Bearing load measuring method based on multi-sensor data fusion
Through the multi-sensor data fusion technology combined with strain method and lift method, multiple sets of strain gauges, overall least squares method and Bayesian fusion technology, the error and deviation problems in bearing load measurement are solved, and the measurement results are achieved with higher accuracy .
Patent Information
- Application Number
- CN202411267472.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-11
- Publication Date
- 2024-11-08
AI Technical Summary
The existing bearing load measurement methods have the problem that the measured value is different from the real value, and the parameter selection deviation in each method leads to deviation of the measurement results, and the data quality is not high.
Multi-sensor data fusion technology is adopted, combined with strain method and overhead lift method, strain method measurement is performed through a combination of multiple sets of strain gauges, and data is processed using the overall least squares method and Bayesian fusion technology to improve measurement accuracy.
It realizes more accurate bearing load measurement, effectively eliminates errors, and improves measurement accuracy and data quality.
Smart Images

Figure CN118913502A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of shaft bearing load testing, and in particular to a method for accurately measuring bearing load. Background Art
[0002] Bearing load measurement is an important basis for inspecting the quality of shaft alignment. At present, the commonly used bearing load measurement method is the jacking method. In addition, the strain method is also commonly used to measure bearing load. The jacking method uses a jacking device to jack up the shaft at the designed jacking point, and simultaneously measures the jacking force and jacking displacement during the process. The jacking load is obtained by drawing a regression line graph using the measured data, and then multiplying it by the jacking coefficient given in the alignment calculation book to obtain the bearing load. The strain method pastes strain gauges at the designed section positions on the shaft segments at both ends of the measured bearing, measures the section strain through the strain gauge, and then obtains the section bending moment, and obtains the bearing load through the moment balance equation. The measurement data of the jacking method and the strain method are limited by the current technical level of the measurement personnel and the measurement process, and it is easy to have a large difference between the measured value and the true value. In addition, the parameters used in each method, such as the jacking coefficient, the selection of the assumed support point position of the bearing, etc., may also have a certain deviation compared with the actual state, resulting in deviation in the measurement results. In summary, due to the existence of various errors in the bearing load measurement method, there will be data quality problems when measuring bearing load using traditional technology.
[0003] Chinese patent CN112816001A discloses a method for synchronously testing the load and displacement of each bearing of a multi-support shaft system based on a jacking process. The method combines the three methods of electronic dynamometer load measurement, jacking method and resistance strain gauge method. On the basis of only one jacking measurement, the other two methods are also used for measurement. Finally, the different load values measured by each method are fused, which can improve the accuracy of each bearing load test. At the same time, the entire jacking process can obtain more bearing load and displacement data reflecting the characteristics of the shaft system. However, this method still has the following disadvantages: it only fuses the measurement results of each method, and the data is not fully utilized; in addition, the accuracy of the measurement results of each method varies greatly, and the accuracy of the results obtained by direct fusion is limited. Summary of the invention
[0004] The main purpose of the present invention is to provide a bearing load test method based on multi-sensor data fusion, which combines the jacking method and the strain method. For the measurement of a single bearing load, the strain method is first measured by combining multiple sets of strain gauges, and the strain method measurement data is fused and processed by multi-sensor data fusion technology to improve the accuracy of the strain method measurement results; then the jacking method is used for measurement, and the strain method measurement is performed synchronously during the jacking measurement, and the strain data and jacking data recorded during the jacking method measurement are fused and processed by multi-sensor data fusion technology to improve the accuracy of the jacking method measurement results; finally, all the measurement data are fused and processed using multi-sensor fusion technology to obtain more accurate measurement results and improve the bearing load measurement accuracy.
