Vibration monitoring system based on transfer function

Abstract

The invention belongs to the field of state monitoring, and particularly discloses a vibration monitoring system based on a transfer function, and the system comprises a data obtaining module which is used for obtaining a signal of a monitoring target in a high-temperature working area; the model building module is used for building a transfer function between the data acquisition module and the monitoring target on the basis of a product simulation model and a fault physics method; according to the method, machine learning is introduced, the application range of a traditional transfer function oriented to a steady system is expanded, high precision is achieved on the nonlinear problem with the time sequence effect, and the time sequence machine learning model is combined on the basis of a product simulation model, a fault physics method and a data mining technology; the application range of a traditional linear steady system transfer function is broken through, and a nonlinear transfer function with a time-allowable effect between the sensor and a monitoring target is established.

Description

Technical Field

[0001] The present invention belongs to the field of condition monitoring, and particularly relates to a vibration monitoring system based on a transfer function. Background Art

[0002] Regarding the vibration monitoring of a special position of a certain product, especially a position where real-time testing cannot be directly carried out through a sensor. For example, a vibration monitoring position of a certain aircraft works in a high-temperature (above 1000 °C) environment for a long time. Traditional contact vibration sensors cannot work and can only work in a low-temperature area near the monitoring target. The position near the monitoring target is directly used to replace the actual target position, and there is a large error between the two. Non-contact sensors are usually too large to be carried, resulting in the inability to achieve vibration monitoring of this position. Summary of the Invention

[0003] The purpose of the present invention is to provide a vibration monitoring system based on a transfer function to solve the problem of vibration monitoring of positions where real-time testing cannot be directly carried out through a sensor as mentioned in the above background art.

[0004] To achieve the above purpose, the present invention provides the following technical solutions:

[0005] A vibration monitoring system based on a transfer function, comprising:

[0006] A data acquisition module for acquiring signals of a monitoring target located in a high-temperature working area;

[0007] A model construction module for establishing a transfer function between the data acquisition module and the monitoring target based on the method of physics of failure with a product simulation model as the basis;

[0008] A data processing module for converting the signals of the monitoring target into vibration information of the monitoring target through the transfer function.

[0009] Preferably, the data acquisition module detects the product through a sensor located in a low-temperature working area.

[0010] Preferably, the model construction module includes:

[0011] Selecting the initial number and layout of sensors;

[0012] Calculating the position data of the sensors through the product simulation model, and dividing the position data into a training data set and an evaluation data set;

[0013] Selecting the type of the transfer function and completing the modeling of the transfer function based on the training data set;

[0014] Combining the evaluation data set to evaluate the accuracy of the transfer function;

[0015] If the evaluation accuracy reaches the expected target, the data processing program is completed; otherwise, the number of sensors, their layout, and the type of transfer function are adjusted for a new round of iteration.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0017] Based on the simulation model of the product, the present invention combines the method of physics of failure and data mining technology with a time series machine learning model, breaks through the applicable range of the transfer function of traditional linear time-invariant systems, and establishes a non-linear transfer function with time-delay effect between the sensor and the monitoring target. This function takes sensor data as input, and converts the sensor signal into the vibration signal of the vibration monitoring target with low delay, and outputs the vibration information of the monitoring target. Thus, when the monitoring target vibrates, the vibration information can be obtained non-contact, realizing vibration monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings. Among them:

[0019] Figure 1 is the system module block diagram of the present invention;

[0020] Figure 2 is the operation flow block diagram of the model construction module of the present invention;

[0021] Figure 3 is the schematic diagram of the relative displacement of the acquisition points of the present invention;

[0022] Figure 4 is the schematic diagram of the vertical direction offset correction of the present invention;

[0023] Figure 5 is the force hammer impact test data graph of the present invention;

[0024] Figure 6 is the shaker excitation test data graph of the present invention;

[0025] Figure 7 is the force hammer impact and shaker excitation test data graph of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0026] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following will make a detailed description of the specific embodiments of the present invention in conjunction with the drawings of the specification.

[0027] In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention. However, the present invention may be practiced in other ways than those specifically described herein, and those skilled in the art can make similar generalizations without departing from the spirit of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0028] Secondly, as used herein, "an embodiment" or "embodiments" refer to specific features, structures, or characteristics that may be included in at least one implementation of the present invention. The appearances of "in an embodiment" in different places in this specification do not all refer to the same embodiment, nor are they separate or alternative embodiments that exclude each other.

