Complex equipment response-oriented hybrid uncertainty propagation method, system, equipment and medium

The hybrid uncertainty propagation method addresses mixed uncertainty in engineering systems by iteratively finding CDF boundaries using sequential simulation and UDRM, enhancing efficiency and accuracy in complex equipment response analysis.

CN119089682AActive Publication Date: 2024-12-06NANJING INST OF TECH
View PDF 6 Cites 0 Cited by

Patent Information

Application Number
CN202411190759.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-28
Publication Date
2024-12-06
Estimated Expiration
2044-08-28

AI Technical Summary

Technical Problem

Existing methods lack efficient strategies for handling mixed uncertainty propagation problems in engineering systems where both random variables and interval numbers coexist, particularly in complex equipment response analysis.

Method used

A hybrid uncertainty propagation method using sequential simulation (SS) and univariate dimensionality reduction integral (UDRM) to iteratively find CDF boundaries, converting high-dimensional problems into moment analysis, and fitting statistical moments with Johnson distributions to enhance accuracy with fewer sample points.

Benefits of technology

Improves calculation efficiency and accuracy in mixed uncertainty propagation by reducing the number of required sample points while maintaining high precision, applicable to complex engineering problems with high nonlinearity and multiple input parameters.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119089682A_ABST
    Figure CN119089682A_ABST
Patent Text Reader

Abstract

The invention discloses a complex equipment response-oriented hybrid uncertainty propagation method, system and device and a medium, and the method comprises the steps: finding a sample which determines a CDF boundary in an iteration mode through a sequence simulation strategy, so as to gradually approach the upper and lower boundaries of a structural response P-box; in the sequence solving process, in each step of CDF solving, a high-dimensional probability propagation problem is converted into a moment analysis problem of a univariate subsystem through a univariate dimension reduction integral method. And after the interval of the system response statistical moment is obtained, fitting the response statistical moment in the interval range through Johnson fitting distribution to obtain the CDF. Compared with a direct simulation method, the sequence simulation strategy disclosed by the invention has the advantages that the response probability box analysis of the complex equipment is more effectively carried out, and under the condition of ensuring the same calculation precision, the relatively accurate system response is obtained by only using less sample point information. According to the method, the sampling efficiency problem in a high-dimensional problem is solved by combining the advantages based on interval sequence simulation and statistical moment analysis.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the field of uncertainty propagation analysis methods, and in particular relates to a hybrid uncertainty propagation method, system, device and medium for complex equipment response. Background Art

[0002] With the development of science and technology and the advancement of industrial technology, engineering technicians have higher and higher requirements for the reliability of equipment. However, there are a lot of uncertainties in engineering equipment, such as load, size, material, processing errors, etc., which have brought great troubles to the improvement of the reliability of engineering equipment. Therefore, it is urgent to carry out propagation analysis of the uncertainty of engineering equipment to lay the foundation for improving the reliability of equipment.

[0003] According to the different cognitive levels and the amount of parameter information, uncertainty is usually divided into random uncertainty and epistemic uncertainty. Random uncertainty is also called objective uncertainty, and relevant scholars have conducted long-term and extensive research. Among them, for the problem of probabilistic uncertainty propagation with sufficient parameter information, the probability density function and cumulative distribution function in probability theory can be used to quantify and propagate random uncertainty. For the probability propagation model, a series of uncertainty propagation methods have been generated, including Monte Carlo simulation method, uncertainty propagation method based on perturbation theory, uncertainty propagation method based on sequence reliability, uncertainty propagation method based on function expansion, and uncertainty propagation method based on moment propagation. For the problem of epistemic uncertainty, related uncertainty propagation methods have also emerged, including interval model, evidence theory, fuzzy theory, and possibility theory. Due to the small amount of relevant parameter sample information in engineering practice, the interval model has been widely used. This has led to uncertainty propagation methods based on interval analysis, uncertainty propagation methods based on interval perturbation, uncertainty propagation methods based on interval sampling, uncertainty propagation methods based on interval optimization and proxy models, and uncertainty propagation methods based on interval sequence simulation.

[0004] According to the above analysis, for random uncertainty and epistemic uncertainty problems, probability model and interval model are two common measurement methods. However, in actual engineering, due to the different ways of obtaining input parameters, measurement difficulty, and measurement methods, mixed uncertainty problems in which random variables and interval numbers exist simultaneously in the system structure are often encountered. At this time, the uncertainty propagation problem is transformed into a probability-interval mixed uncertainty propagation problem. According to the location where the interval variables appear, the probability-interval mixed uncertainty propagation is divided into a separated mixed uncertainty propagation and a coupled mixed uncertainty propagation. The present invention mainly studies the separated mixed uncertainty propagation problem. Many uncertainty propagation methods have emerged for the separated mixed uncertainty problem, including the double-layer nested sampling (DLS) uncertainty propagation method evolved from probability sampling and interval sampling methods, and the mixed uncertainty propagation method based on optimization. At present, there is still a lack of relatively efficient strategies for the separated probability-interval mixed uncertainty propagation problem, which has become a problem to be solved in dealing with mixed uncertainty propagation of engineering problems. Summary of the invention

[0005] The present invention aims at the deficiencies in the prior art and provides a hybrid uncertainty propagation method, system, device and medium for complex equipment response. Through the sequential simulation (SS) strategy, the samples that determine the boundaries of the cumulative distribution function (CDF) are found in an iterative manner, so as to gradually approach the upper and lower boundaries of the structural response probability box (P-box). In the sequence solution process, for each step of the CDF solution, the present invention converts the high-dimensional probability propagation problem into a moment analysis problem of a univariate subsystem through the univariate dimensionality reduction integral method (UDRM). After obtaining the interval of the statistical moment of the system response, the statistical moment of the response is fitted within its interval range to obtain the CDF through the Johnson fitting distribution. Compared with the existing methods, under the condition of ensuring the same calculation accuracy, only less sample point information is used to obtain a more accurate system response. The present invention combines the advantages of interval sequence simulation and statistical moment analysis, and solves the sampling efficiency problem in high-dimensional problems through a high-precision approximation strategy.

[0006] To achieve the above object, the present invention adopts the following technical solutions:

[0007] A hybrid uncertainty propagation method for complex equipment response is applied to a coordination mechanism, wherein the coordination mechanism is a rigid-flexible coupling system, including a motor, a reducer, a coordination manipulator, a posture adjustment electric cylinder assembly, and a conveying mechanism. The method includes the following steps:

[0008] Step 1, establishing a rigid-flexible coupling dynamic model of the coordination mechanism; the input parameters of the rigid-flexible coupling dynamic model of the coordination mechanism include probability uncertainty parameters and interval uncertainty parameters;

[0009] Step 2: Characterize the uncertainty of the output response of the rigid-flexible coupling dynamic model of the coordination mechanism based on P-box;

[0010] Step 3: According to the distribution type of the input parameters, the input parameters of the rigid-flexible coupling dynamics model are layered, with the outer layer being interval uncertainty parameters and the inner layer being probabilistic uncertainty parameters;

[0011] Step 4: Select the interval uncertainty parameters and sample space of the rigid-flexible coupling dynamics model, perform global sampling and local sampling in the sample space, and obtain multiple groups of interval samples, including global samples and local samples;

[0012] Step 5: Randomly select a group of interval samples in the outer layer, use the dimension reduction integral method to simulate and calculate the probability uncertainty parameters of the inner layer, and obtain the P-box response CDF; determine whether all interval samples in the outer layer have been calculated. If so, a series of CDF results are obtained and go to step 6, otherwise continue to step 5;

[0013] Step 6: Based on the series of CDF results obtained in step 5, obtain the lower boundary of the P-box calculated this time, determine the contribution CDF composition corresponding to the lower boundary, find the contribution sample corresponding to the contribution CDF, and complete a sequence simulation;

[0014] Step 7: Determine whether the area difference between the lower bounds of the CDF obtained from the two most recent sequence simulations meets the preset standard. If so, it is considered that the calculation has reached convergence and go to step 8; otherwise, go to step 4 to generate new global samples and local samples;

[0015] Step 8: Output the final P-box response CDF lower boundary;

[0016] Step 9: Perform sequence simulation to obtain the upper boundary of the P-box response CDF. Based on the upper and lower boundaries of the P-box response CDF, obtain the P-box result of the mixed uncertainty propagation of the coordination mechanism, and adjust the positioning accuracy of the coordination robot arm according to the actual working conditions.