[0005] The technical solution adopted by the present invention is:
[0006] A bearing load measurement method based on multi-sensor data fusion comprises the following steps:
[0007] S1. Debugging and installation of strain measurement components: Install a set of strain gauges symmetrically at the design section position, and connect them to the strain gauge according to the full-bridge or half-bridge wiring method;
[0008] S2. Perform a strain measurement by turning the wheel, record the strain value of each measuring point, and obtain the probability density function corresponding to the strain value of each measuring point;
[0009] S3, turn the shaft to adjust it until the strain gauge is in the vertical plane;
[0010] S4. Debugging and installation of jacking method measuring components: Install the jacking device and jacking displacement measuring device at the designed jacking point;
[0011] S5, perform a jacking method measurement, record the jacking force and jacking displacement after the bearing is emptied, and simultaneously record the strain value of each measuring point measured by the strain method after the bearing is emptied, and obtain the probability density function corresponding to the strain value of each measuring point;
[0012] S6, converting the strain value obtained in step S2 into a bearing fulcrum force and obtaining a probability density function of the bearing fulcrum force, and converting the strain value obtained in step S5 into a jacking fulcrum force and obtaining a probability density function of the jacking fulcrum force;
[0013] S7, fitting the jacking fulcrum force obtained in step S6 and the jacking displacement obtained in step S5 by the overall least squares method to obtain the jacking fulcrum load and its probability density function corresponding to the jacking displacement being 0, and fitting the jacking force and the jacking displacement obtained in step S5 by the overall least squares method to obtain the jacking load and its probability density function corresponding to the jacking displacement being 0;
[0014] S8, performing Bayesian fusion on the probability density function of the top-lift support load obtained in step S7 and the probability density function of the top-lift load to obtain a fused top-lift load and its probability density function;
[0015] S9, performing data conversion on the probability density function of the fused top lift load to obtain a probability density function of the top lift bearing force;
[0016] S10. Perform Bayesian fusion on the probability density function of the bearing support force and the probability density function of the lifting bearing force to obtain the fused bearing load.
[0017] In the above scheme, in step S1, the number of design sections measured by the strain method is M, M is an even number and M>3, the design sections should be symmetrically distributed on both sides of the measured bearing, the distance between the two design sections closest to the measured bearing and the measured bearing should be greater than 1m, and the distance between any two design sections should be greater than 1m.
[0018] In the above scheme, in steps S2 and S5, the systematic error and random error should be considered, and the obtained strain value measurement data of each measuring point is converted into the probability density function form of the strain value of each measuring point. The measurement equation of the measurement data is expressed as:
[0019]
[0020] Where: ε represents the measured data, represents the true value of the measurement, k is the systematic error coefficient, and ω is the random error;
[0021] Assume that the systematic error coefficient of the measured data follows a normal distribution Random errors follow a normal distribution According to the maximum likelihood principle, the probability density function of the measured data is:
[0022]
[0023] In the formula: x represents the independent variable, μ ε is the mean of the measured data,
[0024] In the above scheme, in step S6, the method of converting the measured strain value into the bearing fulcrum force or the lifting fulcrum force includes the following steps:
[0025] S6.1. First, calculate the measured strain values of each section to obtain the bending moment of each section, and then obtain Section bending moment combinations, M represents the number of design sections measured by the strain method, M is an even number and M>3, eliminating the section bending moment combinations that cannot be calculated by the force and moment equilibrium conditions, and a total of N groups of measurement results and their probability density functions can be obtained through the force and moment equilibrium equation;
[0026] S6.2, let p i (x) is the probability density function of the i-th group of measurement results, i = [1, 2, ..., N], and defines a distance measure matrix d ij , use Wasserstein distance to calculate the distance between the probability density functions of each group of measurement results:
[0027] d ij =W(p i ,p j ) (3)
[0028] Where W represents the Wasserstein distance, p i represents the probability density function of the i-th group of measurement results, p j represents the probability density function of the jth group of measurement results, i,j = [1,2,…,N];
[0029] S6.3. Define the mutual support s of each group of measurement results i The greater the difference between the probability density functions of the two sets of measurement results, the greater the distance measure, the lower the mutual support, and the mutual support s i Defined as:
[0030]
[0031] In the formula, s i represents the mutual support of the i-th group of measurement results;
[0032] S6.4. The higher the mutual support of each group of measurement results, the higher the corresponding fusion weight ω i The larger the value, the fusion weight can be obtained from the mutual support:
[0033]
[0034] S6.5, the final fusion result is:
[0035]
[0036] Where P(x) represents the probability density function of the final fusion result, that is, the probability density function of the bearing fulcrum force / lifting fulcrum force; the maximum value of the probability density function corresponds to the bearing fulcrum force / lifting fulcrum force.