[0029] As shown in the Figure 1 and Figure 2 accompanying

[0030] Embodiment 1: This embodiment provides a vibration monitoring system based on a transfer function, including:

[0031] A data acquisition module for acquiring signals of a monitoring target located in a high-temperature working area;

[0032] A model construction module for establishing a transfer function between the data acquisition module and the monitoring target based on a product simulation model and a physics-of-failure method;

[0033] A data processing module for converting the signals of the monitoring target into vibration information of the monitoring target through the transfer function.

[0034] Based on a product simulation model and a physics-of-failure method, a transfer function between a sensor and a monitoring target is established. This function takes sensor data as input and converts the sensor signals into vibration signals of the vibration monitoring target with low latency, outputting the vibration information of the monitoring target. Thus, when the monitoring target vibrates, vibration information can be obtained non-contactingly to achieve vibration monitoring.

[0035] Specifically, the data acquisition module detects the product through a sensor located in a low-temperature working area.

[0036] Specifically, the model construction module includes:

[0037] Selecting the initial number and layout of sensors;

[0038] Calculating the position data of the sensors through the product simulation model and dividing the position data into a training data set and an evaluation data set;

[0039] Selecting the type of transfer function and completing the modeling of the transfer function based on the training data set;

[0040] Evaluating the accuracy of the transfer function in combination with the evaluation data set;

[0041] If the evaluation accuracy reaches the expected goal, the data processing program is completed. Otherwise, adjust the number, layout of sensors, and type of transfer function for a new round of iteration.

[0042] As can be seen from the above, by establishing a training data set and an evaluation data set, using the training data set to model the transfer function, and then using the evaluation data set to evaluate the accuracy. When the evaluation accuracy does not reach the expected goal, adjust and optimize the number, layout of sensors, and type of transfer function and its modeling method to improve the accuracy and response rate of the data processing program;

[0043] The modeling methods include, but are not limited to, the transfer function modeling method based on the least squares method and the gradient descent method. The transfer functions include, but are not limited to, linear transfer functions and non-linear transfer functions.

[0044] As shown in Figure 3 and Figure 4 shown:

[0045] Embodiment 2: Use an exciter to perform mechanical vibration excitation on the test piece to measure the vibration response of the lens 1 on the test piece. The test piece is connected to the test platform through a special fixture, and a high-precision laser displacement sensor is used to measure the displacement change generated by the vibration of the lens;

[0046] A total of 5 acquisition points are set on the lens to arrange sensors. The 5 sensors respectively collect the axial displacements at 3 positions of the lens, which are 18 mm away from the center of the lens respectively, 1 lateral displacement, perpendicular to the central axis of the lens, and 1 vertical displacement, perpendicular to the central axis. The 3 acquisition points on the axis are arranged in an equilateral triangle. Acquisition point 1 on the axis and the vertical acquisition point are arranged in the vertical symmetry plane of the lens, and the lateral acquisition point is arranged in the horizontal symmetry plane of the lens;

[0047] At the initial moment, the pose of the lens is in the designed position, and the relative distance values between the 5 acquisition points and the sensors at the initial moment t 0 and a certain moment t i are collected respectively, and then the pose is calculated by the following method;

[0048] At the initial moment, the symmetric center of the lens is the origin O of the coordinate system. The X-axis direction is vertically upward, coinciding with the vertical displacement test point. The Y-axis direction is horizontally to the right, coinciding with the lateral displacement test point. The Z-axis direction follows the right-hand screw rule. The coordinates of the axial displacement test points are (r, 0, 0), where r is the distance from the test point to the center. Assume that the rotation of the lens around the Z-axis will not cause a change in the light refraction quality, and the rotation of the lens around the Z-axis is ignored in the acquisition scheme;

[0049] Collect data:

[0050] Initial time t 0 The distance values returned by the laser displacement sensors at 5 acquisition points are respectively d z1,0 , d z2,0 , d z3,0 , d x,0 , d y,0 , at time t i The distance values returned by the laser displacement sensors at 5 acquisition points are respectively d z1,i , d z2,i , d z3,i , d x,i , d y,i , and the relative displacement values at 5 acquisition points are respectively:

[0051] d z1 = d z1,i - d z1,0

[0052] d z2 = d z2,i - d z2,0

[0053] d z3 = d z3,i - d z3,0

[0054] d x = d x,i - d x,0

[0055] d y = d y,i - d y,0

[0056] Pose calculation:

[0057] According to the relative displacement value d z2 at axial acquisition point 2 and the relative displacement value d z3 at axial acquisition point 3, the relative displacement value at the vertical symmetry plane is calculated by interpolation method:

[0058] d zc = 0.5×(d z2 + d z3 )

