A thin-walled part machining allowance optimization method and a machining allowance optimization evaluation method

Abstract

The present application relates to a kind of thin-walled part machining allowance optimization method, comprising: by measuring and CAD / CAM, the point cloud model of workpiece to be processed and theoretical processing point cloud model are obtained;Establish the machining allowance optimization model of the point cloud of workpiece to be processed and theoretical processing point cloud;The machining allowance of thin-walled part is expressed as the constraint multi-objective optimization problem of non-processing surface coincidence, machining surface allowance uniformity;Solve the constraint multi-objective optimization problem, and calculate the Euclidean transformation parameter.The assembly model of the machining allowance optimization model of thin-walled part is established by the point cloud model measured and theoretical processing model, and it is expressed as the constraint multi-objective optimization problem, and the positioning pose of workpiece to be processed is finally solved, to ensure that there is no machining allowance shortage in the machining process, and ensure that machining allowance is evenly distributed, meet the machining demand of thin-walled part.

Description

Technical Field

[0001] This invention relates to the field of thin-walled part machining and positioning technology, specifically to a method for registering the point cloud of a thin-walled workpiece with the point cloud of a nominal workpiece and a method for optimizing and evaluating machining allowance. Background Technology

[0002] Thin-walled components, such as the skin, are key parts of an aircraft's aerodynamic shape. Their manufacturing and assembly precision (clearance, step difference, shape accuracy, etc.) is crucial for ensuring the aircraft's aerodynamic performance, stealth performance, and internal structural safety.

[0003] Skin parts are typically machined using processes such as wire drawing and chemical milling. They are then assembled from multiple areas of a framework consisting of truss members, beams, and other structural elements using riveting. Due to unavoidable manufacturing and assembly errors within the framework, as well as assembly errors, the edge skins cannot perfectly match the assembly design. Therefore, a certain machining allowance must be reserved during the machining stage to ensure that the assembly intersects with adjacent skins.

[0004] Currently, machining allowances are mostly marked manually, which suffers from large human errors and low efficiency. Machining positioning and accuracy inspection tasks typically involve Euclidean transformation of the measurement model relative to the theoretical model to obtain a reasonable deviation distribution of the workpiece. Then, it can be verified whether the deviations meet engineering requirements. The manufacturing process of thin-walled parts is complex; unreasonable machining positioning and insufficient local machining allowances can lead to serious problems such as local deformation, uneven edge distribution, and low positioning datum accuracy, resulting in workpiece scrap.

[0005] Common workpiece allowance optimization models can be divided into constrained least squares models, constrained minimax models, and range minimization models based on the minimum region criterion. These models do not consider the non-machined surfaces of thin-walled parts. Taking the constrained least squares model as an example, the literature "A Unified Localization Approach for Machining Allowance Optimization of Complex Curved Surfaces, Precision Engineering, 2009, 33(4): 516–523" combines a robust algorithm of multivariate Bernstein polynomials and a Bezier surface segmentation algorithm to propose an algorithm based on recursive quadtree decomposition to solve the machining allowance of various types of workpieces and improve the workpiece positioning speed.

[0006] The paper "Research on the localization of the workpieces with large sculptured surfaces in NC machining, International Journal of Advanced Manufacturing Technology, 2004, 23(5-6):429-435" uses a genetic algorithm and a simplex hybrid global optimization algorithm to solve the precision surface matching problem. This method employs a forward formula to solve the objective, avoiding convergence failures caused by instability in partial derivative calculations.

[0007] The literature “Error calculation for corrective machining with allowance requirements, International Journal of Advanced Manufacturing Technology, 2010, 49: 635-641” ensures the local validity of the optimization variables and imposes constraints on them. In each iteration, a linear search and Palacios adjustment strategy are introduced to find a better point to reduce the value of the objective function, and this point is selected as the starting point for the next iteration.

[0008] The paper "Profile and thickness constrained adaptive localization for manufacturing curved thin-walled parts based on on-machine measurement, The International Journal of Advanced Manufacturing Technology, 2020, 110(1):113-123" uses an equidistant mapping method to create individual point pairs between the measurement points and the nominal shape of the parts. A constraint optimization algorithm for machining allowances is repeatedly executed until the constraints of the profile and thickness tolerance ranges are met.