[0037] In the above scheme, in step S7, the method of fitting the jacking fulcrum force / jacking force and the jacking displacement by the overall least square method to obtain the corresponding jacking fulcrum load / jacking load and its probability density function when the jacking displacement is 0 includes the following steps:
[0038] S7.1. Perform the overall least squares linear fitting and obtain the following linear regression equation:
[0039] y p =a+bx s (7)
[0040] Among them, x s is the lifting displacement, y p is the lifting force or the lifting fulcrum force, a represents the constant term of the fitting, and b represents the fitting coefficient;
[0041] When x s = 0, the lifting force or the lifting fulcrum force is It follows a normal distribution
[0042] Here, the least squares estimates of a and b and σ r The unbiased estimator of is:
[0043]
[0044] In the formula, represents the fitting variance, n represents the total number of data points measured during the lifting and emptying process, t represents the tth data point, and x St represents the lifting displacement corresponding to the tth data point, It represents the lifting force or lifting support force corresponding to the t-th data point, represents the average lifting displacement of n data points, It represents the average lifting force or lifting fulcrum force of n data points;
[0045] The fitting results show the top load P0' and top support load P0" when the top displacement is 0;
[0046] S7.2, Assume that the system error coefficient of the lifting force is Assume the variance of the lifting support force is The probability density functions of the top-lift load P0' and the top-lift support load P0" are:
[0047]
[0048] In the formula, is the probability density function of the top lift load, is the probability density function of the top-lift support load, σ' r 2 is the fitted variance of the top load, σ' r ' 2 is the fitted variance of the lift support load.
[0049] In the above scheme, in step S8, the Bayesian fusion method uses the probability density function of the top-lift support load as the prior function and the probability density function of the top-lift load as the likelihood, and the probability density function of the fused top-lift load is:
[0050]
[0051] In the formula, represents the probability density function of the fused top-lift load, and the maximum value of the probability density function of the fused top-lift load corresponds to the fused top-lift load; represents the probability density function of the top lift load; Represents the probability density function of the top lift support load.
[0052] In the above scheme, in step S9, the method of data conversion is:
[0053] The lifting coefficient given in the calibration calculation book should take into account the systematic error and be converted into a probability density function. Suppose the theoretical lifting coefficient given in the calibration calculation book is K, and the actual lifting coefficient is Assumptions Then the probability density function of the jacking load is multiplied by the probability density function of the jacking coefficient to obtain the probability density function of the jacking bearing force:
[0054]
[0055] In the formula, is the probability density function of the top-lift bearing force, P0 is the fusion top-lift load, is the variance of the probability density function of the fused top lift load.
[0056] In the above scheme, in step S10, the Bayesian fusion method uses the probability density function of the bearing support force as the prior function and the probability density function of the lifting bearing force as the likelihood, and the probability density function of the fused bearing load is:
[0057]
[0058] In the formula, f F (x) represents the probability density function of the fused bearing load, and the maximum value of the probability density function of the fused bearing load corresponds to the fused bearing load; f FS (x) represents the probability density function of the bearing fulcrum force; The probability density function representing the lift bearing force.
[0059] In the above scheme, in step S2, the strain measurement specifically includes the following steps:
[0060] S2.1. Turn the shaft system so that the two resistance strain gauges at each measuring point are placed on a vertical plane;
[0061] S2.2. Adjust each measuring point of the strain gauge to zero balance;
[0062] S2.3, turn the shaft system 90°, measure and record the strain value of each measuring point;
[0063] S2.4. Turn the shaft system by turning the gear to 180°, 270°, and 360° in sequence, and measure and record the strain value of each measuring point respectively;
[0064] S2.5. Reverse the rotation of the shaft system to 270°, 180°, 90°, and 0°, and measure and record the strain value at each measuring point respectively.
[0065] In the above scheme, in step S5, a top-lifting measurement specifically includes the following steps:
[0066] S5.1. The lifting device gradually lifts the shaft at a certain lifting displacement until the bearing and the journal are free. After the shaft is free, the shaft is lifted for a certain distance at a smaller lifting displacement, but it should be ensured that the upper bearing shell is not pushed up.
[0067] S5.2. The lifting device slowly and gradually descends with the same lifting displacement as the lifting rises, until the lifting device is completely free of force and the shaft returns to its original position.
