Method for acquiring pre-rotation angle of pipe flange after pipe bending in marine complex space

Abstract

The invention discloses a method for acquiring a pre-rotation angle of a pipe flange after pipe bending in a marine complex space. The method comprises the following steps: step 1, acquiring a total data set of a training model; 2, a springback value prediction model with a plurality of decision trees is constructed, and a springback angle is obtained according to results of the decision trees through an integrated voting method; the input parameters of the springback value prediction model comprise a bending angle, a relative bending radius, a pipe diameter and a wall thickness; training a springback value prediction model through the total data set in the step 1; 3, the rebound angle of each bent pipe is obtained through the trained rebound value prediction model, and the error value of the angles of the bent pipes is calculated; and 4, according to the minimum error value of the angle of each bent pipe, the flange pre-rotation angle is obtained through a trigonometric function method. According to the method, the pre-rotation angle of the pipe flange is accurately calculated, so that the flange mounting difficulty during subsequent pipe connection can be reduced, the operation efficiency is improved, and the time cost is reduced.

Description

Technical Field

[0001] The invention relates to the technical field of shipbuilding, and in particular to a method for obtaining a pre-rotation angle of a pipe flange after a ship-used complex space bend. Background Art

[0002] Pipe processing is a key link in shipbuilding, and its operation volume accounts for about three-fifths of the entire pipe assembly operation. At present, the production process is mostly organized in the form of "bending first and then welding", but this production process cannot achieve assembly line production, which greatly affects the quality and efficiency of pipe processing. Therefore, in order to improve the one-time assembly rate of pipe processing, whether "welding first and then bending" is the key.

[0003] At present, the main prerequisite for welding before bending is:

[0004] Calculation of flange assembly pre-rotation angle: The flange pre-rotation angle must be determined before flange welding to ensure that the head and tail flanges meet the assembly angle requirements after bending.

[0005] In order to improve the efficiency and quality of pipe processing, most companies have introduced automated flange assembly technology to ensure the quality and efficiency of flange welding. However, assembly line welding technology is only applicable to straight pipe welding and some specific angle bends. For pipes that require complex spatial bending, it is temporarily impossible to directly weld them because the deflection angles of the head and tail flanges cannot be accurately obtained, resulting in the inability to achieve "welding first and then bending".

[0006] The existing "Calculation and Verification Method of Flange Pre-rotation Angle of Pipe Welded First and Bent Later" determines the flange pre-rotation angle of pipe welded first and bent later through trigonometric function calculation method and normal vector calculation method, and summarizes the sample rod simulation verification method to verify the accuracy of the results after pipe bending. However, the calculation of this method does not take into account the influence of the error caused by the springback after the pipe is bent in complex space on the final flange pre-rotation angle result, so the calculated flange pre-rotation angle deviates from the actual production, which affects the subsequent flange connection and causes additional work such as flange correction. Summary of the invention

[0007] In view of the deficiencies in the prior art, the present invention provides a method for obtaining the flange pre-rotation angle of a marine complex space pipe after bending, so as to solve the technical problem that the flange pre-rotation angle cannot be accurately obtained in the prior art.

[0008] The present invention provides a method for obtaining a pre-rotation angle of a pipe flange after a marine complex space bending pipe, comprising the following steps:

[0009] Step 1: Get the total dataset for training the model;

[0010] Step 2: Construct a springback prediction model with several decision trees, wherein the results of several decision trees are used to obtain the springback angle through an integrated voting method; the input parameters of the springback prediction model include: bending angle, relative bending radius, pipe diameter, and wall thickness; the springback prediction model is trained through the total data set of step 1;

[0011] Step 3: Obtain the springback angle of each pipe bend through the trained springback prediction model, and calculate the error value of the pipe bend angle;

[0012] Step 4: According to the minimum error value of each elbow angle, the flange pre-rotation angle is obtained by trigonometric method.

[0013] Furthermore, in step 2, the number of decision trees is set in the range of 55 to 65; the types of the decision trees are one or more types of decision trees.

[0014] Furthermore, the node division criterion of the decision tree is the mean absolute error.

