stress in the same direction.

[0114] For example, Figure 2 This is a schematic diagram of the matrix layout of the reinforcing beams provided in an embodiment of this disclosure. From... Figure 2 It can be seen that the rudder structure using the method disclosed herein can adopt different lattice layout forms, including an overall stiffener-beam frame structure, a grid-like lattice filling structure, and a lattice unit structure with a specific geometric topology. This allows for the partitioned arrangement of stiffeners and lattices, ensuring structural performance while also taking into account lightweight design.

[0115] For example, Figure 3 This is a printed placement diagram and a finished product schematic diagram provided for embodiments of this disclosure. Figure 3 As can be seen, the left side is a schematic diagram of the printed placement model of the rudder wing, and the right side is the actual printed finished product. The finished product matches the model design, indicating that the additive manufacturing method can successfully manufacture a rudder wing structure that meets the design requirements.

[0116] This disclosure also provides an additive manufacturing apparatus for rudder wings. Figure 4A structural block diagram of a compensation-based unsupported rudder additive manufacturing apparatus provided in this disclosure is shown below. Figure 4 As shown, the compensation-based unsupported rudder additive manufacturing apparatus 400 includes:

[0117] Simulation unit 401 is used to perform selective laser melting (SLM) simulation on the rudder to be manufactured using a simulator to obtain simulation deformation data.

[0118] The fusion unit 402 is used to add and fuse the forming error and geometric constraints based on the penalty function to obtain the fusion objective function;

[0119] The anti-deformation element 403 is used to substitute the simulated deformation data, node coordinates and predicted node coordinates corresponding to the discrete node set of the rudder to be manufactured into the fusion objective function, and iteratively optimize to obtain the target anti-deformation data; the discrete node set is determined based on the rudder's three-dimensional manufacturing model.

[0120] Manufacturing unit 404 is used to perform reverse deformation compensation on the three-dimensional manufacturing model of the rudder using the target inverse deformation data, until the target rudder manufactured by the compensated three-dimensional manufacturing model meets the preset conditions, and then outputs the target rudder.

[0121] In one exemplary embodiment, the simulation unit 401 is specifically used for: simulating deformation data including at least one of the following: deformation direction, deformation amount, high-stress region, and overhang angle distribution; using a simulator to perform laser selective melting (SLM) simulation on the rudder to be manufactured to obtain simulated deformation data, including: inputting the SLM parameters and material properties of the rudder to be manufactured into the simulator; the SLM parameters include at least one of the following: laser power, scanning speed, and spot diameter; the material properties include at least one of the following: thermal conductivity, coefficient of linear expansion, and elastic modulus; based on the SLM parameters and material properties, using the simulator to perform thermo-mechanical coupling simulation, extracting the deformation direction, deformation amount, high-stress region, and overhang angle distribution, and determining them as simulated deformation data.

[0122] In one exemplary embodiment, the fusion unit 402 is specifically used for: the forming error being the sum of the squares of the forming coordinates and the predicted node coordinates; the geometric constraints include: deformation constraints and overhang angle constraints.

[0123] In one exemplary embodiment, the fusion unit 402 is specifically used to: accumulate the difference between the node coordinates and the predicted node coordinates for all nodes in the discrete node set and perform a sum of squares to obtain the forming error; when there are deformation constraints and / or overhang angle constraints, determine the corresponding penalty coefficients respectively, and determine the geometric constraints based on the penalty coefficients and the penalty function; and linearly add the forming error and the geometric constraints to obtain the fusion objective function.

[0124] In one exemplary embodiment, the anti-deformation unit 403 is specifically used to: determine the predicted node coordinates based on the node coordinates and each simulated deformation data using an iterative algorithm; the iterative algorithm includes at least one of the following: a genetic algorithm and a gradient descent algorithm; substitute all the simulated deformation data, node coordinates and predicted node coordinates into the fusion objective function to determine the candidate function value; when the candidate function value meets the convergence threshold, obtain the target anti-deformation data based on the candidate function value; when the candidate function value does not meet the convergence threshold, iteratively update the predicted node coordinates until the updated candidate function value meets the convergence threshold, and obtain the target anti-deformation data.

[0125] In one exemplary embodiment, the manufacturing unit 404 is specifically used to: satisfy at least one of the following preset conditions: density, precision, surface roughness, mechanical properties, and internal defects, which meet a preset threshold.

[0126] In one exemplary embodiment, the manufacturing unit 404 is specifically used to: determine the reverse deformation offset using the target reverse deformation data, and adjust the reverse deformation offset of the three-dimensional manufacturing model of the rudder wing to determine the rudder wing compensation model; when the candidate rudder wing manufactured using the rudder wing compensation model meets the preset conditions, it is determined as the target rudder wing; when the candidate rudder wing manufactured using the rudder wing compensation model does not meet the preset conditions, the target reverse deformation data is updated until the updated candidate rudder wing meets the preset conditions, and the target rudder wing is obtained.