[0068] The beneficial effects produced by the present invention are:
[0069] The measurement method of the present invention is aimed at the measurement of a single bearing load, and uses multi-sensor data fusion technology to fuse the top-lift method and the strain method based on the probability density function to achieve the purpose of improving the bearing load measurement accuracy. First, the strain method measurement is performed by combining multiple sets of strain gauges, and the strain method measurement data is fused using the multi-sensor data fusion technology. The measured strain value is converted into the bearing fulcrum force and its probability density function. The strain information obtained by the combination of multiple sets of strain gauges to obtain the bearing fulcrum force can effectively eliminate the error caused by the assumed position offset of the equivalent fulcrum, thereby improving the measurement accuracy of the strain method. Then, the strain method measurement is performed synchronously during the top-lift method measurement process, and the probability density function of the top-lift fulcrum load obtained by the strain method is Bayesian fused with the probability density function of the top-lift load obtained by the top-lift method to obtain the fused top-lift load and its probability density function, thereby achieving the purpose of improving the measurement accuracy of the top-lift force by combining strain information. Then, the uncertainty of the top-lift coefficient is taken into account in the measurement result of the top-lift method to obtain the probability density function of the top-lift bearing force. Finally, the probability density function of the bearing fulcrum force after the strain method fusion is fused with the probability density function of the top-lift bearing force to further improve the bearing load measurement accuracy. The method of the present invention adopts a two-level parallel fusion structure. The first-level fusion improves the measurement accuracy of each measurement method, and the second-level fusion is to fuse the first-level fusion results to further improve the measurement accuracy. The fusion process fully considers the uncertainty of the measurement data obtained by each measurement method, and deeply explores the relationship between the measurement data, thereby achieving the purpose of improving the bearing load measurement accuracy. BRIEF DESCRIPTION OF THE DRAWINGS
[0070] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0071] Figure 1 is a test element arrangement diagram of a bearing load measurement method based on multi-sensor data fusion in an embodiment of the present invention;
[0072] Figure 2 It is a test flow chart of the method of the present invention;
[0073] Figure 3 It is a data processing flow chart of the method of the present invention.
[0074] In the figure: 1. strain gauge; 2. lift displacement measuring device; 3. lift force measuring device; 4. measured bearing; 5-axis. DETAILED DESCRIPTION
[0075] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.
[0076] It should be noted that the illustrations provided in the embodiments of the present invention are only used to illustrate the basic concept of the present invention in a schematic manner. Therefore, the drawings only show components related to the present invention rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout type may also be more complicated.
[0077] In the present invention, it is also necessary to explain that, if the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. appear, the orientation or position relationship indicated is based on the orientation or position relationship shown in the drawings, which is only for the convenience of describing the present application and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present application. In addition, if the terms "first" and "second" appear, they are only used for description and distinction purposes, and cannot be understood as indicating or implying relative importance.
[0078] The present invention proposes a bearing load measurement method based on multi-sensor data fusion, comprising the following steps:
[0079] S1. Debugging and installation of strain measurement components: Install a set of strain gauges 1 symmetrically above and below the designed cross-section position, and connect them to the strain gauge according to the full-bridge or half-bridge wiring method.
[0080] The number of design sections measured by the strain method is M, M is an even number and M>3, the design sections should be symmetrically distributed on both sides of the measured bearing 4, the distance between the two design sections closest to the measured bearing 4 and the measured bearing 4 should be greater than 1m, and the distance between any two design sections should be greater than 1m. In this embodiment, the test elements are arranged as follows Figure 1 As shown, the strain method includes 4 measurement sections, ie, M=4.
[0081] Each group of strain gauges 1 should be installed symmetrically at the designed section position, and ensure that the pasting parts of the strain gauges of all strain sections are in the same plane; the pasting parts should be free of surface defects and the surface should be polished, and moisture-proof treatment should be done after pasting; after pasting, it should be turned 2 to 3 times in both directions to eliminate the internal stress of pasting; then connect it to the strain gauge with shielded wire according to the full-bridge wiring method.
[0082] S2. Perform a strain measurement by turning the wheel, record the strain value of each measuring point, and obtain the probability density function corresponding to the strain value of each measuring point.
[0083] Specifically, the strain method measures the strain value of the designed section, which specifically includes the following steps:
[0084] S2.1. Turn the shaft system so that the two resistance strain gauges at each measuring point are placed on a vertical plane;
[0085] S2.2. Adjust each measuring point of the strain gauge to zero balance;
[0086] S2.3, turn the shaft system 90°, measure and record the strain value of each measuring point;
[0087] S2.4. Turn the shaft system by turning the gear to 180°, 270°, and 360° in sequence, and measure and record the strain value of each measuring point respectively;
[0088] S2.5. Reverse the rotation of the shaft system to 270°, 180°, 90°, and 0°, and measure and record the strain value at each measuring point respectively.
[0089] S3. Turn the crank to adjust the shaft until the strain gauge is in the vertical plane.
[0090] S4. Debugging and installation of measuring components by the jacking method: Install the jacking device 3 and the jacking displacement measuring device 2 at the designed jacking point position.