[0009] However, the optimization and positioning of machining allowances for thin-walled aerospace parts is a hybrid containment and positioning problem. Current common machining allowance optimization models only perform global modeling and do not consider the differences between machined and unmachined surfaces. For thin-walled parts, it is necessary to consider not only the minimum allowance constraints of machined surfaces but also the containment problem of unmachined surfaces. Summary of the Invention

[0010] This invention provides a method for optimizing machining allowance in the processing of thin-walled parts. The aim is to complete workpiece positioning more efficiently, ensure uniform machining allowance, and prevent allowance shortage during the processing, thereby achieving rapid processing of thin-walled parts.

[0011] The technical solution adopted in this invention is as follows: This invention discloses a method for optimizing the machining allowance of thin-walled parts, comprising the following steps:

[0012] Step 1: Measure the point cloud of the thin-walled workpiece to be processed. Obtain the nominal workpiece point cloud through CAD / CAM and use coarse registration to obtain the initial Euclidean transformation parameters.

[0013] Step 2: Establish a machining allowance optimization model between the point cloud of the thin-walled workpiece to be processed and the point cloud of the nominal workpiece;

[0014] Step 3: Establish a constrained multi-objective optimization problem with the constraint that the excess metal on the surface to be processed is not negative, the nominal workpiece point cloud coincides with the non-processed surface of the workpiece point cloud, and the processing surface allowance of the workpiece point cloud is uniformly distributed.

[0015] Step 4: Solve the constrained multi-objective optimization problem, and finally calculate the Euclidean transformation parameters to realize the machining positioning and allowance optimization of thin-walled workpieces.

[0016] Furthermore, in step two, the constructed thin-walled part machining allowance optimization model is as follows:

[0017]

[0018]

[0019]

[0020] Where (R,T) represents the initial Euclidean transformation parameters; d i n This represents the directed distance between the workpiece point cloud and the corresponding points in the nominal workpiece point cloud.

[0021] d i m It can be represented as the machining allowance of the workpiece point cloud;

[0022] This is a constraint term that ensures that the machining allowance will not be insufficient during the machining process.

[0023] Furthermore, in step four, the MOPSO algorithm is used to calculate the thin-walled part machining allowance optimization model established in step two to obtain Euclidean transformation parameters, including the following steps.

[0024] S1. Initialize particle swarm parameters;

[0025] S2. Initialize the position and velocity of all particles;

[0026] S3. Determine if the iteration count has been reached;

[0027] S4. Calculate the memory term, self-cognition term, and group cognition term for each particle;

[0028] S5. Update the optimal fitness and position for each particle;

[0029] S6. Update the optimal fitness and position of the particle swarm;

[0030] S7. Repeat step S3; if not reached, repeat steps S4-S7;

[0031] Once the number of iterations is reached, the Euclidean transformation parameters are obtained.

[0032] Furthermore, in step three, regarding the accurate positioning problem of the non-machined surface of the thin-walled part, in order to make the nominal workpiece point cloud model coincide with the non-machined surface of the nominal workpiece point cloud, equation (4) is adopted as one of the optimization objectives:

[0033]

[0034] To ensure uniform machining allowance on the machined surfaces of thin-walled parts, equation (5) is used as an optimization objective:

[0035]

[0036] To avoid insufficient machining allowance for the workpiece point cloud, equation (6) is used as a constraint:

[0037]

[0038] This invention also discloses an evaluation method for optimizing machining allowances for thin-walled parts, which evaluates the final positioning results by setting evaluation indicators as shown in equation (7).

[0039]

[0040] Where e1 describes the error fluctuation at each point on the unmachined surface, and the smaller the value, the smaller the error on the unmachined surface; e2 describes the overall allowance distribution, indicating the average machining allowance of the workpiece; e3 describes the proportion of points with insufficient machining allowance in the point cloud; n represents the total number of points in the workpiece point cloud, and m represents the total number of points in the workpiece point cloud. The number of points is the nominal workpiece point cloud.