[0059] According to the relative displacement value d z1 at axial acquisition point 1 and the relative displacement value d z2 at axial acquisition point 2, the relative displacement value d z4 at the horizontal symmetry plane is calculated:

[0060]

[0061] According to the relative displacement value d of the axial acquisition point 1 z1 and the relative displacement value d of the axial acquisition point 3 z3 , calculate the relative displacement value d at the horizontal symmetry plane by the interpolation method z5 :

[0062]

[0063] According to the relative displacement value d of the axial acquisition point 1 z1 and the relative displacement value d at the vertical symmetry plane zc , the angle change around the Y-axis can be obtained:

[0064]

[0065] According to the relative displacement value d at the horizontal symmetry plane z4 and the relative displacement value d z5 , the angle change around the X-axis can be obtained:

[0066]

[0067] Assume that the coordinates of the position of the lens acquisition point 1 at the initial moment are (x 1 , y 1 , z 1 ), then the position (x' 1 , y' 1 , z' 1 ) of the acquisition point 1 after the lens position changes can be approximated as:

[0068]

[0069] In the formula, Δd x is the offset correction amount of the uppermost edge point of the measurement surface in the X-axis direction caused by considering the lens thickness D and the deflection of the lens; Δd y is the offset correction amount of the rightmost edge point of the measurement surface in the Y-axis direction caused by considering the lens thickness D and the deflection of the lens, and its calculation formula is as follows:

[0070]

[0071] In the formula, δ xz is the offset amount of the uppermost edge point of the measurement surface in the Z-axis direction caused by considering the deflection of the lens, and δ yz is the offset amount of the rightmost edge point of the measurement surface in the Z-axis direction caused by considering the deflection of the lens;

[0072] δ xz value can be approximately obtained according to the geometric relationships such as the relative displacement value d of the axial acquisition point 1 z1 , the distance L between the acquisition point 1 and the uppermost edge point of the measurement surface 1 , the rotation angle θ around the Y-axis, etc.; δyz The value can be approximately obtained according to the relative displacement value d at the horizontal symmetry plane z4 , and the distance L to the rightmost edge point of the measurement surface 5 , geometric relationships such as the rotation angle φ about the X-axis, etc.

[0073]

[0074] The coordinate changes of the positions of acquisition point 2 and acquisition point 3 are similar to those of acquisition point 1:

[0075]

[0076] For the vertical displacement acquisition point, only the X-axis and Z-axis components change. Assuming that the coordinates of the position of the vertical displacement acquisition point of the lens at the initial moment are (x v , y v , z v ), then the position (x' v , y' v , z' v ) after the position change of the lens can be approximated as:

[0077]

[0078] For the lateral displacement acquisition point, only the Y-axis and Z-axis components change. Assuming that the coordinates of the position of the lateral displacement acquisition point of the lens at the initial moment are (x h , y h , z h ), then the position (x' h , y' h , z' h ) after the position change of the lens can be approximated as:

[0079]

[0080] In the CAT IA software, a cylinder is drawn to replace the lens. The radius of the lens is 1000 mm and the thickness is 100 mm. By translating in three axial directions and rotating by a certain angle about the X-axis and Y-axis respectively, and calculating using the above method, the difference between the finally calculated position of the acquisition point and the theoretical position is that under the condition of a maximum change of about 35 microns, the calculation error does not exceed 1 micron, meeting the measurement requirements.

[0081] Example 3: Three types of sensors are used to measure single-point excitation and single-point response under the condition of impact hammer excitation. The impact hammer is used to perform mechanical vibration excitation on the simulated sample, and the simulated sample is connected to the test platform through a special fixture;

[0082] A total of 5 acquisition points are set on the simulation sample to arrange sensors. The 5 sensors respectively collect the axial displacements at 3 positions of the lens, which are 18 mm away from the center of the simulation sample, 1 lateral displacement, perpendicular to the central axis of the simulation sample, and 1 vertical displacement, perpendicular to the central axis. The 3 acquisition points on the axis are arranged in an equilateral triangle. The acquisition point 1 on the axis and the vertical acquisition point are arranged in the vertical symmetry plane of the lens, and the lateral acquisition point is arranged in the horizontal symmetry plane of the simulation sample. The test steps are as follows:

[0083] Step 1: Connect the data acquisition system, calibrate the measurement system, and ensure that all the equipment used can work properly and calibrate the measurement system;

[0084] Step 2: Pre-acquire to determine reasonable parameters, including sampling frequency, acquisition instrument range setting, sampling duration, and check the pre-acquired data, including linear check, FRF and coherence check, trigger force value, reciprocity check;

[0085] Step 3: Install acceleration sensors, laser displacement sensors, and eddy current sensors at the same position and conduct numerical tests simultaneously;

[0086] Step 4: After determining the effective setting parameters, first use a force hammer to excite the simulation sample at the same position, record and save the data, and repeat the hammering at the fixed point 5 times to test the response value;

[0087] Step 5: Save the original test data;

[0088] Step 6: Conduct precision evaluation and analysis on the measurement results of the three types of sensors, and select the results of the measurement method with the highest precision for the next test.