[0015] Furthermore, in step 3, the specific steps of calculating the error value of the bending angle are:

[0016] Step 31: input the collected data into the springback prediction model to obtain the springback angle;

[0017] Step 32: Calculate the predicted bending angle according to the input bending angle and springback angle;

[0018] Step 33: Calculate the difference between the required bending angle and the predicted bending angle. When the difference is less than the standard value of the bending angle error, use the current difference as the error value of the bending angle and execute subsequent steps. When the difference is greater than or equal to the standard value of the bending angle error, return to step 31 and optimize the input bending angle.

[0019] Furthermore, the specific method for optimizing the input bending angle is:

[0020] The sum of the required bending angle and the springback angle is taken as the bending angle.

[0021] Furthermore, in step 4, the calculation formula of the flange pre-rotation angle is:

[0022] α=β1+β2+arccos(ψ1)+arccos(ψ2)+arccos(ψ3)+……+arccos(ψ n )

[0023] Where α is the flange pre-rotation angle; β1 is the initial angle when the elbow is bent; β2 is the tail angle of the elbow; ψ is the angle between the planes where adjacent bending angles are located.

[0024] Furthermore, the angle between the planes where the adjacent bending angles are located is calculated as follows:

[0025]

[0026] In the formula, the normal vector Among them, a1 is the plane curvature angle, y1 and z1 are the coordinates of a point in the plane where the plane curvature angle a1 is located; the normal vector Among them, a2 is the plane curvature angle, y2 and z2 are the coordinates of a point in the plane where the plane curvature angle a2 is located, the plane where the plane curvature angle a1 is located and the plane where the plane curvature angle a2 is located are the planes where the adjacent bending angles are located; r is the normal vector The plane curvature angle is the sum of the error values ​​of the corresponding required bending angle and the bending angle.

[0027] Beneficial effects of the present invention:

[0028] The present invention can reduce the difficulty of flange installation during subsequent pipe connection by accurately calculating the pre-rotation angle of the pipe flange, improve work efficiency, and reduce time cost. The present invention can realize "welding before bending" for pipes that could only be "bent before welding" due to large pre-rotation angle errors of flanges after complex space bending, greatly improving welding quality and welding efficiency. The present invention can reduce the need for additional production of closed pipes for installation during subsequent connection, reducing material costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The features and advantages of the present invention will be more clearly understood by referring to the accompanying drawings, which are schematic and should not be construed as limiting the present invention in any way. In the accompanying drawings:

[0030] Figure 1 is a flow chart of a specific embodiment of the present invention;

[0031] Figure 2 is a schematic diagram of a springback prediction model in a specific embodiment of the present invention;

[0032] Figure 3 1 is a schematic diagram of a flow chart of calculating the error value of the bend angle in a specific embodiment of the present invention;

[0033] Figure 4 is a schematic diagram of the flange pre-rotation angle in a specific embodiment of the present invention;

[0034] Figure 5 Schematic diagram of the initial angle and the tail angle of a specific embodiment of the present invention. DETAILED DESCRIPTION

[0035] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0036] The present invention is further illustrated below in conjunction with specific embodiments. Those skilled in the art should understand that these embodiments are only used to illustrate the present invention and are not used to limit the scope of the present invention, and modifications to various equivalent forms of the present invention fall within the scope defined by the appended claims of this application.

[0037] like Figure 1 As shown, the present invention provides a method for obtaining the pre-rotation angle of the pipe flange after a marine complex space bend, comprising the following steps:

[0038] Step A1: Get the total dataset for training the model;

[0039] Select steel pipes, determine basic data such as material, diameter, wall thickness, and determine the pipe bending machine and flange model. Perform 30-90 degree bending tests on pipes of the same material, diameter, and wall thickness. Bend the pipes three times at the same angle, record the rebound angle and other parameters, and prepare array sample data for model training;

[0040] like Figure 2 As shown, step A2: construct a springback prediction model with several decision trees, wherein the results of several decision trees are used to obtain the springback angle through an integrated voting method; the input parameters of the springback prediction model include: bending angle, relative bending radius, pipe diameter, and wall thickness; the springback prediction model is trained through the total data set of step 1; the specific process is:

[0041] Step A21: Use bootstrap sampling to randomly select n sample sets from the original data set with replacement as training sets, perform h sampling times in total, and obtain h training sets. At the same time, introduce randomness when selecting the partitioning attributes, that is, randomly select a part of the features for node splitting, and the node partitioning criterion of the decision tree is the mean absolute error;

[0042] Step A22: Each of the h training sets is used to train a base learner, and the base learner is fixed as a decision tree, that is, H decision trees are generated;

[0043] Step A23: Use the generated decision tree to predict the prediction data, summarize the prediction results of all trees, and obtain the final prediction through integrated voting, in which the abnormal predictions that may be generated by individual decision trees are smoothed out to improve the overall prediction accuracy.