[0017] To optimize the above technical solutions, the specific measures taken also include:

[0018] Furthermore, in step 1, the rigid-flexible coupling dynamic model of the coordination mechanism is specifically:

[0019]

[0020] Where M(q,t) is the mass matrix of the coordination mechanism, t is the motion time of the coordination process, q, and are the relative displacement, relative velocity and relative acceleration of the rotary joint in the coordination mechanism, is the generalized damping matrix of the coordination mechanism, is the generalized stiffness matrix of the coordination mechanism, Φ(q,t)=0 is the constraint equation of the coordination mechanism, Φ q Derivative of Φ(q,t) with respect to q, (g) T represents the transpose of the matrix, l is the Lagrange multiplier, and Q is the external force of the coordination mechanism.

[0021] Furthermore, in step 2, the output response of the rigid-flexible coupling dynamic model of the coordination mechanism is the angular displacement angle of rotating the coordination manipulator from the initial position of -3° to the position with an angle of 45° with the horizontal direction, and the CDF of the output response is expressed by F z (Z) indicates that F z (Z) is enclosed by the lower boundary of the CDF F z (Z) and upper boundary In the area formed, The P-box output response Z of the rigid-flexible coupling dynamic model of the coordinated mechanism is written as:

[0022]

[0023] In the formula, The P-box that responds to Z.

[0024] Further, in step 4, the global sample is generated from the entire bounded sample space, and the local sample is generated in the neighborhood of the contributing sample according to the contribution index of the contributing sample; the contributing sample refers to the sample that contributes to the upper and lower boundaries of the P-box;

[0025] The contribution index is calculated as follows:

[0026] For each contribution sample The CDF of the P-box is recorded as Its contribution is evaluated by the ordinate span where the CDF takes its maximum value:

[0027]

[0028] In the formula, It is a contribution sample The contribution index of Indicates the contribution sample The vertical axis span that affects the upper or lower boundary of the CDF, F z (Z(y)) represents the total ordinate span of the upper and lower boundaries of the P-box.

[0029] Furthermore, the neighborhood of the contributing sample is centered at The radius is r * The circular area or multi-dimensional spherical area of ​​; the radius is calculated as follows:

[0030]

[0031] Where b represents the neighborhood relative to the entire bounded sample space W α The size of V(W α ) is a bounded sample space W α The volume or area of ​​, m is the number of interval uncertainty parameters, and it is also the dimension of the bounded sample space.

[0032] Furthermore, the neighborhood of the contributing sample is centered at The side length is 2r * A square or multidimensional cubic region; r * The calculation formula is as follows:

[0033]

[0034] Where b represents the neighborhood relative to the entire bounded sample space W α The size of V(W α ) is a bounded sample space W α of

[0035] volume or area, m is the number of interval uncertainty parameters and also the dimension of the bounded sample space.

[0036] Furthermore, the neighborhood of the contributing sample is centered at The side length is 2r i * ,i=1,2,L,m rectangular or multidimensional rectangular solid area, m is the number of interval uncertainty parameters, r i * The calculation formula is as follows:

[0037]

[0038] Where b represents the neighborhood relative to the entire bounded sample space W α The size of For interval Width.

[0039] Further, the value of b is set between 1 / 12 and 1 / 4.

[0040] Furthermore, in step 5, the inner layer probability uncertainty parameter is simulated and calculated using the dimension reduction integration method to obtain the P-box response CDF as follows:

[0041] Step 5.1: Use the single variable dimension reduction integral method to reduce the dimension near the mean value of the probability uncertainty parameter, and transform the multivariable probability uncertainty parameter system into a subsystem containing only a single variable, which can be expressed as:

[0042]

[0043] Among them, E l (Z) represents the lth order moment of the response, l = 1, 2, 3, 4; m i is the mean of the ith probability uncertainty parameter, yes It is an abbreviation of the decomposed univariate subsystem X. i represents the i-th probability uncertainty parameter, Y (d) represents the interval uncertainty parameter, is the binomial expansion, j is the order number, Represents the function Find the jth order moment, m represents the number of probability uncertainty parameters, g0(m1,m2,L,m m ,Y (d) ) indicates that all probability uncertainty parameters are initialized to the mean;

[0044] Step 5.2: Use Gauss-Hermite integration to calculate the first four moments of the univariate subsystem, expressed as:

[0045]

[0046] In the formula, E{·} represents the function The l-th order moment, r represents the number of Gaussian integration nodes, U i,k and w k are Gaussian integration nodes and node weights respectively;

[0047] Step 5.3, the first four-order center moments are calculated iteratively by coordinating the first four-order origin moments of the output response of the rigid-flexible coupling dynamic model of the mechanism, which can be expressed as:

[0048]

[0049] Among them, E l (Y (d)), l = 1, 2, 3, 4 are the first four origin moments of the output response of the rigid-flexible coupling dynamic model of the coordination mechanism, M l (Y (d) ), l = 1, 2, 3, 4 are the first four central moments of the output response of the rigid-flexible coupling dynamic model of the coordination mechanism;

[0050] Step 5.4: Use the Johnson fitting method to obtain the P-box response CDF of the first four-order central moments of the output response of the rigid-flexible coupling dynamic model of the coordination mechanism. The Johnson fitting method is expressed as:

[0051]

[0052] Among them, g and δ are shape parameters, is the size parameter, x is the position parameter, w~N(0,1) obeys the standard normal distribution, Y is the transformation function, w represents the Johnson distribution, and h is the Johnson variable of the output response of the rigid-flexible coupling dynamic model of the coordination mechanism.

[0053] Further, in step 7, the determination of whether the area difference between the CDF lower bounds obtained from the two most recent sequence simulations meets the preset standard is specifically:

[0054] Will As the lower bound of the CDF obtained by the k-th sequence simulation, As the CDF lower bound obtained by the k-1th sequence simulation, the area difference between the CDF lower bounds obtained by two adjacent sequence simulations is used as the convergence judgment criterion, which is as follows:

[0055]

[0056] in, is the CDF lower bound obtained by the k-th sequence simulation and the CDF lower bound obtained by the k-1th sequence simulation The Minkowski-L1 distance between Indicates the contribution sample The differential of

[0057] Minkowski-L1 distance as the absolute error of the k-th sequence simulation The convergence criterion is:

[0058]

[0059] Where e0 is the preset value of error control.

[0060] Furthermore, in step 5.2, the number r of the Gaussian integration nodes is 6.

[0061] Furthermore, the preset value of the error control is e0=1′10 -3 .