[0091] Specifically, the lifting displacement measuring device 2 adopts a lifting displacement sensor, and its supporting position should ensure that it is not affected by the lifting device 3; the lifting displacement sensor is adjusted to the zero point, at which time the lifting device 3 should just contact the shaft, but the lifting force is zero.
[0092] S5. Perform a jacking method measurement, record the jacking force and jacking displacement after the bearing is emptied, and simultaneously record the strain value of each measuring point measured by the strain method after the bearing is emptied, and obtain the probability density function corresponding to the strain value of each measuring point.
[0093] Specifically, a top-lift measurement includes the following steps:
[0094] S5.1. The lifting device 3 gradually lifts the shaft 5 at a certain lifting displacement until the bearing and the journal are free. After the shaft is free, the shaft is lifted for a distance at a smaller lifting displacement, but it should be ensured that the upper bearing shell is not pushed up.
[0095] S5.2, the lifting device 3 is slowly and gradually lowered with the same lifting displacement as the lifting rise, until the lifting device 3 is completely free of force and the shaft returns to its original position. During this period, the lifting force, lifting displacement, and cross-sectional strain are recorded.
[0096] In steps S2 and S5, both systematic errors and random errors should be considered, and the obtained strain value measurement data of each measuring point are converted into the probability density function form of the strain value of each measuring point. The measurement equation of the measurement data is expressed as:
[0097]
[0098] Where: ε represents the measured data, represents the true value of the measurement, k is the systematic error coefficient, and ω is the random error;
[0099] Assume that the systematic error coefficient of the measured data follows a normal distribution Random errors follow a normal distribution According to the maximum likelihood principle, the probability density function of the measured data is:
[0100]
[0101] In the formula: x represents the independent variable, μ ε is the mean of the measured data,
[0102] S6, converting the strain value obtained in step S2 into a bearing fulcrum force and obtaining a probability density function of the bearing fulcrum force, and converting the strain value obtained in step S5 into a jacking fulcrum force and obtaining a probability density function of the jacking fulcrum force. The purpose of this step is to first fuse the strain method measurement data through multi-sensor data fusion technology to improve the accuracy of the strain method measurement results.
[0103] Specifically, the method of converting the measured strain value into the bearing fulcrum force or the lifting fulcrum force includes the following steps:
[0104] S6.1. First, calculate the measured strain values of each section to obtain the bending moment of each section, and then obtain Section bending moment combinations, M represents the number of design sections measured by the strain method, M is an even number and M>3, eliminating the section bending moment combinations that cannot be calculated by the force and moment equilibrium conditions, and a total of N groups of measurement results and their probability density functions can be obtained through the force and moment equilibrium equation;
[0105] S6.2, let p i (x) is the probability density function of the i-th group of measurement results, i = [1, 2, ..., N], and defines a distance measure matrix d ij , use Wasserstein distance to calculate the distance between the probability density functions of each group of measurement results:
[0106] d ij =W(p i ,p j ) (3)
[0107] Where W represents the Wasserstein distance, p i represents the probability density function of the i-th group of measurement results, p jrepresents the probability density function of the jth group of measurement results, i,j = [1,2,…,N];
[0108] S6.3. Define the mutual support s of each group of measurement results i The greater the difference between the probability density functions of the two sets of measurement results, the greater the distance measure, the lower the mutual support, and the mutual support s i Defined as:
[0109]
[0110] In the formula, s i represents the mutual support of the i-th group of measurement results;
[0111] S6.4. The higher the mutual support of each group of measurement results, the higher the corresponding fusion weight ω i The larger the value, the fusion weight can be obtained from the mutual support:
[0112]
[0113] S6.5, the final fusion result is:
[0114]
[0115] Where P(x) represents the probability density function of the final fusion result, that is, the probability density function of the bearing fulcrum force / lifting fulcrum force; the maximum value of the probability density function corresponds to the bearing fulcrum force / lifting fulcrum force.
[0116] S7. Fit the jacking fulcrum force obtained in step S6 and the jacking displacement obtained in step S5 by the overall least squares method to obtain the jacking fulcrum load and its probability density function corresponding to the jacking displacement being 0, and fit the jacking force and the jacking displacement obtained in step S5 by the overall least squares method to obtain the jacking load and its probability density function corresponding to the jacking displacement being 0.