[0044] The root mean square error (RMSE) is used to measure the prediction effect of the training model. RMSE represents the difference between the predicted rebound amount of the model and the actual rebound amount. The specific formula is as follows:

[0045]

[0046] Where m is the number of instances in the evaluated data set; H is the prediction function, that is, the prediction model of the three regression variables; the data set is regarded as a matrix X, then X (i) is the total input of the i-th instance in the dataset, y (i) is the label of the data.

[0047] Step A3: Obtain the springback angle of each pipe bend through the trained springback prediction model, and calculate the error value of the pipe bend angle;

[0048] The specific steps for calculating the error value of the bending angle are:

[0049] Step A31: Input the collected data: bending angle θ, relative bending radius R, pipe diameter D, wall thickness t into the springback prediction model to obtain the springback angle Δθ i , where i is the number of times returned from step A33, initially 0;

[0050] Step A32: According to the input bending angle θ and springback angle Δθ i Calculate the predicted bending angle θ′ using the following formula:

[0051] θv=θ-Δθi

[0052] Step A33: Calculate the difference ф0 between the required bending angle θ0 and the predicted bending angle θ′. When the difference ф0 is less than the standard value ф of the bending angle error, use the current difference ф0 as the error value ф′ of the bending angle and execute the subsequent steps. When the difference ф0 is greater than or equal to less than the standard value ф of the bending angle error, return to step A31 and optimize the input bending angle. At this time, i is increased by 1. After returning to step A31, since the bending angle θ is optimized, the springback prediction model will output a new springback angle Δθ i , and then go through the subsequent process until the difference ф0 in step A33 is less than the standard value ф of the bending angle error.

[0053] The specific method for optimizing the input bending angle is:

[0054] The required bending angle θ0 and the springback angle Δθ i The sum of is taken as the bending angle θ.

[0055] Step A4: According to the minimum error value of each bend angle, determine the pre-rotation angle of the "weld first, bend later" flange by using the trigonometric function calculation method and the normal vector calculation method. Figure 5 As shown in the figure, the flange pre-rotation angle refers to the angle α between the head flange (the parallel plane of the double hole horizontal line through the center of the head flange) and the tail flange (the parallel plane of the double hole horizontal line through the center of the tail flange) when the pipe is welded in a straight pipe.

[0056] like Figure 5 As shown, assuming that point A(x1,y1,z1) is in the plane where the plane angle a1 is located, the corresponding normal vector is Assume that point B(x2,y2,z2) is in the plane where the plane angle a2 is located, and the corresponding normal vector is Among them, the plane bending angle = the required bending angle θ0 in the same plane + the minimum error φ′ of the angle.

[0057] Let ψ be the angle between the planes where the adjacent bending angles are located, such as Figure 4 The angle between the midplane P1P2P3 and the plane P4P3P2 is the rotation angle ψ1. Similarly, the angle between the plane curvature a3 and the plane curvature a2 is the rotation angle ψ2. n Angle a with plane n- The angle between the planes is the rotation angle ψ n , r is the normal vector The angle between , so we can get:

[0058] The calculation formula of the plane angle between the plane angles is:

[0059]

[0060] Flange pre-rotation angle calculation formula:

[0061] α=β1+β2+arccos(ψ1)+arccos(ψ2)+arccos(ψ3)+……+arccos(ψ n )

[0062] Where α is the flange pre-rotation angle; β1 is the initial angle when the pipe is bent.