[0062] The present invention also proposes a hybrid uncertainty propagation system for complex equipment response, which is applied to a coordination mechanism. The coordination mechanism is a rigid-flexible coupling system, including a motor, a reducer, a coordination manipulator, a posture adjustment electric cylinder assembly and a conveying mechanism. The hybrid uncertainty propagation system includes:

[0063] A modeling module is used to establish a rigid-flexible coupling dynamics model of a coordination mechanism; the input parameters of the rigid-flexible coupling dynamics model of a coordination mechanism include a probability uncertainty parameter and an interval uncertainty parameter;

[0064] Uncertainty characterization module, used to characterize the uncertainty of the output response of the rigid-flexible coupling dynamics model of the coordination mechanism based on P-box;

[0065] A stratification module is used to stratify the input parameters of the rigid-flexible coupling dynamics model according to the distribution type of the input parameters, with the outer layer being interval uncertainty parameters and the inner layer being probabilistic uncertainty parameters;

[0066] A sampling module is used to select interval uncertainty parameters and sample space of the rigid-flexible coupling dynamics model, perform global sampling and local sampling in the sample space, and obtain multiple groups of interval samples, including global samples and local samples;

[0067] The lower boundary generation module is used to select any group of interval samples in the outer layer, simulate and calculate the inner layer probability uncertainty parameters using the dimensionality reduction integral method, and obtain the P-box response CDF; determine whether all the interval samples in the outer layer have been calculated, if so, obtain a series of CDF results, otherwise continue to select the interval samples that have not been calculated for simulation calculation to obtain the P-box response CDF until all the interval samples in the outer layer are calculated; based on a series of CDF results, obtain the P-box lower boundary of this calculation, and determine the contribution CDF composition corresponding to the lower boundary, find the contribution samples corresponding to the contribution CDF, and complete a sequence simulation; determine whether the area difference between the CDF lower bounds obtained from the most recent two sequence simulations meets the preset standard, if so, it is considered that the calculation has reached convergence, and the final P-box response CDF lower bound is output; otherwise, control the sampling module to generate new global samples and local samples;

[0068] The upper boundary generation module is used to perform sequence simulation to obtain the upper boundary of the P-box response CDF;

[0069] The adjustment module is used to obtain the P-box result of the mixed uncertainty propagation of the coordination mechanism based on the upper and lower boundaries of the P-box response CDF, and adjust the positioning accuracy of the coordination robot arm according to the actual working conditions.

[0070] To optimize the above technical solutions, the specific measures taken also include:

[0071] Furthermore, the rigid-flexible coupling dynamic model of the coordination mechanism is specifically:

[0072]

[0073] Where M(q,t) is the mass matrix of the coordination mechanism, t is the motion time of the coordination process, q, and are the relative displacement, relative velocity and relative acceleration of the rotary joint in the coordination mechanism, is the generalized damping matrix of the coordination mechanism, is the generalized stiffness matrix of the coordination mechanism, Φ(q,t)=0 is the constraint equation of the coordination mechanism, Φ q Derivative of Φ(q,t) with respect to q, (g) T represents the transpose of the matrix, l is the Lagrange multiplier, and Q is the external force of the coordination mechanism.

[0074] Furthermore, the output response of the rigid-flexible coupling dynamic model of the coordination mechanism is the angular displacement angle of rotating the coordination manipulator from the initial position of -3° to the position with an angle of 45° with the horizontal direction, and the CDF of the output response is expressed as F z (Z) indicates that F z (Z) is enclosed by the lower boundary of the CDF F z (Z) and upper boundary In the area formed, The P-box output response Z of the rigid-flexible coupling dynamic model of the coordinated mechanism is written as:

[0075]

[0076] In the formula, The P-box that responds to Z.

[0077] Furthermore, the global samples are generated from the entire bounded sample space, and the local samples are generated in the neighborhood of the contributing samples according to the contribution index of the contributing samples; the contributing samples refer to the samples that contribute to the upper and lower boundaries of the P-box;

[0078] The contribution index is calculated as follows:

[0079] For each contribution sample The CDF of the P-box is recorded as Its contribution is evaluated by the ordinate span where the CDF takes its maximum value:

[0080]

[0081] In the formula, It is a contribution sample The contribution index of Indicates the contribution sample The vertical axis span that affects the upper or lower boundary of the CDF, F z (Z(y)) represents the total ordinate span of the upper and lower boundaries of the P-box.

[0082] Furthermore, the neighborhood of the contributing sample is centered at The radius is r * The circular area or multi-dimensional spherical area of ​​; the radius is calculated as follows:

[0083]

[0084] Where b represents the neighborhood relative to the entire bounded sample space W α The size of V(W α ) is a bounded sample space W α The volume or area of ​​, m is the number of interval uncertainty parameters, and it is also the dimension of the bounded sample space.

[0085] Furthermore, the neighborhood of the contributing sample is centered at The side length is 2r * A square or multidimensional cubic region; r * The calculation formula is as follows:

[0086]

[0087] Where b represents the neighborhood relative to the entire bounded sample space W α The size of V(W α ) is a bounded sample space W α of

[0088] volume or area, m is the number of interval uncertainty parameters and also the dimension of the bounded sample space.

[0089] Furthermore, the neighborhood of the contributing sample is centered at The side length is 2r i * ,i=1,2,L,m rectangular or multidimensional rectangular solid area, m is the number of interval uncertainty parameters, r i * The calculation formula is as follows:

[0090]

[0091] Where b represents the neighborhood relative to the entire bounded sample space W α The size of For interval Width.

[0092] Further, the value of b is set between 1 / 12 and 1 / 4.

[0093] Furthermore, the inner layer probability uncertainty parameter is simulated and calculated using the dimension reduction integration method to obtain the P-box response CDF as follows:

[0094] The single variable dimension reduction integral method is used for the probability uncertainty parameter, and the dimension reduction is expanded near the mean, and the multivariable probability uncertainty parameter system is transformed into a subsystem containing only a single variable, which is expressed as:

[0095]

[0096] Among them, E l (Z) represents the lth order moment of the response, l = 1, 2, 3, 4; m i is the mean of the ith probability uncertainty parameter, yes It is an abbreviation of the decomposed univariate subsystem X. i represents the i-th probability uncertainty parameter, Y (d) represents the interval uncertainty parameter, is the binomial expansion, j is the order number, Represents the function Find the jth order moment, m represents the number of probability uncertainty parameters, g0(m1,m2,L,m m ,Y (d) ) indicates that all probability uncertainty parameters are initialized to the mean;

[0097] The Gauss-Hermite integral is used to calculate the first four moments of the univariate subsystem, expressed as:

[0098]

[0099] In the formula, E{·} represents the function The l-th order moment, r represents the number of Gaussian integration nodes, U i,k and w k are Gaussian integration nodes and node weights respectively;

[0100] The first four-order central moments are calculated iteratively by coordinating the first four-order origin moments of the output response of the rigid-flexible coupling dynamic model of the mechanism, and are expressed as:

[0101]

[0102] Among them, E l (Y (d) ), l = 1, 2, 3, 4 are the first four origin moments of the output response of the rigid-flexible coupling dynamic model of the coordination mechanism, M l (Y (d) ), l = 1, 2, 3, 4 are the first four central moments of the output response of the rigid-flexible coupling dynamic model of the coordination mechanism;

[0103] The first four-order central moments of the output response of the rigid-flexible coupling dynamic model of the coordination mechanism are used to obtain the P-box response CDF using the Johnson fitting method. The Johnson fitting method is expressed as:

[0104]

[0105] Among them, g and δ are shape parameters, is the size parameter, x is the position parameter, w~N(0,1) obeys the standard normal distribution, Y is the transformation function, w represents the Johnson distribution, and h is the Johnson variable of the output response of the rigid-flexible coupling dynamic model of the coordination mechanism.