[0041] The beneficial effects of this invention are as follows:

[0042] 1. The thin-walled part machining allowance optimization model of this application obtains the point cloud model of the workpiece to be processed and the point cloud model of the nominal workpiece by measurement. The optimization and positioning of the machining allowance of the thin-walled workpiece is described as the non-machined surface of the workpiece to be processed and the nominal workpiece coincides. Under the premise of ensuring sufficient allowance, the machining allowance is evenly distributed.

[0043] Compared with existing machining allowance optimization methods, this method not only includes the inclusion problem of non-machined surfaces, but also the mixing problem of machined surfaces. It is suitable for machining thin-walled parts and can obtain better workpiece positioning results.

[0044] 2. This invention uses the MOPSO algorithm to solve the established constrained multi-objective optimization problem. Compared with the traditional ICP algorithm, it is simple in principle, has a small amount of computation, a fast convergence speed, and low requirements for computer hardware. In addition, the leap in MOPSO algorithm makes it easier to find the global optimum.

[0045] This invention provides reasonable evaluation indicators for the results after positioning. Compared with the overlap rate and other indicators of traditional point cloud registration algorithms, the evaluation indicators of this invention are more inclined to evaluate whether there is still insufficient machining allowance, overlap of non-machined surfaces, and uniformity of machining allowance after workpiece positioning. Therefore, the evaluation indicators are more reasonable than those of traditional point cloud algorithms, thereby guiding the machining and allowance optimization process of thin-walled parts. Attached Figure Description

[0046] Figure 1 A schematic diagram of the point cloud of the workpiece to be processed and the point cloud of the nominal workpiece;

[0047] Figure 2 In a specific embodiment of the present invention, d n Definition diagram;

[0048] Figure 3 In a specific embodiment of the present invention, d m Definition diagram;

[0049] Figure 4 This is a schematic diagram illustrating how the unmachined surface of the workpiece to be processed coincides with the unmachined surface of the nominal workpiece in a specific embodiment of the present invention;

[0050] Figure 5 This is a schematic diagram illustrating how the machining allowance on the machined surface is uniform in a specific embodiment of the present invention.

[0051] Figure 6 This is a flowchart of the MOPSO algorithm in a specific embodiment of the present invention;

[0052] Figure 7 This is a set of Pareto solutions obtained through iterative calculation in a specific embodiment of the present invention;

[0053] Figure 8 The final Euclidean transformed d obtained in a specific embodiment of the present invention n With d m The probability distribution. Detailed Implementation

[0054] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0055] Example 1

[0056] This embodiment selects an aircraft skin part, wherein the measurement point cloud is acquired using a Leica scanner. Figure 1 (a)), the nominal workpiece point cloud is obtained from the CAD model. Figure 1 (b)).

[0057] The coordinate system of the measured point cloud is defined as the measured point cloud coordinate system, and the coordinate system of the nominal workpiece point cloud is defined as the nominal workpiece point cloud coordinate system.

[0058] The method claimed in this invention can calculate the Euclidean transformation parameters between the measurement point cloud coordinate system and the nominal workpiece point cloud coordinate system, and finally calculate the optimized machining allowance, ensuring that the machining allowance is uniform and there is no shortage of allowance during workpiece machining.

[0059] Step 1 – Let coordinate system M be the measurement coordinate system of the workpiece point cloud, and coordinate system D be the model coordinate system of the nominal workpiece point cloud. Figure 2 ).

[0060] The position and orientation of coordinate system D relative to coordinate system M can be represented as g = (R, T) ∈ SE (3), where R ∈ SO (3) is the rotation matrix. is the translation vector, SE(3) is the special Euclidean group, and SO(3) is the characteristic orthogonal group.

[0061] The rotation matrix can be represented as R = R(α)R(β)R(γ), where α, β, and γ are the rotation angles about the X, Y, and Z axes, respectively, and can be expressed as:

[0062]

[0063]

[0064]

[0065] If p is the coordinate of a point in coordinate system D, then its coordinates in coordinate system M can be represented as gp.

