Spline pair misalignment wear prediction method based on laser sensor

Abstract

The embodiment of the invention provides a spline countershaft laser monitoring wear prediction and adjustment method. The method comprises the steps that an adjustable laser sensor and an axial and radial grating ruler are installed; calibrating the position of the position-adjustable laser sensor group; the angle misalignment amount and the radial misalignment amount of the spline pair are obtained through calculation; the mass wear rate is calculated through a weighing method; the average contact pressure is calculated through the torque and the spline basic parameters; simulating the working condition through a finite element to obtain a simulated slip amount; predicting the wear degree of the tooth surface of the spline pair through the Ruiz parameter; evaluating and establishing a sample space of the wear rate data, the Ruiz parameters and corresponding parameters, and training and testing samples in the sample space based on a neural network; storing the learned neural network model as a database; the database predicts the amount of wear based on the feature values without detaching the spline pair. Through laser monitoring and neural network autonomous learning, the shaft misalignment amount is monitored under the condition that the spline pair is not disassembled, and the abrasion loss of the spline pair is predicted.

Description

Technical Field

[0001] The invention relates to a method for predicting misalignment wear of a spline pair based on a laser sensor, and belongs to the technical field of micro-tribology in mechanical engineering. Background Art

[0002] As a key component in mechanical transmission, the spline pair's design and structure give it unique advantages: exceptional load-bearing capacity, excellent alignment, and good guidance. These advantages make it an irreplaceable role in various transmission systems. However, due to factors such as manufacturing errors, assembly errors, and overturning moments, spline pairs can experience radial misalignment, angular misalignment, and axial float. Fretting wear refers to the wear caused by the relative motion of two contacting surfaces at small amplitudes. In spline pairs, fretting wear caused by radial and angular misalignment can gradually increase the clearance between the key teeth and change the key tooth shape, thereby affecting transmission efficiency and accuracy. In the actual use of high-speed motors as hub-drive motors, spline pairs often fail due to fretting wear caused by misalignment. This causes severe wear to both the motor and reducer splines, and some reducer spline ends wear unevenly. Spline pair failure can significantly shorten the equipment life.

[0003] In recent years, with the explosive development of cutting-edge technologies such as machine learning and neural networks, the remarkable capabilities of computers for autonomous learning and decision-making have increasingly attracted attention from both academia and industry, replacing traditional models that rely on extensive experiments and simulations. Database models built using autonomous learning neural networks demonstrate significant advantages over traditional experimental and simulation methods. Furthermore, artificial intelligence algorithms demonstrate exceptional value thanks to their high computational speed and low cost.

[0004] Previous studies on spline wear have typically measured it using finite element methods or experimental methods, and predicted it using statistical or other methods. Spline wear is subject to complex loads, and the kinematic wear model is also relatively complex. If misalignment measurements fail to accurately reflect the misalignment of high-speed shafts, leading to erroneous predictions and incorrect results, compensatory measures may be counterproductive and potentially lead to more serious consequences. Summary of the Invention

[0005] The purpose of the present invention is to solve the problem that it is difficult to monitor the operation of existing spline pairs in real time, and the prediction process cannot correspond to the actual operation in real time, thereby providing a spline pair laser monitoring wear prediction and adjustment method based on neural network and finite element simulation.

[0006] The present invention adopts a spline pair laser monitoring wear prediction and adjustment method based on neural network and finite element simulation, which is characterized by the following steps:

[0007] (1) The method includes a horizontal adjustable laser sensor group, a vertical adjustable laser sensor group, and an auxiliary calibration mechanism; the horizontal adjustable laser sensor group includes a horizontal laser sensor 1, a horizontal laser sensor 2, a horizontal laser sensor 3, and a horizontal laser sensor 4; the vertical adjustable laser sensor group includes a vertical laser sensor 1, a vertical laser sensor 2, a vertical laser sensor 3, and a vertical laser sensor 4; the horizontal adjustable laser sensor group and the vertical adjustable laser sensor group are aligned with a rotation axis; the horizontal adjustable laser sensor group and the vertical adjustable laser sensor group each include four laser sensors, in order of position, the horizontal laser sensor 1 and the vertical laser sensor 1 are collectively referred to as laser sensor 1, the horizontal laser sensor 2 and the vertical laser sensor 2 are collectively referred to as laser sensor 2, the horizontal laser sensor 3 and the vertical laser sensor 3 are collectively referred to as laser sensor 3, and the horizontal laser sensor 4 and the vertical laser sensor 4 are collectively referred to as laser sensor 4;

[0008] (2) When the rotating shaft is not working, measure the distance between the laser sensor 1 and the laser sensor 2 , measure the distance between the laser sensor 3 and the laser sensor 4 , the diameter of the rotating shaft , the distance between the laser sensor 1 and the laser sensor 4 , the distance between the laser sensor 2 and the laser sensor 3 ; The distance between the horizontal plane laser sensor 1 and the horizontal plane laser sensor 2 is The distance between the vertical laser sensor 1 and the vertical laser sensor 2 is The data in the horizontal adjustable laser sensor group are the corresponding suffix 1, and the data in the vertical adjustable laser sensor group are the corresponding suffix 2; the definitions of the remaining corresponding data are the same as 、 The same are 、 、 、 、 、 ;