[0106] Furthermore, the determination of whether the area difference between the CDF lower bounds obtained from the two most recent sequence simulations meets a preset standard is specifically as follows:

[0107] Will As the lower bound of the CDF obtained by the k-th sequence simulation, As the CDF lower bound obtained by the k-1th sequence simulation, the area difference between the CDF lower bounds obtained by two adjacent sequence simulations is used as the convergence judgment criterion, which is as follows:

[0108]

[0109] in, is the CDF lower bound obtained by the k-th sequence simulation and the CDF lower bound obtained by the k-1th sequence simulation The Minkowski-L1 distance between Indicates the contribution sample The differential of

[0110] Minkowski-L1 distance as the absolute error of the k-th sequence simulation The convergence criterion is:

[0111]

[0112] Where ε0 is the preset value of error control.

[0113] Furthermore, the value of the number r of the Gaussian integration nodes is 6.

[0114] Furthermore, the preset value of the error control ε0=1×10 -3 .

[0115] The present invention also proposes an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the hybrid uncertainty propagation method for complex equipment response as described above is implemented.

[0116] The present invention also provides a computer-readable storage medium storing a computer program, wherein the computer program enables a computer to execute the hybrid uncertainty propagation method for complex equipment response as described above.

[0117] The beneficial effects of the present invention are:

[0118] 1. For the probability-interval mixed uncertainty propagation analysis problem, the outer layer uses the sequence simulation method to quickly find the samples that determine the CDF boundary in an iterative manner. Compared with the MCS (Monet Carlo Method) method, the calculation efficiency is greatly improved with a smaller error. The CDF solution of each step in the inner layer, the present invention converts the high-dimensional probability propagation problem into a moment analysis problem of a single variable subsystem through a single variable dimensionality reduction integral method. After obtaining the interval of the system response statistical moment, the statistical moment of the response is fitted within its interval to obtain the CDF through the Johnson fitting distribution. Compared with the MCS method, only very few sample points are required to ensure higher accuracy.

[0119] 2. This method is a non-embedded method and is generally applicable to complex engineering problems such as those with high nonlinearity, multiple input parameters, and non-monotonic mechanisms and structural dynamic responses. BRIEF DESCRIPTION OF THE DRAWINGS

[0120] Figure 1 It is a flow chart of the present invention.

[0121] Figure 2 It is a schematic diagram of the rigid-flexible coupling system of the present invention.

[0122] Figure 3 CDF of the contributing samples and their responses for this invention.

[0123] Figure 4 The neighborhood of the contributing samples of the present invention.

[0124] Figure 5 Schematic diagram of the convergence criterion of the present invention. DETAILED DESCRIPTION

[0125] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0126] Embodiment 1

[0127] The present invention proposes a hybrid uncertainty propagation method for complex equipment response, which is applied to a coordination mechanism, wherein the coordination mechanism is a rigid-flexible coupling system, such as Figure 2 As shown in the figure, it includes a motor, a reducer, a coordination robot arm, a posture adjustment electric cylinder assembly and a conveying mechanism. The conveying mechanism is mainly coordinated to a specific position by coordinating the robot arm and the posture adjustment electric cylinder, and a certain in-place accuracy is ensured. Its specific actions are as follows: the robot arm is connected to the rotating shaft through the reducer and can rotate around the rotating shaft, and the posture adjustment limit bumper and the posture adjustment electric cylinder support are fixedly connected to the robot arm. The front end of the posture adjustment electric cylinder is hinged to the conveying mechanism, and the rear end is hinged to the posture adjustment electric cylinder support. The conveying mechanism is installed at the end of the robot arm through the posture adjustment shaft, and performs a small angle rotation around the robot arm under the action of the posture adjustment electric cylinder. The limit position of the movement is guaranteed by the limit bumper and the upper and lower limit blocks of the conveying mechanism. Under the joint action of the motor and the posture adjustment electric cylinder assembly, the axis of the conveying mechanism can be aligned with the transmission receiving device at any angle, thereby realizing the accurate transmission of the object. When mixed uncertainty parameters such as geometric dimensions and physical property parameters exist, it will have a certain impact on the in-place accuracy of the robot arm. Therefore, it is necessary to carry out mixed uncertainty propagation analysis of the rigid-flexible coupling dynamics model.

[0128] The hybrid uncertainty propagation method for complex equipment response proposed in this invention has the following overall process: Figure 1 As shown, the following steps are included:

[0129] Step 1: Establish a rigid-flexible coupling dynamic model of the coordination mechanism; the input parameters of the rigid-flexible coupling dynamic model of the coordination mechanism include probabilistic uncertainty parameters and interval uncertainty parameters.

[0130] During the movement, there will be contact collision between the shaft hole and the limit block, which will generate contact force and damping torque. The contact force model is composed of the Lankarani-Nikravesh normal contact force model F n and the improved Coulomb friction model F t They are calculated as follows:

[0131]

[0132] In the formula, F n is the actual collision force, K is the stiffness coefficient of the two contacting objects, δ is the penetration depth of the contact body, n is the collision index, C is the contact collision damping coefficient, is the relative collision velocity at the contact point. a1 is the sliding friction coefficient, c d is the dynamic correction coefficient, v t is the relative tangential velocity. c is the rotational damping torque, a2 is the friction coefficient of the damping torque, is the angular velocity of rotation, sign is the sign function, when hour, when hour, when hour, a3 is the damping coefficient of the damping moment.

[0133] Since the robot arm is a thin plate structure, it will deform to a certain extent during the movement. Therefore, the robot arm is considered as a flexible body in the modeling process. Combined with the kinematic relationship, force and constraint of the mechanism, based on the principle of virtual power, the rigid-flexible coupling dynamic model of the coordination mechanism is established as follows:

[0134]

[0135] Where M(q,t) is the mass matrix of the coordination mechanism, t is the motion time of the coordination process, q, and are the relative displacement, relative velocity and relative acceleration of the rotary joint in the coordination mechanism, is the generalized damping matrix of the coordination mechanism, is the generalized stiffness matrix of the coordination mechanism, Φ(q,t)=0 is the constraint equation of the coordination mechanism, Φ q Derivative of Φ(q,t) with respect to q, (g) T represents the transpose of the matrix, l is the Lagrange multiplier, and Q is the external force of the coordination mechanism.

[0136] In the rigid-flexible coupling dynamics model of the coordination mechanism, material parameters such as elastic modulus, density, initial position of the mechanism, Coulomb friction coefficient, torque friction coefficient, damping coefficient, load mass and other parameters are uncertain. The values ​​and types of uncertainty parameters are shown in Table 1. The input parameters of the rigid-flexible coupling dynamics model of the coordination mechanism include probabilistic uncertainty parameters and interval uncertainty parameters. Probabilistic uncertainty parameters include elastic modulus, density, initial position, load mass and moment of inertia, and interval uncertainty parameters include Coulomb friction coefficient, dry friction coefficient, damping coefficient, mass center eccentricity along the x direction, mass center eccentricity along the y direction and mass center eccentricity along the z direction.

[0137] Table 1 Uncertainty parameters and their distribution types

[0138]

[0139]

[0140] Step 2: Based on the P-box, the uncertainty of the output response of the rigid-flexible coupling dynamics model of the coordination mechanism is characterized; the output response of the rigid-flexible coupling dynamics model of the coordination mechanism is the angular displacement angle of the coordination robot arm from the initial -3° position to the position with an angle of 45° with the horizontal direction, and the CDF of the output response is expressed by F z (Z) indicates that F z (Z) is enclosed by the lower boundary of the CDF F z (Z) and upper boundary In the area formed, The P-box output response Z of the rigid-flexible coupling dynamic model of the coordinated mechanism is written as:

[0141]

[0142] In the formula, The P-box that responds to Z.

[0143] Step 3: According to the distribution type of the input parameters in Table 1, the input parameters of the rigid-flexible coupling dynamics model are layered, with the outer layer being the interval uncertainty parameters and the inner layer being the probabilistic uncertainty parameters.