[0066] If q is the coordinate of a point on the model surface in coordinate system M, then its coordinates in coordinate system D can be represented as g. -1 q.

[0067] The measured point cloud is recorded as The parameters of the nominal model surface are denoted as S(u,v).

[0068] Step 2 – Establishing a machining allowance optimization model for the workpiece point cloud and the nominal workpiece point cloud, specifically including:

[0069] Definition 1: Definition d n =(gp-q)gn q d n The direction is perpendicular to the curved surface S, such as Figure 2 As shown, where n q It is the unit outward normal vector at point q.

[0070] Definition 2: Let the boundary of surface S be denoted as The boundary of a point cloud is denoted as The directed distance function of the point-plane boundary is defined as follows: in m q It is the surface S and The boundary curve and tangent plane and

[0071] n q The surface S at point q * The unit normal vector, t q It is a boundary curve At point q * The unit tangent vector ( Figure 3 ).

[0072] Several criteria need to be met when optimizing the machining allowance of thin-walled parts.

[0073] First, it is necessary to satisfy the non-machined surface of the workpiece point cloud and the nominal workpiece point cloud. Alignment Ideally, the non-machined surfaces of the workpiece point cloud should coincide with the nominal non-machined surfaces of the workpiece point cloud. Figure 4 That is, for point P in the workpiece point cloud and point q corresponding to the nominal workpiece point cloud, the following conditions must be met:

[0074]

[0075] Therefore, in order to make the workpiece point cloud coincide as much as possible with the non-machined surface of the nominal workpiece point cloud, equation (4) is adopted as one of the optimization objectives:

[0076]

[0077] Secondly, the machining allowance must be sufficient. For the machining allowance of the workpiece point cloud, it should be guaranteed to be adequate. Let the machining allowance for each point be a. i ,Right now:

[0078] a i ≥0, (i=1,2,...,n) (5)

[0079] Secondly, a uniform machining allowance standard needs to be met. If the machining allowance is evenly distributed, the cutting force and elastic deformation of the machining system will be more uniform during machining. Figure 5 This protects the billet from damage caused by excessive vibration. Ideally, the machining allowance 'a' at each point... i The variance is 0, that is:

[0080]

[0081] When the non-machined surfaces coincide, the machining allowance a is then available. i It can be regarded as d i m .

[0082] In summary, the final optimized machining allowance model can be expressed as:

[0083]

[0084]

[0085]

[0086] Step 3 – Use the MOPSO algorithm to calculate the thin-walled part machining allowance optimization model established in Step 2. The flowchart is as follows: Figure 6 As shown. The pseudocode form is:

[0087]

[0088]

[0089] In this embodiment, the particle swarm size is set to 20, the inertia factor is set to w1 = 0.2, and the learning factor is set to c1 = c2 = 0.5. Through iterative calculation, a Pareto solution set containing 20 non-dominant solutions is obtained, as follows: Figure 7 As shown.

[0090] d i m and d i n The probability distributions are as follows: Figure 8 (a) and Figure 8 As shown in (b);

[0091] d is calculated based on its probability distribution. i n The distribution has a kurtosis of -0.3301, a skewness of 0.7722, and a mean of 0.04769 mm. i n More than 90% of the distribution falls within the range of -0.05 mm to 0.2 mm, indicating that the error of most points on the non-machined surfaces of the workpiece point cloud is less than 0.2 mm.

[0092] d i m The distribution has a kurtosis of 0.8576, a skewness of 1.3812, and a mean of 0.76525. Where d i m Approximately 3.357% of the distribution exceeds 1.5 mm, which corresponds to the top of the workpiece to be processed, indicating a larger machining allowance.

[0093] From this Pareto solution set, one solution is selected as the final solution, and the Euclidean transformation parameters are obtained as follows:

[0094]

[0095] Step 4 – Analyze the results of the optimized machining allowance using three evaluation indicators.