[0009] (3) Any time after the rotating shaft starts working is called Time, unit frequency is the first preset value, the frequency count is a positive integer Frequency; the laser sensor 1 is Read data at any time , the laser sensor 2 is Read data at any time , the laser sensor 3 is Read data at any time , the laser sensor 4 is Read data at any time , first calculation radius , Second calculation radius The average unit frequency data of the horizontal plane laser sensor 1 in the horizontal plane adjustable laser sensor group is: ; The unit frequency average data of the remaining corresponding data is calculated and The same are 、 、 、 、 、 、 、 、 、 ;

[0010] (4) The z-axis is perpendicular to the horizontal plane, the x-axis is the axial direction, and the y-axis is in the direction that satisfies the spatial coordinate system conditions with the z-axis and the x-axis; when the rotating shaft is not working, the z-axis, x-axis, and y-axis positions of the horizontal position adjustable laser sensor group are calibrated; the y-axis position is adjusted so that the horizontal plane laser sensor 1, the horizontal plane laser sensor 2, the horizontal plane laser sensor 3, and the horizontal plane laser sensor 4 read data; the displacement of each direction of the auxiliary calibration mechanism is 0; the z-axis position is adjusted for measurement so that 、 、 、 , the error allowable value is the second preset value; adjust the y-axis fine-tuning module to measure, so that , the laser sensor readings are subtracted from each other satisfy ; Ensure that the Y-axis position and Z-axis position of the horizontal adjustable laser sensor group remain unchanged; measure the position coordinates of the four laser sensors through the axial radial grating ruler, so that ,; The displacement of the auxiliary calibration mechanism in each direction is 0, and the horizontal plane deflection angle is ; Adjust the x-axis fine-tuning module for measurement , ; The calibration is considered completed;

[0011] (5) When the rotating shaft is not working, calibrate the z-axis, x-axis, and y-axis positions of the vertical position adjustable laser sensor group; adjust the z-axis position so that the vertical laser sensor 1, the vertical laser sensor 2, the vertical laser sensor 3, and the vertical laser sensor 4 read data; the displacement of the auxiliary calibration mechanism in each direction is 0; adjust the y-axis position for measurement so that 、 、 、 , the error allowable value is the second preset value; adjust the z-axis fine-tuning module to measure, so that , the laser sensor readings are subtracted from each other satisfy ; Ensure that the y-axis position and z-axis position of the vertical adjustable laser sensor group remain unchanged; measure the position coordinates of the four laser sensors through the axial radial grating ruler, so that ,; The displacement of the auxiliary calibration mechanism in each direction is 0, and the vertical deflection angle is ; Adjust the x-axis fine-tuning module for measurement , ; The calibration is considered completed;

[0012] (6) When the rotating shaft is not working, the angle misalignment of the spline pair is adjusted by the auxiliary calibration mechanism and radial misalignment of spline pair is 0, and the axis position at this time is called the reference axis of the rotation axis;

[0013] (7) When the rotating shaft is working, the horizontal position adjustable laser sensor group collects data in real time 、 、 、 Collect data in real time with the vertical position adjustable laser sensor group 、 、 、 ; Collect the load information of the spline pair during operation, including using the torque sensor to collect the torque during the spline pair load-bearing process , transportation time , Carrying speed ;

[0014] (8) A simple traversal method is used for real-time calculation, and the unit frequency average data of the horizontal laser sensor 1 is The data flow is used Use it as a variable to record the current maximum value; whenever new data arrives, compare it with the current maximum value. If the new data is greater, update the maximum value; similarly, calculate in real time 、 、 、 、 、 、 The maximum value of the corresponding data flow , , 、 , , , The horizontal laser sensor 1 and the horizontal laser sensor 3 are offset relative to the reference axis of the rotation axis. , the vertical laser sensor 2 and the vertical laser sensor 4 are offset relative to the reference axis of the rotation axis The horizontal laser sensor 1 and the horizontal laser sensor 3 are offset relative to the reference axis of the rotation axis. The vertical laser sensor 2 and the vertical laser sensor 4 are offset relative to the reference axis of the rotation axis. ;

[0015] (9) Calculate after the data becomes stable. If and If both are positive or negative, then , is the same direction offset, the radial misalignment of the spline pair horizontal plane: , horizontal plane angular misalignment of the spline pair: At this time, the axial misalignment of the spline pair horizontal plane is is a small effect, and As mentioned above, the radial misalignment of the horizontal plane Angular misalignment with the horizontal plane of the spline pair Expression, so take ;like and If the value is opposite, , is the reverse offset, the radial misalignment of the spline pair horizontal plane: , horizontal plane misalignment of spline pair At this time, the axial misalignment of the spline pair horizontal plane is The maximum influence is the axial misalignment of the spline pair horizontal plane. ;