[0144] Step 4: Select the interval uncertainty parameters and sample space of the rigid-flexible coupling dynamics model, perform global sampling and local sampling (the first local sampling is 0) in the sample space, and obtain multiple groups of interval samples. where n s-initial =50 is the initial sample number, at which point all interval variables degenerate into scalars, and the system is transformed into a system containing only random variables; interval samples include global samples and local samples (there are no local samples in the initial sequence). Global samples are generated from the entire bounded sample space, and the local samples are generated in the neighborhood of the contributing samples according to the contribution index of the contributing samples; the contributing samples refer to samples that contribute to the upper and lower boundaries of the P-box; in each sequence simulation of the outer interval, samples that contribute to the upper and lower boundaries of the P-box need to be identified. The upper and lower boundaries obtained in each sequence simulation must be composed of some parts of the existing response CDF. Therefore, these input samples that can constitute the CDF response boundary components are called contributing samples.

[0145] The contribution index is calculated as follows:

[0146] like Figure 3 As shown, for each contributing sample The CDF of the P-box is recorded as Its contribution is evaluated by the ordinate span where the CDF takes its maximum value:

[0147]

[0148] In the formula, It is a contribution sample The contribution index of Indicates the contribution sample The vertical axis span that affects the upper or lower boundary of the CDF, F z (Z(y)) represents the total ordinate span of the upper and lower boundaries of the P-box.

[0149] like Figure 4 As shown, the contributing sample points Neighborhood Can be defined in different forms or shapes. Figure 4 The circular neighborhood shown in (a) is the center of the circle Radius r * ) or a multidimensional spherical region (for interval uncertainty parameters greater than 2). Figure 4 The square neighborhood shown in (b) is a The side length is 2r * A square or multidimensional cubic region of . Figure 4 The rectangular neighborhood shown in (c) is centered at The side length is 2r i * ,i=1,2,L,m rectangular or multidimensional rectangular solid area, m is the number of interval parameters. Appropriate size set helps to speed up the convergence of sequence simulation. For ease of description, the ratio b is introduced to represent the neighborhood relative to the entire bounded space W α The size of the circular neighborhood and the square neighborhood is defined as Where V(W α ) is a bounded space W α The volume or area of ​​​​the α For a rectangular area, Determine, among which For interval Through the measurement of the interval uncertainty problem, the present invention suggests setting b between 1 / 12 and 1 / 4.

[0150] use Figure 4(c) The rectangular neighborhood form, the relative neighborhood radius b = 0.1, the global sampling number of each sequence simulation is 20, and the local sampling number is determined by the contributing samples. The simulation times of the sequence simulation are shown in Table 2.

[0151] Table 2 Simulation times of sequence simulation process

[0152]

[0153]

[0154] Step 5, randomly select a group of interval samples in the outer layer, use the dimension reduction integral method to simulate and calculate the inner layer probability uncertainty parameters, and obtain the P-box response CDF; determine whether all the interval samples in the outer layer have been calculated, if so, a series of CDF results are obtained, and proceed to step 6, otherwise continue to step 5. The P-box response CDF obtained by simulating and calculating the inner layer probability uncertainty parameters using the dimension reduction integral method is specifically:

[0155] Step 5.1: Use the single variable dimension reduction integral method to reduce the dimension near the mean value of the probability uncertainty parameter, and transform the multivariable probability uncertainty parameter system into a subsystem containing only a single variable, which can be expressed as:

[0156]

[0157] Among them, E l (Z) represents the lth order moment of the response, l = 1, 2, 3, 4; m i is the mean of the ith probability uncertainty parameter, yes It is an abbreviation of the decomposed univariate subsystem X. i represents the i-th probability uncertainty parameter, Y (d) represents the interval uncertainty parameter, is the binomial expansion, j is the order number, Represents the function Find the jth order moment, m represents the number of probability uncertainty parameters, g0(m1,m2,L,m m ,Y (d) ) indicates that all probability uncertainty parameters are initialized to the mean;

[0158] Step 5.2: Use Gauss-Hermite integration to calculate the first four moments of the univariate subsystem, expressed as:

[0159]

[0160] In the formula, E{·} represents the function The l-th order moment, r represents the number of Gaussian integration nodes, which is set to r=6 in this embodiment. i,k and w k are Gaussian integration nodes and node weights respectively;

[0161] Step 5.3, the first four-order center moments are calculated iteratively by coordinating the first four-order origin moments of the output response of the rigid-flexible coupling dynamic model of the mechanism, which can be expressed as:

[0162]

[0163] Among them, E l (Y (d) ), l = 1, 2, 3, 4 are the first four origin moments of the output response of the rigid-flexible coupling dynamic model of the coordination mechanism, M l (Y (d) ), l = 1, 2, 3, 4 are the first four central moments of the output response of the rigid-flexible coupling dynamic model of the coordination mechanism;

[0164] Step 5.4: Use the Johnson fitting method to obtain the P-box response CDF of the first four-order central moments of the output response of the rigid-flexible coupling dynamic model of the coordination mechanism. The Johnson fitting method is expressed as:

[0165]

[0166] Among them, g and δ are shape parameters, is the size parameter, x is the position parameter, w~N(0,1) obeys the standard normal distribution, Y is the transformation function, w represents the Johnson distribution, and h is the Johnson variable of the output response of the rigid-flexible coupling dynamic model of the coordination mechanism.

[0167] Step 6: Based on the series of CDF results obtained in step 5, obtain the lower boundary of the P-box calculated this time, determine the contribution CDF composition corresponding to the lower boundary, find the contribution sample corresponding to the contribution CDF, and complete a sequence simulation;

[0168] Step 7: Determine whether the area difference between the CDF lower bounds obtained from the two most recent sequence simulations meets the preset standard. If so, it is considered that the calculation has reached convergence and go to step 8; otherwise, go to step 4 to generate new global samples and local samples. Determine whether the area difference between the CDF lower bounds obtained from the two most recent sequence simulations meets the preset standard as follows:

[0169] Will As the lower bound of the CDF obtained by the k-th sequence simulation, As the CDF lower bound obtained by the k-1th sequence simulation, the area difference between the CDF lower bounds obtained by two adjacent sequence simulations is used as the convergence judgment criterion, which is as follows:

[0170]

[0171] in, is the CDF lower bound obtained by the k-th sequence simulation and the CDF lower bound obtained by the k-1th sequence simulation The Minkowski-L1 distance between Indicates the contribution sample The differential of

[0172] Minkowski-L1 distance as the absolute error of the k-th sequence simulation The convergence criterion is:

[0173]

[0174] Where, e0 is the preset value of error control, e0 = 1′10 -3 .

[0175] Step 8: Output the final P-box response CDF lower boundary;

[0176] Step 9. Perform sequence simulation to obtain the upper boundary of the P-box response CDF. The sequence simulation process of the upper boundary is similar to that of the lower boundary. Based on the upper and lower boundaries of the P-box response CDF, the P-box result of the mixed uncertainty propagation of the coordination mechanism is obtained, and the positioning accuracy of the coordination robot arm is adjusted according to the actual working conditions.

[0177] According to the above calculation steps, the P-box calculated by the present invention is compared with DLS (Double Layer Sampling) and MCS-UDRM as shown in Tables 3 and 4.