[0117] Specifically, the method of fitting the jacking fulcrum force / jacking force and the jacking displacement by the overall least square method to obtain the corresponding jacking fulcrum load / jacking load and its probability density function when the jacking displacement is 0 includes the following steps:
[0118] S7.1. Perform the overall least squares linear fitting and obtain the following linear regression equation:
[0119] y p =a+bx s (7)
[0120] Among them, x s is the lifting displacement, y p is the lifting force or the lifting fulcrum force, a represents the constant term of the fitting, and b represents the fitting coefficient;
[0121] When x s = 0, the lifting force or the lifting fulcrum force is It follows a normal distribution
[0122] Here, the least squares estimates of a and b and σ r The unbiased estimator of is:
[0123]
[0124] In the formula, represents the fitting variance, n represents the total number of data points measured during the lifting and emptying process, t represents the tth data point, and x St represents the lifting displacement corresponding to the t-th data point, It represents the lifting force or lifting support force corresponding to the t-th data point, represents the average lifting displacement of n data points, It represents the average lifting force or lifting fulcrum force of n data points;
[0125] The fitting results show the top load P0' and top support load P0" when the top displacement is 0;
[0126] S7.2, Assume that the system error coefficient of the lifting force is Assume that the variance of the lifting support force is The probability density functions of the top-lift load P0' and the top-lift support load P0" are:
[0127]
[0128] In the formula, is the probability density function of the top lift load, is the probability density function of the top-lift support load, σ' r 2 is the fitted variance of the top load, σ' r ' 2 is the fitted variance of the lift support load.
[0129] S8. Perform Bayesian fusion on the probability density function of the top-lift support load obtained in step S7 and the probability density function of the top-lift load to obtain a fused top-lift load and its probability density function.
[0130] The method of the present invention adopts the strain method and the jacking method to measure the force at the jacking fulcrum, and performs Bayesian fusion on the measurement data of the two methods through S7-S8, so as to improve the measurement accuracy of the jacking force in the jacking method measurement of the de-airing process.
[0131] Specifically, the Bayesian fusion method uses the probability density function of the top-lift support load as the prior function and the probability density function of the top-lift load as the likelihood. The probability density function of the fused top-lift load is:
[0132]
[0133] In the formula, represents the probability density function of the fused top-lift load, and the maximum value of the probability density function of the fused top-lift load corresponds to the fused top-lift load; represents the probability density function of the top lift load; Represents the probability density function of the top lift support load.
[0134] S9. Perform data conversion on the probability density function of the fused top-lift load to obtain the probability density function of the top-lift bearing force.
[0135] Specifically, the data conversion method is:
[0136] The lifting coefficient given in the calibration calculation book should take into account the systematic error and be converted into a probability density function. Suppose the theoretical lifting coefficient given in the calibration calculation book is K, and the actual lifting coefficient is Assumptions Then the probability density function of the jacking load is multiplied by the probability density function of the jacking coefficient to obtain the probability density function of the jacking bearing force:
[0137]
[0138] In the formula, is the probability density function of the top-lift bearing force, P0 is the fusion top-lift load, is the variance of the probability density function of the fused top lift load.
[0139] The function of step S9 is to take the uncertainty of the lifting coefficient into consideration in the probability density function of the measurement result of the lifting method.
[0140] S10. Perform Bayesian fusion on the probability density function of the bearing support force and the probability density function of the lifting bearing force to obtain the fused bearing load.
[0141] Specifically, the Bayesian fusion method uses the probability density function of the bearing support force as the prior function and the probability density function of the top bearing force as the likelihood. The probability density function of the fused bearing load is:
[0142] f F (x)∝f FS (x)f FJ (x)(12)
[0143] In the formula, f F(x) represents the probability density function of the fused bearing load, and the maximum value of the probability density function of the fused bearing load corresponds to the fused bearing load; f FS (x) represents the probability density function of the bearing support force; f FJ (x) represents the probability density function of the lift bearing force.
[0144] The purpose of step S10 is to fuse the measurement results obtained by fusing the strain method and the top-lift method, so as to further improve the measurement accuracy.
[0145] It should be pointed out that, according to the needs of implementation, the various steps / components described in this application can be split into more steps / components, and two or more steps / components or partial operations of steps / components can be combined into new steps / components to achieve the purpose of the present invention.
[0146] The order of execution of each step in the above embodiment does not mean the order of execution. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present application.
[0147] It should be understood that those skilled in the art can make improvements or changes based on the above description, and all these improvements and changes should fall within the scope of protection of the appended claims of the present invention.