[0016] (10) Calculate after the data becomes stable. If and If both are positive or negative, then , is the same direction offset, the radial misalignment of the vertical surface of the spline pair: , the angular misalignment of the vertical surface of the spline pair: At this time, the axial misalignment of the vertical surface of the spline pair is is a small effect, and As mentioned above, the radial misalignment of the vertical plane Angular misalignment with the vertical surface of the spline pair Expression, so take ;like and If the value is opposite, , is the reverse offset, the radial misalignment of the vertical surface of the spline pair: , angular misalignment of the vertical surface of the spline pair At this time, the axial misalignment of the vertical surface of the spline pair is The maximum influence is the axial misalignment of the vertical surface of the spline pair. ;

[0017] (11) The angular misalignment of the spline pair Record as matrix [ 、 ], the radial misalignment of the spline pair Record as matrix [ 、 ]、Axial misalignment of spline pair Record as matrix [ 、 ]; the angle misalignment of the spline pair , the radial misalignment of the spline pair , the axial misalignment of the spline pair , the torque , the transportation time , the carrying speed Combine them into the characteristic value of the spline pair carrying condition data and save them as the first sample size;

[0018] (12) Measure the mass before the test by weighing ; Quality after test ; Measured mass change ; Calculate wear rate ; and save the wear rate as the second sample size;

[0019] (13) Establish a three-dimensional model based on the geometric design parameters of the involute spline pair using the present method, appropriately simplify the model based on the material mechanical parameters, perform finite element simulation analysis based on the characteristic values of the spline pair's load data, and measure the axial slip distance. , tangential slip distance , axial contact length , radial contact length ; Determine the friction coefficient based on the relevant material coefficient ;

[0020] (14) Combined with finite element simulation, the axial slip distance is obtained , tangential slip distance Calculate the relative slip between meshing teeth of a spline pair , combined with finite element simulation axial contact length , radial contact length Calculate the average contact pressure in the contact area of the tooth surfaces , through the relative slip between the meshing teeth of the spline pair and average contact pressure The Ruiz parameter is calculated to quantitatively predict the wear degree of the spline tooth surface, the maximum wear position and the estimated life.

[0021] Relative slip between meshing teeth of spline pair ; Average contact pressure ;

[0022] Ruiz parameters ; and save the Ruiz parameter as the third sample size;

[0023] (15) Repeat steps (6) to (14) to establish a first sample space based on the first sample size, the second sample size, and the third sample size, and divide the first sample space into a training set and a test set. The first sample space includes multiple sets of spline pair carrying data information; the first sample space includes data information outside the training set as the test set; and the percentage of the number of the test set to the number of the first sample space is a third preset value.

[0024] (16) Using a neural network to train the training set in the first sample space, a prediction model for predicting the spline pair wear rate and Ruiz parameter based on the characteristic value of the spline pair carrying condition data information is obtained; secondly, the characteristic value of the spline pair carrying condition data information contained in the test set is input into the prediction model to obtain the corresponding predicted value of the spline pair wear rate and Ruiz parameter; comparing the predicted value of the spline pair wear rate and Ruiz parameter with the standard error absolute value of the spline pair wear rate and Ruiz parameter contained in the test set, and judging whether the standard error absolute value of the prediction accuracy of the prediction model meets the fourth preset value, and the judgment condition is: if the standard error absolute value is greater than the fourth preset value, then continue to increase the number of the first sample space in step (11), and then execute the subsequent steps, and repeat the above steps until the standard error absolute value meets the fourth preset value. The loop ends when the value is less than or equal to the fourth preset value; if the predicted value of the spline pair wear rate and the Ruiz parameter predicted by the prediction model and the standard error absolute value of the spline pair wear rate and the Ruiz parameter in the test set are less than or equal to the fourth preset value, it is considered that the mapping relationship between the spline pair carrying condition data information and the wear rate, and the mapping relationship between the spline pair carrying condition data information and the Ruiz parameter are obtained; save the mapping relationship, and at the same time save the prediction model as a database; generate an expanded data set through random noise injection, parameter interpolation and extrapolation, and combined parameters, use the expanded data set to retrain the neural network to obtain the mapping relationship, and use the test set to test the new mapping relationship. If the standard error absolute value of the spline pair wear rate and the Ruiz parameter in the test set is less than or equal to the fourth preset value, then it is stored in the database to expand the database to obtain a new database;

[0025] (17) The characteristic values of the data information of the spline pair carrying condition during the test described in step (7) are directly input into the database described in step (16). The database will automatically predict the wear rate of the spline pair and the Ruiz parameter to determine whether the predicted wear is within the allowable range of the fifth preset value.

[0026] (18) When the working conditions are limited, only the horizontal position adjustable laser sensor group can be selected for use. The angle misalignment of the spline pair during operation is Record as matrix [ ], the radial misalignment Record as matrix [ ], after the data becomes stable, the horizontal plane misalignment of the spline pair is approximately , the radial misalignment of the spline pair horizontal plane is approximately .