[0178] Table 3 Comparison of P-box calculation results of several hybrid uncertainty propagation methods

[0179]

[0180] Table 4 Comparison of the fourth-order moments of several hybrid uncertainty propagation methods

[0181]

[0182]

[0183] According to Table 2, for this example, the outer upper bound sequence simulation reached convergence 8 times, and the required sampling number was 330. The lower bound sequence simulation reached convergence 10 times, and the required sampling number was 424. The total number of samplings in the outer layer was 804. Combined with Table 3, it can be seen that the P-box result calculated by the method of the present invention has a maximum error of 0.25% compared with DLS and MCS-UDRM, while the sampling number of DLS is 10. 4 '10 4 and the sampling times of MCS-UDRM is 10 4 '30, the method of the present invention only takes 804'30 to achieve such a high accuracy. Further, according to Table 4, the error of the first four central moments calculated by the method of the present invention is about 2% compared with the DLS results, and only the error of the third central moment is large, but according to the calculation results of MCS-UDRM, it can be seen that the error is mainly caused by the inner layer dimensionality reduction integral, and the outer layer sequence simulation still has high efficiency and accuracy.

[0184] It can be seen that for the rigid-flexible coupling dynamic system in engineering practice with high parameter dimensions, the accuracy of the method of the present invention can still be guaranteed, and the calculation efficiency is greatly improved, which shows that the method of the present invention has high universality.

[0185] Embodiment 2

[0186] The present invention proposes a hybrid uncertainty propagation system for complex equipment response corresponding to the method of embodiment 1, which is applied to a coordination mechanism. The coordination mechanism is a rigid-flexible coupling system, including a motor, a reducer, a coordination manipulator, a posture adjustment electric cylinder assembly and a conveying mechanism. The hybrid uncertainty propagation system includes:

[0187] A modeling module is used to establish a rigid-flexible coupling dynamics model of a coordination mechanism; the input parameters of the rigid-flexible coupling dynamics model of a coordination mechanism include a probability uncertainty parameter and an interval uncertainty parameter;

[0188] Uncertainty characterization module, used to characterize the uncertainty of the output response of the rigid-flexible coupling dynamics model of the coordination mechanism based on P-box;

[0189] A stratification module is used to stratify the input parameters of the rigid-flexible coupling dynamics model according to the distribution type of the input parameters, with the outer layer being interval uncertainty parameters and the inner layer being probabilistic uncertainty parameters;

[0190] A sampling module is used to select interval uncertainty parameters and sample space of the rigid-flexible coupling dynamics model, perform global sampling and local sampling in the sample space, and obtain multiple groups of interval samples, including global samples and local samples;

[0191] The lower boundary generation module is used to select any group of interval samples in the outer layer, simulate and calculate the inner layer probability uncertainty parameters using the dimensionality reduction integral method, and obtain the P-box response CDF; determine whether all the interval samples in the outer layer have been calculated, if so, obtain a series of CDF results, otherwise continue to select the interval samples that have not been calculated for simulation calculation to obtain the P-box response CDF until all the interval samples in the outer layer are calculated; based on a series of CDF results, obtain the P-box lower boundary of this calculation, and determine the contribution CDF composition corresponding to the lower boundary, find the contribution samples corresponding to the contribution CDF, and complete a sequence simulation; determine whether the area difference between the CDF lower bounds obtained from the most recent two sequence simulations meets the preset standard, if so, it is considered that the calculation has reached convergence, and the final P-box response CDF lower bound is output; otherwise, control the sampling module to generate new global samples and local samples;

[0192] The upper boundary generation module is used to perform sequence simulation to obtain the upper boundary of the P-box response CDF;

[0193] The adjustment module is used to obtain the P-box result of the mixed uncertainty propagation of the coordination mechanism based on the upper and lower boundaries of the P-box response CDF, and adjust the positioning accuracy of the coordination robot arm according to the actual working conditions.

[0194] The implementation methods of each module and module function in the system are completely consistent with the steps of the method in Example 1, so they will not be repeated here.

[0195] Embodiment 3

[0196] The present invention proposes an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the hybrid uncertainty propagation method for complex equipment response as described in Embodiment 1 is implemented.

[0197] Embodiment 4

[0198] The present invention provides a computer-readable storage medium storing a computer program, wherein the computer program enables a computer to execute the hybrid uncertainty propagation method for complex equipment response as described in the first embodiment.

[0199] In the embodiments disclosed in the present application, the computer storage medium can be a tangible medium that can contain or store a program for use by or in conjunction with an instruction execution system, device or equipment. The computer storage medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the above. More specific examples of computer storage media can include electrical connections based on one or more lines, portable computer disks, hard disks, random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memory), optical fibers, portable compact disk read-only memories (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above.

[0200] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in this application can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0201] The above are only preferred embodiments of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should be regarded as the protection scope of the present invention.

Claims

1. A hybrid uncertainty propagation method for complex equipment response, applied to coordination mechanism, characterized by: The coordination mechanism is a rigid-flexible coupling system, including a motor, a reducer, a coordination robot arm, a posture adjustment electric cylinder assembly and a conveying mechanism. The method includes the following steps: Step 1, establishing a rigid-flexible coupling dynamic model of the coordination mechanism; the input parameters of the rigid-flexible coupling dynamic model of the coordination mechanism include probability uncertainty parameters and interval uncertainty parameters; Step 2: Characterize the uncertainty of the output response of the rigid-flexible coupling dynamic model of the coordination mechanism based on P-box; Step 3: According to the distribution type of the input parameters, the input parameters of the rigid-flexible coupling dynamics model are layered, with the outer layer being interval uncertainty parameters and the inner layer being probabilistic uncertainty parameters; Step 4: Select the interval uncertainty parameters and sample space of the rigid-flexible coupling dynamics model, perform global sampling and local sampling in the sample space, and obtain multiple groups of interval samples, including global samples and local samples; Step 5: Randomly select a group of interval samples in the outer layer, use the dimension reduction integral method to simulate and calculate the probability uncertainty parameters of the inner layer, and obtain the P-box response CDF; determine whether all interval samples in the outer layer have been calculated. If so, a series of CDF results are obtained and go to step 6, otherwise continue to step 5; Step 6: Based on the series of CDF results obtained in step 5, obtain the lower boundary of the P-box calculated this time, determine the contribution CDF composition corresponding to the lower boundary, find the contribution sample corresponding to the contribution CDF, and complete a sequence simulation; Step 7: Determine whether the area difference between the lower bounds of the CDF obtained from the two most recent sequence simulations meets the preset standard. If so, it is considered that the calculation has reached convergence and go to step 8; otherwise, go to step 4 to generate new global samples and local samples; Step 8: Output the final P-box response CDF lower boundary; Step 9: Perform sequence simulation to obtain the upper boundary of the P-box response CDF. Based on the upper and lower boundaries of the P-box response CDF, obtain the P-box result of the mixed uncertainty propagation of the coordination mechanism, and adjust the positioning accuracy of the coordination robot arm according to the actual working conditions.

2. The hybrid uncertainty propagation method for complex equipment response according to claim 1, characterized in that: In step 1, the rigid-flexible coupling dynamic model of the coordination mechanism is specifically: Where M(q,t) is the mass matrix of the coordination mechanism, t is the motion time of the coordination process, q, and are the relative displacement, relative velocity and relative acceleration of the rotary joint in the coordination mechanism, is the generalized damping matrix of the coordination mechanism, is the generalized stiffness matrix of the coordination mechanism, Φ(q,t)=0 is the constraint equation of the coordination mechanism, Φ q Derivative of Φ(q,t) with respect to q, (·) T represents the transpose of the matrix, λ is the Lagrange multiplier, and Q is the external force of the coordination mechanism.

3. The hybrid uncertainty propagation method for complex equipment response according to claim 1, characterized in that: In step 2, the output response of the rigid-flexible coupling dynamic model of the coordination mechanism is the angular displacement angle of rotating the coordination robot arm from the initial -3° position to the position with an angle of 45° with the horizontal direction, and the CDF of the output response is expressed by F z (Z) indicates that F z (Z) is enclosed by the lower boundary of the CDF F z (Z) and upper boundary In the area formed, The P-box output response Z of the rigid-flexible coupling dynamic model of the coordinated mechanism is written as: In the formula, The P-box that responds to Z.