Claims
1. A bearing load measurement method based on multi-sensor data fusion, characterized in that: The following steps are involved: S1. Debugging and installation of strain measurement components: Install a set of strain gauges symmetrically at the design section position, and connect them to the strain gauge according to the full-bridge or half-bridge wiring method; S2. Perform a strain measurement by turning the wheel, record the strain value of each measuring point, and obtain the probability density function corresponding to the strain value of each measuring point; S3, turn the shaft to adjust it until the strain gauge is in the vertical plane; S4. Debugging and installation of jacking method measuring components: Install the jacking device and jacking displacement measuring device at the designed jacking point; S5, perform a jacking method measurement, record the jacking force and jacking displacement after the bearing is emptied, and simultaneously record the strain value of each measuring point measured by the strain method after the bearing is emptied, and obtain the probability density function corresponding to the strain value of each measuring point; S6, converting the strain value obtained in step S2 into a bearing fulcrum force and obtaining a probability density function of the bearing fulcrum force, and converting the strain value obtained in step S5 into a jacking fulcrum force and obtaining a probability density function of the jacking fulcrum force; S7, fitting the jacking fulcrum force obtained in step S6 and the jacking displacement obtained in step S5 by the overall least squares method to obtain the jacking fulcrum load and its probability density function corresponding to the jacking displacement being 0, and fitting the jacking force and the jacking displacement obtained in step S5 by the overall least squares method to obtain the jacking load and its probability density function corresponding to the jacking displacement being 0; S8, performing Bayesian fusion on the probability density function of the top-lift support load obtained in step S7 and the probability density function of the top-lift load to obtain a fused top-lift load and its probability density function; S9, performing data conversion on the probability density function of the fused top lift load to obtain a probability density function of the top lift bearing force; S10. Perform Bayesian fusion on the probability density function of the bearing support force and the probability density function of the lifting bearing force to obtain a fused bearing load.
2. The bearing load measurement method based on multi-sensor data fusion according to claim 1 is characterized in that: In step S1, the number of design sections measured by the strain method is M, where M is an even number and M>3. The design sections should be symmetrically distributed on both sides of the measured bearing. The distance between the two design sections closest to the measured bearing and the measured bearing should be greater than 1m, and the distance between any two design sections should be greater than 1m.
3. The bearing load measurement method based on multi-sensor data fusion according to claim 1 is characterized in that: In steps S2 and S5, the systematic error and random error should be considered, and the obtained strain value measurement data of each measuring point is converted into the probability density function form of the strain value of each measuring point. The measurement equation of the measurement data is expressed as: Where: ε represents the measured data, represents the true value of the measurement, k is the systematic error coefficient, and ω is the random error; Assume that the systematic error coefficient of the measured data follows a normal distribution Random errors follow a normal distribution According to the maximum likelihood principle, the probability density function of the measured data is: In the formula: x represents the independent variable, μ ε is the mean of the measured data, 4. The bearing load measurement method based on multi-sensor data fusion according to claim 1 is characterized in that: In step S6, the method of converting the measured strain value into the bearing fulcrum force or the lifting fulcrum force includes the following steps: S6.
1. First, calculate the measured strain values of each section to obtain the bending moment of each section, and then obtain Section bending moment combinations, M represents the number of design sections measured by the strain method, M is an even number and M>3, eliminating the section bending moment combinations that cannot be calculated by the force and moment equilibrium conditions, and a total of N groups of measurement results and their probability density functions can be obtained through the force and moment equilibrium equation; S6.2, let p i (x) is the probability density function of the i-th group of measurement results, i = [1, 2, ..., N], and defines a distance measure matrix d ij , use Wasserstein distance to calculate the distance between the probability density functions of each group of measurement results: d ij =W(p i ,p j ) (3) Where W represents the Wasserstein distance, p i represents the probability density function of the i-th group of measurement results, p j represents the probability density function of the jth group of measurement results, i,j = [1,2,…,N]; S6.
3. Define the mutual support s of each group of measurement results i The greater the difference between the probability density functions of the two sets of measurement results, the greater the distance measure, the lower the mutual support, and the mutual support s i Defined as: In the formula, s i represents the mutual support of the i-th group of measurement results; S6.
4. The higher the mutual support of each group of measurement results, the higher the corresponding fusion weight ω i The larger the value, the fusion weight can be obtained from the mutual support: S6.5, the final fusion result is: Where P(x) represents the probability density function of the final fusion result, that is, the probability density function of the bearing fulcrum force / lifting fulcrum force; the maximum value of the probability density function corresponds to the bearing fulcrum force / lifting fulcrum force.