[0096]

[0097] Where e1 describes the error fluctuation at each point on the unmachined surface, the smaller the value, the smaller the error on the unmachined surface; e2 describes the overall allowance distribution, indicating the average machining allowance of the workpiece; e3 describes the proportion of points in the point cloud with insufficient machining allowance; n is the number of points in the point cloud; m is the number of points that satisfy... The number of points.

[0098] In this embodiment, e1 indicates that the average error per point on the non-machined surface is 0.157844 mm, e2 indicates that the average machining allowance of the workpiece is 0.765255 mm, and e3 indicates that the points that do not meet the allowance constraint account for 0.07609% of the total number of points in the point cloud. This shows that the average error on the non-machined surface is approximately 0.158 mm, while the workpiece meets the constraint of sufficient machining allowance.

[0099] <![CDATA[e1]]> <![CDATA[e2]]> <![CDATA[e3]]> 0.157844mm 0.765255mm 0.07609％

[0100] This application proposes a novel model to address this challenge by observing the characteristics of thin-walled parts. The proposed model optimizes machining allowances by imposing constraints to ensure a uniform distribution of machining allowances without causing shortages.

[0101] Those skilled in the art will readily understand that the above description is merely one embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for optimizing machining allowances for thin-walled parts, characterized in that, Includes the following steps: Step 1: Measure the point cloud of the thin-walled workpiece to be processed. Obtain the nominal workpiece point cloud through CAD / CAM and use coarse registration to obtain the initial Euclidean transformation parameters. Step 2: Establish a machining allowance optimization model for the workpiece point cloud and the nominal workpiece point cloud; Step 3: Establish a constrained multi-objective optimization problem with sufficient machining allowance as the constraint, the nominal workpiece point cloud and the non-machined surface of the workpiece point cloud coincide, and the machining allowance of the workpiece point cloud is uniformly distributed; sufficient machining allowance means that the excess metal on the surface to be machined is not negative. Step 4: Solve the constrained multi-objective optimization problem, and finally calculate the Euclidean transformation parameters to realize the machining positioning and allowance optimization of thin-walled workpieces; In step two, the constructed optimization model for machining allowance of thin-walled parts is as follows: ； Where (R,T) represents the initial Euclidean transformation parameters; This represents the directed distance between the workpiece point cloud and the corresponding point in the nominal workpiece point cloud; This is represented as the machining allowance of the workpiece point cloud; This is a constraint term that ensures that the machining allowance will not be insufficient during the machining process. In step three, to address the problem of accurate positioning of the unmachined surface of the thin-walled part, in order to make the nominal workpiece point cloud model coincide with the unmachined surface of the nominal workpiece point cloud, equation (4) is adopted as one of the optimization objectives: （4）； To ensure uniform machining allowance on the machined surfaces of thin-walled parts, equation (5) is used as an optimization objective: （5）； To avoid insufficient machining allowance for the point cloud of the workpiece, equation (6) is used as a constraint: （6）。 2. The method for optimizing machining allowances for thin-walled parts according to claim 1, characterized in that, In step four, the MOPSO algorithm is used to calculate the thin-walled part machining allowance optimization model established in step two to obtain Euclidean transformation parameters, including the following steps: S1. Initialize particle swarm parameters; S2. Initialize the position and velocity of all particles; S3. Determine if the iteration count has been reached; S4. Calculate the memory term, self-cognition term, and group cognition term for each particle; S5. Update the optimal fitness and position for each particle; S6. Update the optimal fitness and position of the particle swarm; S7. Repeat step S3; if not reached, repeat steps S4-S7; Once the number of iterations is reached, the Euclidean transformation parameters are obtained.

3. The method for optimizing machining allowances for thin-walled parts according to claim 1, characterized in that, It also includes a step to evaluate the final positioning results, which is done by setting the following evaluation indicators: (7) in, e 1 describes the error fluctuation at various points on the unmachined surface; the smaller the value, the smaller the error on the unmachined surface. e 2 describes the overall allowance distribution, indicating the average machining allowance of the workpiece; e 3 describes the proportion of points with insufficient processing margin in the point cloud; This indicates the number of points in the point cloud of the workpiece. Indicating the workpiece point cloud The number of dots.

Images