[0027] (19) When the working conditions are limited, only the vertical position adjustable laser sensor group can be selected for use, and the angle misalignment of the spline pair during operation can be adjusted. Record as matrix [ ], the radial misalignment Record as matrix [ ], after stabilization, the vertical angle misalignment of the spline pair is calculated. , Vertical radial misalignment . BRIEF DESCRIPTION OF THE DRAWINGS

[0028] The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following detailed description, they are used to explain the embodiments of the present invention, but do not constitute a limitation of the embodiments of the present invention. In the accompanying drawings:

[0029] Figure 1 This is the overall flow chart of the method for predicting misalignment wear of spline pairs based on laser sensors.

[0030] Figure 2 Flowchart of the calibration position of the laser sensor group with adjustable horizontal position

[0031] Figure 3 This is the flow chart of the calibration position of the vertical position adjustable laser sensor group

[0032] Figure 4 It is the target database prediction creation and application flow chart

[0033] Figure 5 This is the flowchart for expanding the dataset

[0034] Figure 6 This is a schematic diagram of the laser sensor layout

[0035] Figure 7 This is a schematic diagram of the measurement data of the adjustable laser sensor group

[0036] Figure 8 This is the calibration diagram of laser sensor 1 and laser sensor 3 Figure 1

[0037] Figure 9 This is the calibration diagram of laser sensor 2 and laser sensor 4 Figure 1

[0038] Figure 10 This is the calibration diagram of laser sensor 1 and laser sensor 3 Figure 2

[0039] Figure 11 This is the calibration diagram of laser sensor 2 and laser sensor 4 Figure 2

[0040] Figure 12 yes Calculation diagram

[0041] Figure 13 yes Calculation diagram

[0042] Figure 14 yes Calculation diagram

[0043] Figure 15 yes Calculation diagram

[0044] Figure 16 This is a schematic diagram of data distribution provided by an embodiment of the present invention.

[0045] Figure 17 This is a data comparison diagram provided by an embodiment of the present invention

[0046] Figure 18 This is a schematic diagram of a neural network provided by an embodiment of the present invention.

[0047] Figure 19 This is the error histogram of 20 intervals provided by the embodiment of the present invention.

[0048] Figure 20 The error gradient map provided by the embodiment of the present invention is

[0049] Figure 21 This is a training status diagram provided by an embodiment of the present invention.

[0050] Figure 22 This is a working condition prediction result diagram provided by an embodiment of the present invention

[0051] Figure 23 This is a working condition debugging suggestion diagram provided by an embodiment of the present invention DETAILED DESCRIPTION

[0052] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only 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 ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0053] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, but they are not intended to limit the present invention.

[0054] A method for predicting and adjusting the wear of a spline pair by laser monitoring based on neural network and finite element simulation is characterized by comprising the following steps:

[0055] (1) When the rotating shaft is not working, measure the distance between the laser sensor 1 and the laser sensor 2 , measure the distance between the laser sensor 3 and the laser sensor 4 , the diameter of the rotating shaft , the distance between the laser sensor 1 and the laser sensor 4 , the distance between the laser sensor 2 and the laser sensor 3 ; Determine the distance between the horizontal surface laser sensor 1 and the horizontal surface laser sensor 2 , the distance between the vertical laser sensor 1 and the vertical laser sensor 2 is measured to be The data in the horizontal adjustable laser sensor group are the corresponding suffix 1, and the data in the vertical adjustable laser sensor group are the corresponding suffix 2; the definitions of the remaining corresponding data are the same as 、 The same are 、 、 、 、 、 , measure all the data;

[0056] (2) Determine the laser sensor 1 at Read data at all times , the laser sensor 2 is Read data at all times , the laser sensor 3 is Read data at all times , the laser sensor 4 is Read data at all times ; Calculate the first calculation radius , Second calculation radius ; Calculate the average unit frequency data of the horizontal plane laser sensor 1 in the horizontal plane adjustable laser sensor group: ; The unit frequency average data of the remaining corresponding data is calculated and The same are 、 、 、 、 、 、 、 、 、 , determine and calculate all the data;

[0057] (3) When the rotating shaft is not working, calibrate the z-axis, x-axis, and y-axis positions of the horizontal position adjustable laser sensor group; adjust the y-axis position so that the horizontal laser sensor 1, the horizontal laser sensor 2, the horizontal laser sensor 3, and the horizontal laser sensor 4 read data; the displacement of each direction of the auxiliary calibration mechanism is 0; adjust the z-axis position for measurement so that 、 、 、 , the error allowable value is the second preset value; adjust the y-axis fine-tuning module to measure, so that , the laser sensor readings are subtracted from each other satisfy ; Ensure that the Y-axis position and Z-axis position of the horizontal adjustable laser sensor group remain unchanged; measure the position coordinates of the four laser sensors through the axial radial grating ruler, so that ,; The displacement of the auxiliary calibration mechanism in each direction is 0, and the horizontal plane deflection angle is ; Adjust the x-axis fine-tuning module for measurement , ; The calibration is considered completed;