5. The bearing load measurement method based on multi-sensor data fusion according to claim 1 is characterized in that: In step S7, the method of fitting the jacking fulcrum force / jacking force and the jacking displacement by the overall least square method to obtain the corresponding jacking fulcrum load / jacking load and its probability density function when the jacking displacement is 0 includes the following steps: S7.
1. Perform the overall least squares linear fitting and obtain the following linear regression equation: y p =a+bx s (7) Among them, x s is the lifting displacement, y p is the lifting force or the lifting fulcrum force, a represents the constant term of the fitting, and b represents the fitting coefficient; When x s = 0, the lifting force or the lifting fulcrum force is It follows a normal distribution Here, the least squares estimates of a and b and σ r The unbiased estimator of is: In the formula, represents the fitting variance, n represents the total number of data points measured during the lifting and emptying process, t represents the tth data point, and x St represents the lifting displacement corresponding to the tth data point, It represents the lifting force or lifting support force corresponding to the t-th data point, represents the average lifting displacement of n data points, It represents the average lifting force or lifting fulcrum force of n data points; The fitting results show the top load P0' and top support load P0" when the top displacement is 0; S7.2, Assume that the system error coefficient of the lifting force is Assume that the variance of the lifting support force is The probability density functions of the top-lift load P0' and the top-lift support load P0" are: In the formula, is the probability density function of the top lift load, is the probability density function of the top-lift support load, σ' r 2 is the fitted variance of the top load, σ' r ' 2 is the fitted variance of the lift support load.
6. The bearing load measurement method based on multi-sensor data fusion according to claim 1 is characterized in that: In step S8, the Bayesian fusion method uses the probability density function of the top-lift support load as the prior function and the probability density function of the top-lift load as the likelihood. The probability density function of the fused top-lift load is: In the formula, represents the probability density function of the fused top-lift load, and the maximum value of the probability density function of the fused top-lift load corresponds to the fused top-lift load; represents the probability density function of the top lift load; Represents the probability density function of the top lift support load.
7. The bearing load measurement method based on multi-sensor data fusion according to claim 1 is characterized in that: In step S9, the method of data conversion is: The lifting coefficient given in the calibration calculation book should take into account the systematic error and be converted into a probability density function. Suppose the theoretical lifting coefficient given in the calibration calculation book is K, and the actual lifting coefficient is Assumptions Then the probability density function of the jacking load is multiplied by the probability density function of the jacking coefficient to obtain the probability density function of the jacking bearing force: In the formula, f FJ (x) is the probability density function of the top-lift bearing force, P0 is the fusion top-lift load, is the variance of the probability density function of the fused top lift load.
8. The bearing load measurement method based on multi-sensor data fusion according to claim 1 is characterized in that: In step S10, the Bayesian fusion method uses the probability density function of the bearing support force as the prior function and the probability density function of the lifting bearing force as the likelihood. The probability density function of the fused bearing load is: f F (x)∝f FS (x)f FJ (x)(12) In the formula, f F (x) represents the probability density function of the fused bearing load, and the maximum value of the probability density function of the fused bearing load corresponds to the fused bearing load; f FS (x) represents the probability density function of the bearing support force; f FJ (x) represents the probability density function of the lift bearing force.
9. The bearing load measurement method based on multi-sensor data fusion according to claim 1, characterized in that: In step S2, the strain measurement specifically includes the following steps: S2.
1. Turn the shaft system so that the two resistance strain gauges at each measuring point are placed on a vertical plane; S2.
2. Adjust each measuring point of the strain gauge to zero balance; S2.3, turn the shaft system 90°, measure and record the strain value of each measuring point; S2.
4. Turn the shaft system by turning the gear to 180°, 270°, and 360° in sequence, and measure and record the strain value of each measuring point respectively; S2.
5. Reverse the rotation of the shaft system to 270°, 180°, 90°, and 0°, and measure and record the strain value at each measuring point respectively.
10. The bearing load measurement method based on multi-sensor data fusion according to claim 1, characterized in that: In step S5, a top-lift measurement specifically includes the following steps: S5.
1. The lifting device gradually lifts the shaft at a certain lifting displacement until the bearing and the journal are free. After the shaft is free, the shaft is lifted for a certain distance at a smaller lifting displacement, but it should be ensured that the upper bearing shell is not pushed up. S5.
2. The lifting device slowly and gradually descends with the same lifting displacement as the lifting rises, until the lifting device is completely free of force and the shaft returns to its original position.
Citation Information
Patent Citations
Method for synchronously testing load displacement of each bearing of multi-supporting shaft system based on jacking process
CN112816001A