[0127] In one exemplary embodiment, the manufacturing unit 404 is further configured to: monitor the target rudder using sensors and provide feedback.

[0128] In one exemplary embodiment, the manufacturing unit 404 is further configured to: include at least one of the following sensors: a coaxial molten pool monitoring sensor and a laser rangefinder sensor.

[0129] In summary, this disclosure provides a compensation-based additive manufacturing method and apparatus for unsupported rudder wings. This disclosure utilizes a simulator to perform selective laser melting (SLM) simulation on the rudder wing to be manufactured, obtaining simulated deformation data. Based on a penalty function, forming errors and geometric constraints are added and fused to obtain a fusion objective function. The simulated deformation data, node coordinates, and predicted node coordinates corresponding to the discrete node set of the rudder wing to be manufactured are substituted into the fusion objective function, and iterative optimization is performed to obtain target inverse deformation data. The discrete node set is determined based on the rudder wing's 3D manufacturing model. The target inverse deformation data is used to perform reverse deformation compensation on the rudder wing's 3D manufacturing model until the target rudder wing manufactured by the compensated 3D manufacturing model meets preset conditions, at which point the target rudder wing is output. Thus, compared to existing manual experience-based compensation methods, this disclosure eliminates reliance on subjective experience, accurately captures deformation patterns in unsupported manufacturing scenarios through SLM simulation, and simultaneously achieves synergistic optimization of forming accuracy and geometric constraints by combining the fusion objective function constructed with the penalty function, outputting target inverse deformation data adapted to process characteristics. This disclosure can significantly improve the accuracy and stability of anti-deformation compensation, effectively suppressing problems such as molten pool collapse and excessive deformation, while adapting to the manufacturing needs of rudder wings of different sizes and structures and diverse application scenarios. In summary, the technical solution provided by this disclosure can improve the accuracy of compensation, accurately match the deformation law of the process, and adapt to a variety of application scenarios.

[0130] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this disclosure.

[0131] The basic principles of this disclosure have been described above with reference to specific embodiments. However, it should be noted that the advantages, benefits, and effects mentioned in this disclosure are merely examples and not limitations, and should not be considered as essential features of each embodiment of this disclosure. Furthermore, the specific details disclosed above are for illustrative and facilitative purposes only, and are not limitations. These details do not limit the scope of this disclosure to the necessity of employing the aforementioned specific details for implementation.

[0132] The block diagrams of devices, apparatuses, devices, and systems disclosed herein are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As those skilled in the art will recognize, these devices, apparatuses, devices, and systems can be connected, arranged, and configured in any manner. Words such as “comprising,” “including,” “having,” etc., are open-ended terms meaning “including but not limited to,” and are used interchangeably with them. The terms “or” and “and” as used herein refer to the terms “and / or,” and are used interchangeably with them unless the context clearly indicates otherwise. The term “such as” as used herein refers to the phrase “such as but not limited to,” and is used interchangeably with it.

[0133] Additionally, as used herein, the “or” used in a list of items beginning with “at least one” indicates a separate list, such that a list of, for example, “at least one of A, B, or C” means A or B or C, or AB or AC or BC, or ABC (i.e., A and B and C). Furthermore, the word “exemplary” does not imply that the described example is preferred or better than other examples.

[0134] It should also be noted that in the systems and methods of this disclosure, the components or steps can be decomposed and / or recombined. These decompositions and / or recombinations should be considered as equivalent solutions to this disclosure.

[0135] Various changes, substitutions, and modifications can be made to the technology described herein without departing from the teachings defined by the appended claims. Furthermore, the scope of the claims of this disclosure is not limited to the specific aspects of the processes, machines, manufactures, events, means, methods, and actions described above. Currently existing or later-developed processes, machines, manufactures, events, means, methods, or actions that perform substantially the same function or achieve substantially the same result as the corresponding aspects described herein can be utilized. Therefore, the appended claims include such processes, machines, manufactures, events, means, methods, or actions within their scope.

[0136] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use this disclosure. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of this disclosure. Therefore, this disclosure is not intended to be limited to the aspects shown herein, but rather to be carried out within the widest scope consistent with the principles and novel features disclosed herein.

[0137] The above description has been given for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of this disclosure to the forms disclosed herein. Although numerous exemplary aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations therein.

Claims

1. A compensation-based, unsupported vane wing additive manufacturing method, characterized by, The method comprises: Performing selective laser melting (SLM) simulation on the to-be-manufactured rudder vane by using a simulator to obtain simulation deformation data; Adding and fusing forming errors and geometric constraints based on a penalty function to obtain a fused target function; Substituting each of the simulation deformation data corresponding to a discrete node set of the to-be-manufactured rudder vane, the node coordinates and the predicted node coordinates into the fused target function, and iteratively optimizing to obtain target counter-deformation data; the discrete node set is determined based on a three-dimensional manufacturing model of the rudder vane; Performing reverse deformation compensation on the three-dimensional manufacturing model of the rudder vane by using the target counter-deformation data until the target rudder vane manufactured by the compensated three-dimensional manufacturing model of the rudder vane meets a preset condition, and outputting the target rudder vane.