[0058] (4) When the rotating shaft is not working, calibrate the z-axis, x-axis, and y-axis positions of the vertical position adjustable laser sensor group; adjust the z-axis position so that the vertical laser sensor 1, the vertical laser sensor 2, the vertical laser sensor 3, and the vertical laser sensor 4 read data; the displacement of the auxiliary calibration mechanism in each direction is 0; adjust the y-axis position for measurement so that 、 、 、 , the error allowable value is the second preset value; adjust the z-axis fine-tuning module to measure, so that , the laser sensor readings are subtracted from each other satisfy ; Ensure that the y-axis position and z-axis position of the vertical adjustable laser sensor group remain unchanged; measure the position coordinates of the four laser sensors through the axial radial grating ruler, so that ,; The displacement of the auxiliary calibration mechanism in each direction is 0, and the vertical deflection angle is ; Adjust the x-axis fine-tuning module for measurement , ; The calibration is considered completed;

[0059] (5) When the rotating shaft is working, the horizontal position adjustable laser sensor group collects data in real time 、 、 、 Collect data in real time with the vertical position adjustable laser sensor group 、 、 、 ; Collect the load information of the spline pair during operation, including using the torque sensor to collect the torque during the spline pair load-bearing process , transportation time , Carrying speed ;

[0060] (6) A simple traversal method is used for real-time calculation, and the unit frequency average data of the horizontal laser sensor 1 is The data flow is used Use it as a variable to record the current maximum value; whenever new data arrives, compare it with the current maximum value. If the new data is greater, update the maximum value; similarly, calculate in real time 、 、 、 、 、 、 The maximum value of the corresponding data flow , , 、 , , , The horizontal laser sensor 1 and the horizontal laser sensor 3 are offset relative to the reference axis of the rotation axis. , the vertical laser sensor 2 and the vertical laser sensor 4 are offset relative to the reference axis of the rotation axis The horizontal laser sensor 1 and the horizontal laser sensor 3 are offset relative to the reference axis of the rotation axis. The vertical laser sensor 2 and the vertical laser sensor 4 are offset relative to the reference axis of the rotation axis. ;

[0061] (7) Calculate after the data becomes stable. If and If both are positive or negative, then , is the same direction offset, the radial misalignment of the spline pair horizontal plane: , horizontal plane angular misalignment of the spline pair: At this time, the axial misalignment of the spline pair horizontal plane is is a small effect, and As mentioned above, the radial misalignment of the horizontal plane Angular misalignment with the horizontal plane of the spline pair Expression, so take ;like and If the value is opposite, , is the reverse offset, the radial misalignment of the spline pair horizontal plane: , horizontal plane misalignment of spline pair At this time, the axial misalignment of the spline pair horizontal plane is The maximum influence is the axial misalignment of the spline pair horizontal plane. ;

[0062] (8) Calculate after the data becomes stable. If and If both are positive or negative, then , is the same direction offset, the radial misalignment of the vertical surface of the spline pair: , the angular misalignment of the vertical surface of the spline pair: At this time, the axial misalignment of the vertical surface of the spline pair is is a small effect, and As mentioned above, the radial misalignment of the vertical plane Angular misalignment with the vertical surface of the spline pair Expression, so take ;like and If the value is opposite, , is the reverse offset, the radial misalignment of the vertical surface of the spline pair: , the vertical surface misalignment of the spline pair At this time, the axial misalignment of the vertical surface of the spline pair is The maximum influence is the axial misalignment of the vertical surface of the spline pair. ;

[0063] (9) The angular misalignment of the spline pair Record as matrix [ 、 ], the radial misalignment of the spline pair Record as matrix [ 、 ]、Axial misalignment of spline pair Record as matrix [ 、 ]; the angle misalignment of the spline pair , the radial misalignment of the spline pair , the axial misalignment of the spline pair , the torque , the transportation time , the carrying speed Combine them into the characteristic value of the spline pair carrying condition data and save them as the first sample size;

[0064] (10) Measure the mass before the test by weighing ; Quality after test ; Measured mass change ; Calculate wear rate ; and save the wear rate as the second sample size;

[0065] (11) Establish a three-dimensional model based on the geometric design parameters of the involute spline pair using the method, appropriately simplify the model based on the material mechanical parameters, perform finite element simulation analysis based on the characteristic values of the spline pair load data, and measure the axial slip distance , tangential slip distance , axial contact length , radial contact length ; Determine the friction coefficient based on the relevant material coefficient ;

[0066] (12) Combined with finite element simulation, the axial slip distance is obtained , tangential slip distance Calculate the relative slip between meshing teeth of a spline pair , combined with finite element simulation axial contact length , radial contact length Calculate the average contact pressure in the contact area of the tooth surfaces , through the relative slip between the meshing teeth of the spline pair and average contact pressure The Ruiz parameter is calculated to quantitatively predict the wear degree of the spline tooth surface, the maximum wear position and the estimated life.