[0063] like Figure 4As shown, first, make a line a′d′ parallel to the double-hole horizontal line ab on the head end flange through the center of the head end flange, and make a line c′d′ parallel to the double-hole horizontal line cd on the tail end flange through the center of the tail end flange, and then make a plane perpendicular to the head end flange through c′d′, and the intersection of the head end flange plane is c″d″″, that is, the c″′d″″ dotted line on the head end flange represents the projection of the tail end flange c″d′ on the head end flange, where c″d″ intersects with a′b′ at point o, and ∠c″oa′, that is, “α”, is the flange pre-rotation angle.

[0064] like Figure 5 As shown, the angle between the horizontal plane a′b′P1P2 where the parallel line a′b′ of the positive horizontal line ab of the double holes of the head end flange is located and the plane P1P2P3 where the curved angle α1 is located is the initial angle β1 of the head end flange. The angle between the plane c′d′P3P4 where the parallel line c′d′ of the positive horizontal line cd of the double holes of the tail end flange is located and the plane P2P3P4 where the curved angle α2 is located is the tail end angle β2 of the tail end flange. The initial angle β1 and the tail end angle β2 can be calculated by the normal vector method.

[0065] Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A method for obtaining the pre-rotation angle of the pipe flange after a marine complex space bend, characterized in that: The steps include: Step 1: Get the total dataset for training the model; Step 2: Construct a springback prediction model with several decision trees, wherein the results of several decision trees are used to obtain the springback angle through an integrated voting method; the input parameters of the springback prediction model include: bending angle, relative bending radius, pipe diameter, and wall thickness; the springback prediction model is trained through the total data set of step 1; Step 3: Obtain the springback angle of each pipe bend through the trained springback prediction model, and calculate the error value of the pipe bend angle; Step 4: According to the minimum error value of each elbow angle, the flange pre-rotation angle is obtained by trigonometric method.

2. The method for obtaining the pre-rotation angle of the pipe flange after the ship's complex space bending according to claim 1 is characterized in that: In step 2, the number of decision trees is set in a range of 55 to 65; the types of the decision trees are one or more types of decision trees.

3. The method for obtaining the pre-rotation angle of the pipe flange after the ship's complex space bending according to claim 1 or 2, characterized in that: The node division standard of the decision tree is the mean absolute error.

4. The method for obtaining the pre-rotation angle of the pipe flange after the ship's complex space bending as claimed in claim 1 is characterized in that: In step 3, the specific steps of calculating the error value of the bending angle are: Step 31: input the collected data into the springback prediction model to obtain the springback angle; Step 32: Calculate the predicted bending angle according to the input bending angle and springback angle; Step 33: Calculate the difference between the required bending angle and the predicted bending angle. When the difference is less than the standard value of the bending angle error, use the current difference as the error value of the bending angle and execute subsequent steps. When the difference is greater than or equal to the standard value of the bending angle error, return to step 31 and optimize the input bending angle.

5. The method for obtaining the pre-rotation angle of the pipe flange after the ship's complex space bending as claimed in claim 1 is characterized in that: The specific method for optimizing the input bending angle is: The sum of the required bending angle and the springback angle is taken as the bending angle.

6. The method for obtaining the pre-rotation angle of the pipe flange after the ship's complex space bending as claimed in claim 1 is characterized in that: In step 4, the calculation formula of the flange pre-rotation angle is: α=β1+β2+arccos(ψ1)+arccos(ψ2)+arccos(ψ3)+……+arccos(ψ n ) Where, α is the flange pre-rotation angle; β1 is the initial angle of the elbow; β2 is the tail angle of the elbow; ψ is the angle between the planes where adjacent bending angles are located.

7. The method for obtaining the pre-rotation angle of the pipe flange after the ship's complex space bending as claimed in claim 6 is characterized in that: The calculation formula for the angle between the planes where the adjacent bending angles are located is: In the formula, the normal vector Among them, a1 is the plane curvature angle, y1 and z1 are the coordinates of a point in the plane where the plane curvature angle a1 is located; the normal vector Among them, a2 is the plane curvature angle, y2 and z2 are the coordinates of a point in the plane where the plane curvature angle a2 is located, the plane where the plane curvature angle a1 is located and the plane where the plane curvature angle a2 is located are the planes where the adjacent bending angles are located; r is the normal vector The plane bending angle is the sum of the error values ​​of the corresponding required bending angle and the bending angle.

Images