4. The hybrid uncertainty propagation method for complex equipment response according to claim 3, characterized in that: In step 4, the global samples are generated from the entire bounded sample space, and the local samples are generated in the neighborhood of the contributing samples according to the contribution index of the contributing samples; the contributing samples refer to the samples that contribute to the upper and lower boundaries of the P-box; The contribution index is calculated as follows: For each contribution sample The CDF of the P-box is recorded as Its contribution is evaluated by the ordinate span where the CDF takes its maximum value: In the formula, It is a contribution sample The contribution index of Indicates the contribution sample The vertical axis span that affects the upper or lower boundary of the CDF, F z (Z(y)) represents the total ordinate span of the upper and lower boundaries of the P-box.

5. The hybrid uncertainty propagation method for complex equipment response according to claim 4, characterized in that: The neighborhood of the contributing sample is centered at The radius is r * The circular area or multi-dimensional spherical area of ​​; the radius is calculated as follows: Where β represents the neighborhood relative to the entire bounded sample space Ω α The size of V(Ω α ) is a bounded sample space Ω α The volume or area of ​​, m is the number of interval uncertainty parameters, and it is also the dimension of the bounded sample space.

6. The hybrid uncertainty propagation method for complex equipment response according to claim 4, characterized in that: The neighborhood of the contributing sample is centered at The side length is 2r * A square or multidimensional cubic region; r * The calculation formula is as follows: Where β represents the neighborhood relative to the entire bounded sample space Ω α The size of V(Ω α ) is a bounded sample space Ω α The volume or area of ​​, m is the number of interval uncertainty parameters, and it is also the dimension of the bounded sample space.

7. The hybrid uncertainty propagation method for complex equipment response according to claim 4, characterized in that: The neighborhood of the contributing sample is centered at The side length is is a rectangular or multidimensional rectangular solid area, m is the number of interval uncertainty parameters, The calculation formula is as follows: Where β represents the neighborhood relative to the entire bounded sample space Ω α The size of For interval Width.

8. The hybrid uncertainty propagation method for complex equipment response according to any one of claims 5 to 7, characterized in that: The value of β is set between 1 / 12 and 1 / 4.

9. The hybrid uncertainty propagation method for complex equipment response according to claim 1, characterized in that: In step 5, the inner layer probability uncertainty parameter is simulated and calculated using the dimension reduction integration method to obtain the P-box response CDF as follows: Step 5.1: Use the single variable dimension reduction integral method to reduce the dimension near the mean value of the probability uncertainty parameter, and transform the multivariable probability uncertainty parameter system into a subsystem containing only a single variable, which can be expressed as: Among them, E l (Z) represents the lth order moment of the response, l = 1, 2, 3, 4; μ i is the mean of the ith probability uncertainty parameter, yes It is an abbreviation of the decomposed univariate subsystem X. i represents the i-th probability uncertainty parameter, Y (d) represents the interval uncertainty parameter, is the binomial expansion, j is the order number, Represents the function Find the jth order moment, m represents the number of probability uncertainty parameters, g0(μ1,μ2,…,μ m ,Y (d) ) indicates that all probability uncertainty parameters are initialized to the mean; Step 5.2: Use Gauss-Hermite integration to calculate the first four moments of the univariate subsystem, expressed as: In the formula, E{·} represents the function The l-th order moment, r represents the number of Gaussian integration nodes, U i,k and ω k are Gaussian integration nodes and node weights respectively; Step 5.3, the first four-order center moments are calculated iteratively by coordinating the first four-order origin moments of the output response of the rigid-flexible coupling dynamic model of the mechanism, which can be expressed as: Among them, E l (Y (d) ), l = 1, 2, 3, 4 are the first four origin moments of the output response of the rigid-flexible coupling dynamic model of the coordination mechanism, M l (Y (d) ), l = 1, 2, 3, 4 are the first four central moments of the output response of the rigid-flexible coupling dynamic model of the coordination mechanism; Step 5.4: Use the Johnson fitting method to obtain the P-box response CDF of the first four-order central moments of the output response of the rigid-flexible coupling dynamic model of the coordination mechanism. The Johnson fitting method is expressed as: Among them, γ and δ are shape parameters, is the size parameter, ξ is the position parameter, w~N(0,1) obeys the standard normal distribution, Ψ is the transformation function, w represents the Johnson distribution, and h is the Johnson variable of the output response of the rigid-flexible coupling dynamic model of the coordination mechanism.

10. The hybrid uncertainty propagation method for complex equipment response according to claim 1, characterized in that: In step 7, the determination of whether the area difference between the CDF lower bounds obtained from the two most recent sequence simulations meets the preset standard is specifically as follows: Will As the lower bound of the CDF obtained by the k-th sequence simulation, As the CDF lower bound obtained by the k-1th sequence simulation, the area difference between the CDF lower bounds obtained by two adjacent sequence simulations is used as the convergence judgment criterion, which is as follows: in, is the CDF lower bound obtained by the k-th sequence simulation and the CDF lower bound obtained by the k-1th sequence simulation The Minkowski-L1 distance between Indicates the contribution sample The differential of Minkowski-L1 distance as the absolute error of the k-th sequence simulation The convergence criterion is: Where ε0 is the preset value of error control.

11. The hybrid uncertainty propagation method for complex equipment response according to claim 9, characterized in that: In step 5.2, the value of the number r of the Gaussian integration nodes is 6.

12. The hybrid uncertainty propagation method for complex equipment response according to claim 10, characterized in that: The preset value of the error control ε0=1×10 -3 .

13. A hybrid uncertainty propagation system for complex equipment response, applied to coordination organizations, characterized in that: The coordination mechanism is a rigid-flexible coupling system, including a motor, a reducer, a coordination robot arm, a posture adjustment electric cylinder assembly and a conveying mechanism. The hybrid uncertainty propagation system includes: A modeling module is used to establish a rigid-flexible coupling dynamics model of a coordination mechanism; the input parameters of the rigid-flexible coupling dynamics model of a coordination mechanism include a probability uncertainty parameter and an interval uncertainty parameter; Uncertainty characterization module, used to characterize the uncertainty of the output response of the rigid-flexible coupling dynamics model of the coordination mechanism based on P-box; A stratification module is used to stratify the input parameters of the rigid-flexible coupling dynamics model according to the distribution type of the input parameters, with the outer layer being interval uncertainty parameters and the inner layer being probabilistic uncertainty parameters; A sampling module is used to select interval uncertainty parameters and sample space of the rigid-flexible coupling dynamics model, perform global sampling and local sampling in the sample space, and obtain multiple groups of interval samples, including global samples and local samples; The lower boundary generation module is used to select any group of interval samples in the outer layer, simulate and calculate the inner layer probability uncertainty parameters using the dimensionality reduction integral method, and obtain the P-box response CDF; determine whether all the interval samples in the outer layer have been calculated, if so, obtain a series of CDF results, otherwise continue to select the interval samples that have not been calculated for simulation calculation to obtain the P-box response CDF until all the interval samples in the outer layer are calculated; based on a series of CDF results, obtain the P-box lower boundary of this calculation, and determine the contribution CDF composition corresponding to the lower boundary, find the contribution samples corresponding to the contribution CDF, and complete a sequence simulation; determine whether the area difference between the CDF lower bounds obtained from the most recent two sequence simulations meets the preset standard, if so, it is considered that the calculation has reached convergence, and the final P-box response CDF lower bound is output; otherwise, control the sampling module to generate new global samples and local samples; The upper boundary generation module is used to perform sequence simulation to obtain the upper boundary of the P-box response CDF; The adjustment module is used to obtain the P-box result of the mixed uncertainty propagation of the coordination mechanism based on the upper and lower boundaries of the P-box response CDF, and adjust the positioning accuracy of the coordination robot arm according to the actual working conditions.