[0067] Relative slip between meshing teeth of spline pair ; Average contact pressure ;

[0068] Ruiz parameters ; and save the Ruiz parameter as the third sample size;

[0069] (13) Repeat steps (5) to (13) to establish a first sample space based on the first sample size, the second sample size, and the third sample size, and divide the first sample space into a training set and a test set. The first sample space includes multiple sets of spline pair carrying data information; the first sample space includes data information outside the training set as the test set; and the percentage of the number of the test set to the number of the first sample space is a third preset value.

[0070] (14) Using a neural network to train the training set in the first sample space, a prediction model for predicting the spline pair wear rate and Ruiz parameter based on the characteristic value of the spline pair carrying condition data information is obtained; secondly, the characteristic value of the spline pair carrying condition data information contained in the test set is input into the prediction model to obtain the corresponding predicted value of the spline pair wear rate and Ruiz parameter; comparing the predicted value of the spline pair wear rate and Ruiz parameter with the standard error absolute value of the spline pair wear rate and Ruiz parameter contained in the test set, and judging whether the standard error absolute value of the prediction accuracy of the prediction model meets the fourth preset value, and the judgment condition is: if the standard error absolute value is greater than the fourth preset value, then continue to increase the number of the first sample space in step (11), and then execute the subsequent steps, and repeat the above steps until the standard error absolute value meets the fourth preset value. The loop ends when the value is less than or equal to the fourth preset value; if the predicted value of the spline pair wear rate and the Ruiz parameter predicted by the prediction model and the standard error absolute value of the spline pair wear rate and the Ruiz parameter in the test set are less than or equal to the fourth preset value, it is considered that the mapping relationship between the spline pair carrying condition data information and the wear rate, and the mapping relationship between the spline pair carrying condition data information and the Ruiz parameter are obtained; save the mapping relationship, and at the same time save the prediction model as a database; generate an expanded data set through random noise injection, parameter interpolation and extrapolation, and combined parameters, use the expanded data set to retrain the neural network to obtain the mapping relationship, and use the test set to test the new mapping relationship. If the standard error absolute value of the spline pair wear rate and the Ruiz parameter in the test set is less than or equal to the fourth preset value, then it is stored in the database to expand the database to obtain a new database;

[0071] (15) The characteristic values of the data information of the spline pair carrying condition during the test described in step (5) are directly input into the database described in step (14). The database will automatically predict the wear rate of the spline pair and the Ruiz parameter to determine whether the predicted wear is within the allowable range of the fifth preset value.

[0072] (16) The experimental data are input into the database described in (14), and the output parameter data distribution diagram, data comparison diagram, neural network diagram, error histogram of 20 intervals, error gradient diagram, training status diagram, working condition prediction result diagram and working condition debugging suggestion diagram are shown in the attached figure of the specification.

[0073] The specific embodiments of the present invention are described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the above embodiments. Various changes can be made within the scope of knowledge possessed by ordinary technicians in this field without departing from the scope of the present invention.