2. The method of claim 1, wherein, The simulation deformation data comprises at least one of a deformation direction, a deformation amount, a high stress area and a overhang angle distribution; the simulation deformation data is obtained by performing laser selective melting (SLM) simulation on the to-be-manufactured rudder vane by using a simulator, comprising: Inputting SLM parameters and material properties of the to-be-manufactured rudder vane into the simulator; the SLM parameters comprise at least one of laser power, scanning speed and spot diameter; the material properties comprise at least one of thermal conductivity, linear expansion coefficient and elastic modulus; Performing thermal-mechanical coupling simulation by using the simulator based on the SLM parameters and the material properties, extracting the deformation direction, the deformation amount, the high stress area and the overhang angle distribution, and determining the simulation deformation data.

3. The method of claim 1, wherein, The forming error is a forming sum of squares error of the node coordinates and the predicted node coordinates; the geometric constraint comprises a deformation constraint and an overhang angle constraint.

4. The method of claim 1, wherein, The adding and fusing of the forming error and the geometric constraint based on the penalty function to obtain the fused target function comprises: For all nodes of the discrete node set, accumulating and squaring the difference between the node coordinates and the predicted node coordinates to obtain the forming error; When there is a deformation constraint and / or an overhang angle constraint, respectively determining a corresponding penalty coefficient, and determining the geometric constraint based on the penalty coefficient and the penalty function; Linearly adding the forming error and the geometric constraint to obtain the fused target function.

5. The method of claim 1, wherein, The substituting of each of the simulation deformation data corresponding to the discrete node set of the to-be-manufactured rudder vane, the node coordinates and the predicted node coordinates into the fused target function, and the iteratively optimizing to obtain the target counter-deformation data, comprises: Determining the predicted node coordinates by using an iterative algorithm based on the node coordinates and each of the simulation deformation data; the iterative algorithm comprises at least one of a genetic algorithm and a gradient descent algorithm; Substituting all of each of the simulation deformation data, the node coordinates and the predicted node coordinates into the fused target function to determine a candidate function value; When the candidate function value meets a convergence threshold, obtaining the target counter-deformation data based on the candidate function value; When the candidate function value does not meet the convergence threshold, iteratively updating the predicted node coordinates until the updated candidate function value meets the convergence threshold, and obtaining the target counter-deformation data.

6. The method of claim 1, wherein, The preset conditions include that at least one of compactness, accuracy, surface roughness, mechanical property and internal defect meets a preset threshold value.

7. The method of claim 1, wherein, The reverse deformation compensation of the rudder wing three-dimensional manufacturing model by using the target reverse deformation data is performed until the target rudder wing manufactured by the compensated rudder wing three-dimensional manufacturing model meets the preset conditions, and the target rudder wing is output. The reverse deformation offset amount is determined by using the target reverse deformation data, and the rudder wing three-dimensional manufacturing model is adjusted by reverse deformation offset to determine a rudder wing compensation model; When the candidate rudder wing manufactured by using the rudder wing compensation model meets the preset conditions, the target rudder wing is determined; When the candidate rudder wing manufactured by using the rudder wing compensation model does not meet the preset conditions, the target reverse deformation data is updated until the updated candidate rudder wing meets the preset conditions, and the target rudder wing is obtained.

8. The method of claim 1, wherein, The method further includes: The target rudder wing is monitored by using a sensor and feedback is performed.

9. The method of claim 8, wherein, The sensor includes at least one of a coaxial molten pool monitoring sensor and a laser ranging sensor.

10. A compensation-based, unsupported vane wing additive manufacturing apparatus, comprising: The device includes: An emulation unit is configured to perform selective laser melting (SLM) emulation of a to-be-manufactured rudder wing by using an emulator to obtain emulation deformation data; A fusion unit is configured to add and fuse forming error and geometric constraint based on a penalty function to obtain a fusion target function; A reverse deformation unit is configured to substitute each of the emulation deformation data, node coordinates and predicted node coordinates corresponding to a discrete node set of the to-be-manufactured rudder wing into the fusion target function to obtain target reverse deformation data through iterative optimization; the discrete node set is determined based on a rudder wing three-dimensional manufacturing model; A manufacturing unit is configured to perform reverse deformation compensation of the rudder wing three-dimensional manufacturing model by using the target reverse deformation data until a target rudder wing manufactured by the compensated rudder wing three-dimensional manufacturing model meets preset conditions, and the target rudder wing is output.

Images