14. The hybrid uncertainty propagation system for complex equipment response according to claim 13, characterized in that: The rigid-flexible coupling dynamic model of the coordination mechanism is specifically: Where M(q,t) is the mass matrix of the coordination mechanism, t is the motion time of the coordination process, q, and are the relative displacement, relative velocity and relative acceleration of the rotary joint in the coordination mechanism, is the generalized damping matrix of the coordination mechanism, is the generalized stiffness matrix of the coordination mechanism, Φ(q,t)=0 is the constraint equation of the coordination mechanism, Φ q Derivative of Φ(q,t) with respect to q, (g) T represents the transpose of the matrix, l is the Lagrange multiplier, and Q is the external force of the coordination mechanism.

15. The hybrid uncertainty propagation system for complex equipment response according to claim 13, characterized in that: The output response of the rigid-flexible coupling dynamic model of the coordination mechanism is to rotate the coordination robot arm from the initial position of -3° to the position with an angle of 45° with the horizontal direction to the angular displacement angle, and output the CDF of the response, which is expressed by F z (Z) indicates that F z (Z) is enclosed by the lower boundary of the CDF F z (Z) and upper boundary In the area formed, The P-box output response Z of the rigid-flexible coupling dynamic model of the coordinated mechanism is written as: In the formula, The P-box that responds to Z.

16. The hybrid uncertainty propagation system for complex equipment response according to claim 15, characterized in that: The global samples are generated from the entire bounded sample space, and the local samples are generated in the neighborhood of the contributing samples according to the contribution index of the contributing samples; the contributing samples refer to the samples that contribute to the upper and lower boundaries of the P-box; The contribution index is calculated as follows: For each contribution sample The CDF of the P-box is recorded as Its contribution is evaluated by the ordinate span where the CDF takes its maximum value: In the formula, It is a contribution sample The contribution index of Indicates the contribution sample The vertical axis span that affects the upper or lower boundary of the CDF, F z (Z(y)) represents the total ordinate span of the upper and lower boundaries of the P-box.

17. The hybrid uncertainty propagation system for complex equipment response according to claim 16, characterized in that: The neighborhood of the contributing sample is centered at The radius is r * The circular area or multi-dimensional spherical area of ​​; the radius is calculated as follows: Where β represents the neighborhood relative to the entire bounded sample space Ω α The size of V(Ω α ) is a bounded sample space Ω α The volume or area of ​​, m is the number of interval uncertainty parameters, and it is also the dimension of the bounded sample space.

18. The hybrid uncertainty propagation system for complex equipment response according to claim 16, characterized in that: The neighborhood of the contributing sample is centered at The side length is 2r * A square or multidimensional cubic region; r * The calculation formula is as follows: Where β represents the neighborhood relative to the entire bounded sample space Ω α The size of V(Ω α ) is a bounded sample space Ω α The volume or area of ​​, m is the number of interval uncertainty parameters, and it is also the dimension of the bounded sample space.

19. The hybrid uncertainty propagation system for complex equipment response according to claim 16, characterized in that: The neighborhood of the contributing sample is centered at The side length is is a rectangular or multidimensional rectangular solid area, m is the number of interval uncertainty parameters, The calculation formula is as follows: Where β represents the neighborhood relative to the entire bounded sample space Ω α The size of For interval Width.

20. The hybrid uncertainty propagation system for complex equipment response according to any one of claims 17 to 19, characterized in that: The value of β is set between 1 / 12 and 1 / 4.

21. The hybrid uncertainty propagation system for complex equipment response according to claim 13, characterized in that: The inner layer probability uncertainty parameters are simulated and calculated using the dimension reduction integration method to obtain the P-box response CDF as follows: The single variable dimension reduction integral method is used for the probability uncertainty parameter, and the dimension reduction is expanded near the mean, and the multivariable probability uncertainty parameter system is transformed into a subsystem containing only a single variable, which is expressed as: Among them, E l (Z) represents the lth order moment of the response, l = 1, 2, 3, 4; μ i is the mean of the ith probability uncertainty parameter, yes It is an abbreviation of the decomposed univariate subsystem X. i represents the i-th probability uncertainty parameter, Y (d) represents the interval uncertainty parameter, is the binomial expansion, j is the order number, Represents the function Find the jth order moment, m represents the number of probability uncertainty parameters, g0(μ1,μ2,…,μ m ,Y (d) ) indicates that all probability uncertainty parameters are initialized to the mean; The Gauss-Hermite integral is used to calculate the first four moments of the univariate subsystem, expressed as: In the formula, E{·} represents the function The l-th order moment, r represents the number of Gaussian integration nodes, U i,k and ω k are Gaussian integration nodes and node weights respectively; The first four-order central moments are calculated iteratively by coordinating the first four-order origin moments of the output response of the rigid-flexible coupling dynamic model of the mechanism, and are expressed as: Among them, E l (Y (d) ), l = 1, 2, 3, 4 are the first four origin moments of the output response of the rigid-flexible coupling dynamic model of the coordination mechanism, M l (Y (d) ), l = 1, 2, 3, 4 are the first four central moments of the output response of the rigid-flexible coupling dynamic model of the coordination mechanism; The first four-order central moments of the output response of the rigid-flexible coupling dynamic model of the coordination mechanism are used to obtain the P-box response CDF using the Johnson fitting method. The Johnson fitting method is expressed as: Among them, γ and δ are shape parameters, is the size parameter, ξ is the position parameter, w~N(0,1) obeys the standard normal distribution, Ψ is the transformation function, w represents the Johnson distribution, and h is the Johnson variable of the output response of the rigid-flexible coupling dynamic model of the coordination mechanism.

22. The hybrid uncertainty propagation system for complex equipment response according to claim 13, characterized in that: The determination of whether the area difference between the CDF lower bounds obtained from the two most recent sequence simulations meets the preset standard is specifically as follows: Will As the lower bound of the CDF obtained by the k-th sequence simulation, As the CDF lower bound obtained by the k-1th sequence simulation, the area difference between the CDF lower bounds obtained by two adjacent sequence simulations is used as the convergence judgment criterion, which is as follows: in, is the CDF lower bound obtained by the k-th sequence simulation and the CDF lower bound obtained by the k-1th sequence simulation The Minkowski-L1 distance between Indicates the contribution sample The differential of Minkowski-L1 distance as the absolute error of the k-th sequence simulation The convergence criterion is: Where ε0 is the preset value of error control.

23. The hybrid uncertainty propagation system for complex equipment response according to claim 21, characterized in that: The value of the number r of the Gaussian integration nodes is 6.

24. The hybrid uncertainty propagation system for complex equipment response according to claim 22, characterized in that: The preset value of the error control ε0=1×10 -3 .

25. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the hybrid uncertainty propagation method for complex equipment response as described in any one of claims 1 to 12 is implemented.

26. A computer-readable storage medium storing a computer program, characterized in that: The computer program enables a computer to execute the hybrid uncertainty propagation method for complex equipment response according to any one of claims 1 to 12.

Citation Information

Patent Citations

  • Unified probability interval hybrid uncertainty propagation analysis method

    CN110781622A

  • Sensitivity analysis method under hybrid uncertainty

    CN113051851A

  • Improved multi-flexible body system dynamics prediction method with hybrid uncertainty

    CN113255150A

  • Hybrid uncertainty mechanical property analysis method, electronic equipment and medium

    CN114925480A

  • Dynamic response and reliability prediction method for hybrid uncertainty system

    CN115495871A