Claims

1. A method for predicting misalignment wear of spline pairs based on laser sensors, characterized by: The following steps are involved: (1) The method includes a horizontal adjustable laser sensor group, a vertical adjustable laser sensor group, and an auxiliary calibration mechanism; the horizontal adjustable laser sensor group includes a horizontal laser sensor 1, a horizontal laser sensor 2, a horizontal laser sensor 3, and a horizontal laser sensor 4; the vertical adjustable laser sensor group includes a vertical laser sensor 1, a vertical laser sensor 2, a vertical laser sensor 3, and a vertical laser sensor 4; the horizontal adjustable laser sensor group and the vertical adjustable laser sensor group are aligned with a rotation axis; the horizontal adjustable laser sensor group and the vertical adjustable laser sensor group each include four laser sensors, in order of position, the horizontal laser sensor 1 and the vertical laser sensor 1 are collectively referred to as laser sensor 1, the horizontal laser sensor 2 and the vertical laser sensor 2 are collectively referred to as laser sensor 2, the horizontal laser sensor 3 and the vertical laser sensor 3 are collectively referred to as laser sensor 3, and the horizontal laser sensor 4 and the vertical laser sensor 4 are collectively referred to as laser sensor 4; (2) When the rotating shaft is not working, measure the distance between the laser sensor 1 and the laser sensor 2 , measure the distance between the laser sensor 3 and the laser sensor 4 , the diameter of the rotating shaft , the distance between the laser sensor 1 and the laser sensor 4 , the distance between the laser sensor 2 and the laser sensor 3 ; The distance between the horizontal plane laser sensor 1 and the horizontal plane laser sensor 2 is The distance between the vertical laser sensor 1 and the vertical laser sensor 2 is The data in the horizontal adjustable laser sensor group are the corresponding suffix 1, and the data in the vertical adjustable laser sensor group are the corresponding suffix 2; the definitions of the remaining corresponding data are the same as 、 The same are 、 、 、 、 、 ; (3) Any time after the rotating shaft starts working is called Time, unit frequency is the first preset value, the frequency count is a positive integer Frequency; the laser sensor 1 is Read data at any time , the laser sensor 2 is Read data at any time , the laser sensor 3 is Read data at any time , the laser sensor 4 is Read data at any time , first calculation radius , Second calculation radius The average unit frequency data of the horizontal plane laser sensor 1 in the horizontal plane adjustable laser sensor group is: ; The unit frequency average data of the remaining corresponding data is calculated and The same are 、 、 、 、 、 、 、 、 、 ; (4) The z-axis is perpendicular to the horizontal plane, the x-axis is the axial direction, and the y-axis is in the direction that satisfies the spatial coordinate system conditions with the z-axis and the x-axis; when the rotating shaft is not working, the z-axis, x-axis, and y-axis positions of the horizontal position adjustable laser sensor group are calibrated; the y-axis position is adjusted so that the horizontal plane laser sensor 1, the horizontal plane laser sensor 2, the horizontal plane laser sensor 3, and the horizontal plane laser sensor 4 read data; the displacement of each direction of the auxiliary calibration mechanism is 0; the z-axis position is adjusted for measurement so that 、 、 、 , the error allowable value is the second preset value; adjust the y-axis fine-tuning module to measure, so that , the laser sensor readings are subtracted from each other satisfy ; Ensure that the Y-axis position and Z-axis position of the horizontal adjustable laser sensor group remain unchanged; measure the position coordinates of the four laser sensors through the axial radial grating ruler, so that ,; The displacement of the auxiliary calibration mechanism in each direction is 0, and the horizontal plane deflection angle is ; Adjust the x-axis fine-tuning module for measurement , ; The calibration is considered completed; (5) When the rotating shaft is not working, calibrate the z-axis, x-axis, and y-axis positions of the vertical position adjustable laser sensor group; adjust the z-axis position so that the vertical laser sensor 1, the vertical laser sensor 2, the vertical laser sensor 3, and the vertical laser sensor 4 read data; the displacement of the auxiliary calibration mechanism in each direction is 0; adjust the y-axis position for measurement so that 、 、 、 , the error allowable value is the second preset value; adjust the z-axis fine-tuning module to measure, so that , the laser sensor readings are subtracted from each other satisfy ; Ensure that the y-axis position and z-axis position of the vertical adjustable laser sensor group remain unchanged; measure the position coordinates of the four laser sensors through the axial radial grating ruler, so that ,; The displacement of the auxiliary calibration mechanism in each direction is 0, and the vertical deflection angle is ; Adjust the x-axis fine-tuning module for measurement , ; The calibration is considered completed; (6) When the rotating shaft is not working, the angle misalignment of the spline pair is adjusted by the auxiliary calibration mechanism and radial misalignment of spline pair is 0, and the axis position at this time is called the reference axis of the rotation axis; (7) When the rotating shaft is working, the horizontal position adjustable laser sensor group collects data in real time 、 、 、 Collect data in real time with the vertical position adjustable laser sensor group 、 、 、 ; Collect the load information of the spline pair during operation, including using the torque sensor to collect the torque during the spline pair load-bearing process , transportation time , Carrying speed ; (8) A simple traversal method is used for real-time calculation, and the unit frequency average data of the horizontal laser sensor 1 is The data flow is used Use it as a variable to record the current maximum value; Whenever new data arrives, it is compared with the current maximum value. If the new data is greater, the maximum value is updated; similarly, real-time calculation 、 、 、 、 、 、 The maximum value of the corresponding data flow , , 、 , , , The horizontal laser sensor 1 and the horizontal laser sensor 3 are offset relative to the reference axis of the rotation axis. , the vertical laser sensor 2 and the vertical laser sensor 4 are offset relative to the reference axis of the rotation axis The horizontal laser sensor 1 and the horizontal laser sensor 3 are offset relative to the reference axis of the rotation axis. The vertical laser sensor 2 and the vertical laser sensor 4 are offset relative to the reference axis of the rotation axis. ; (9) Calculate after the data becomes stable. If and If both are positive or negative, then , is the same direction offset, the radial misalignment of the spline pair horizontal plane: , horizontal plane angular misalignment of the spline pair: At this time, the axial misalignment of the spline pair horizontal plane is is a small effect, and As mentioned above, the radial misalignment of the horizontal plane Angular misalignment with the horizontal plane of the spline pair Expression, so take ;like and If the value is opposite, , is the reverse offset, the radial misalignment of the spline pair horizontal plane: , horizontal plane misalignment of spline pair At this time, the axial misalignment of the spline pair horizontal plane is The maximum influence is the axial misalignment of the spline pair horizontal plane. ; (10) Calculate after the data becomes stable. If and If both are positive or negative, then , is the same direction offset, the radial misalignment of the vertical surface of the spline pair: , the angular misalignment of the vertical surface of the spline pair: At this time, the axial misalignment of the vertical surface of the spline pair is is a small effect, and As mentioned above, the radial misalignment of the vertical plane Angular misalignment with the vertical surface of the spline pair Expression, so take ;like and If the value is opposite, , is the reverse offset, the radial misalignment of the vertical surface of the spline pair: , angular misalignment of the vertical surface of the spline pair At this time, the axial misalignment of the vertical surface of the spline pair is The maximum influence is the axial misalignment of the vertical surface of the spline pair. ; (11) The angular misalignment of the spline pair Record as matrix [ 、 ], the radial misalignment of the spline pair Record as matrix [ 、 ]、Axial misalignment of spline pair Record as matrix [ 、 ]; the angle misalignment of the spline pair , the radial misalignment of the spline pair , the axial misalignment of the spline pair , the torque , the transportation time , the carrying speed Combine them into the characteristic value of the spline pair carrying condition data and save them as the first sample size; (12) Measure the mass before the test by weighing ; Quality after test ; Measured mass change ; Calculate wear rate ; And save the wear rate as the second sample size; (13) Establish a three-dimensional model based on the geometric design parameters of the involute spline pair using the present method, appropriately simplify the model based on the material mechanical parameters, perform finite element simulation analysis based on the characteristic values of the spline pair's load data, and measure the axial slip distance. , tangential slip distance , axial contact length , radial contact length ; Determination of friction coefficient based on relevant material coefficients ; (14) Combined with finite element simulation, the axial slip distance is obtained , tangential slip distance Calculate the relative slip between meshing teeth of a spline pair , combined with finite element simulation axial contact length , radial contact length Calculate the average contact pressure in the contact area of the tooth surfaces , through the relative slip between the meshing teeth of the spline pair and average contact pressure Calculate Ruiz parameters to quantitatively predict the wear degree of spline tooth surface, the maximum wear position and estimate the life. ; Average contact pressure ; Ruiz parameters ; and save the Ruiz parameter as the third sample size; (15) Repeat steps (6) to (14) to establish a first sample space based on the first sample size, the second sample size, and the third sample size, and divide the first sample space into a training set and a test set. The first sample space includes multiple sets of spline pair carrying data information; the first sample space includes data information outside the training set as the test set; and the percentage of the number of the test set to the number of the first sample space is a third preset value. (16) Using a neural network to train the training set in the first sample space, a prediction model for predicting the spline pair wear rate and Ruiz parameter based on the characteristic value of the spline pair carrying condition data information is obtained; secondly, the characteristic value of the spline pair carrying condition data information contained in the test set is input into the prediction model to obtain the corresponding predicted value of the spline pair wear rate and Ruiz parameter; comparing the predicted value of the spline pair wear rate and Ruiz parameter with the standard error absolute value of the spline pair wear rate and Ruiz parameter contained in the test set, and judging whether the standard error absolute value of the prediction accuracy of the prediction model meets the fourth preset value, and the judgment condition is: if the standard error absolute value is greater than the fourth preset value, then continue to increase the number of the first sample space in step (11), and then execute the subsequent steps, and repeat the above steps until the standard error absolute value meets the fourth preset value. The loop ends when the value is less than or equal to the fourth preset value; if the predicted value of the spline pair wear rate and the Ruiz parameter predicted by the prediction model and the standard error absolute value of the spline pair wear rate and the Ruiz parameter in the test set are less than or equal to the fourth preset value, it is considered that the mapping relationship between the spline pair carrying condition data information and the wear rate, and the mapping relationship between the spline pair carrying condition data information and the Ruiz parameter are obtained; save the mapping relationship, and at the same time save the prediction model as a database; generate an expanded data set through random noise injection, parameter interpolation and extrapolation, and combined parameters, use the expanded data set to retrain the neural network to obtain the mapping relationship, and use the test set to test the new mapping relationship. If the standard error absolute value of the spline pair wear rate and the Ruiz parameter in the test set is less than or equal to the fourth preset value, then it is stored in the database to expand the database to obtain a new database; (17) The characteristic values of the data information of the spline pair carrying condition during the test described in step (7) are directly input into the database described in step (16). The database will automatically predict the wear rate of the spline pair and the Ruiz parameter to determine whether the predicted wear is within the allowable range of the fifth preset value.

2. When the working conditions are limited, only the horizontal position adjustable laser sensor group can be selected for use. The angle misalignment of the spline pair during operation Record as matrix [ ], the radial misalignment Record as matrix [ ], after the data becomes stable, the horizontal plane misalignment of the spline pair is approximately , the radial misalignment of the spline pair horizontal plane is approximately .

3. When the working conditions are limited, only the vertical position adjustable laser sensor group can be selected for use, and the angle misalignment of the spline pair during operation can be adjusted. Record as matrix [ ], the radial misalignment Record as matrix [ ], after stabilization, the vertical angle misalignment of the spline pair is calculated. , Vertical radial misalignment .

Images