[0089] The comparison of the measurement data of the three types of sensors is shown in the following table:

[0090]

[0091]

[0092] Result: The trends of the measurement data of the three types of sensors at the same position are basically the same. All three types of sensors have achieved effective measurement. The peak value of the eddy current sensor has the largest relative time delay, which is 0.6 ms later than the maximum value of the acceleration sensor and 0.2 ms later than the laser displacement sensor, and the waveform distortion is obvious with serrations. The measurement object is limited to metals; therefore, the eddy current sensor will not be used in subsequent tests.

[0093] Basic data test of transfer function:

[0094] Fix and install the simulation sample (lens) on the load-bearing platform through a fixture according to the coordinate system defined below. The axial direction is the Z-axis, and the positive direction is forward;

[0095] When placed horizontally (test state), the upward direction is the positive Y direction;

[0096] When placed horizontally (test state), when viewed from the back to the front, the right side is the positive X direction;

[0097] Arrange 5 laser displacement sensors, among which:

[0098] Arrange 3 in the Z direction, named U3-1 at the +Y position, U3-2 at the -X position, and U3-3 at the +X position;

[0099] Arrange 1 in the Y direction, named U2-1;

[0100] Arrange 1 in the X direction, named U1-1.

[0101] There are 3 acceleration sensors, named A1, A2, and A3 respectively. The test steps are as follows:

[0102] Step 1: Install the sensors;

[0103] Step 2: Connect the data acquisition system and calibrate the measurement system: Ensure that all the equipment used can work properly and calibrate the measurement system;

[0104] Step 3: Pre-acquire to determine reasonable parameters, including sampling frequency, range setting of the acquisition instrument, sampling duration, and check the pre-acquired data, including linearity check, FRF and coherence check, trigger force value, reciprocity check;

[0105] Step 4: After determining the effective setting parameters, first use a force hammer to excite the simulated sample, record and save the data, repeat hitting the fixed point with the force hammer 5 times, and test the response value;

[0106] Step 5: Adopt the multi-point excitation method, repeat hitting 5 times, and test the response value;

[0107] Step 6: Save the original test data;

[0108] Step 7: Randomly hit the force hammer 5 times, test the response value, and save the original data;

[0109] Step 8: Single-point excitation of the shaker + hitting with the force hammer, test the response value, and save the original data;

[0110] Step 9: End the measurement until the required state is completed, save the original data and establish a transfer function model.

[0111] The results are as attached Figure 5 、 6 and attached Figure 7As shown: Comparative tests for measuring vibration displacement signals and basic data tests for transfer functions are carried out using three types of sensors. The test process is controllable, the test data is reliable. When the sampling frequency is 20KHz, the signal spectrogram is smooth, there is no missing data point, and the acquisition is reliable.

[0112] Importantly, it should be noted that the construction and arrangement of the present application shown in multiple different exemplary embodiments are merely illustrative. Although only a few embodiments are described in detail in this disclosure, those who refer to this disclosure should easily understand that many modifications are possible without substantially departing from the novel teachings and advantages of the subject matter described in this application (for example, the dimensions, scales, structures, shapes and proportions of various components, and parameter values (such as temperature, pressure, etc.), installation arrangements, use of materials, color, changes in orientation, etc.). For example, an element shown as integrally formed may be composed of multiple parts or elements, the position of the element may be inverted or otherwise changed, and the nature or number or position of discrete elements may be altered or changed. Therefore, all such modifications are intended to be included within the scope of the present invention. The order or sequence of any process or method steps may be changed or reordered according to alternative embodiments. In the claims, any "means-plus-function" clause is intended to cover the structure that performs the recited function herein, and not only structural equivalents but also equivalent structures. Other substitutions, modifications, changes and omissions may be made in the design, operating conditions and arrangement of the exemplary embodiments without departing from the scope of the present invention. Therefore, the present invention is not limited to a particular embodiment, but extends to various modifications that still fall within the scope of the appended claims.

[0113] In addition, in order to provide a concise description of the exemplary embodiments, not all features of the actual embodiments may be described (i.e., those features that are not relevant to the currently considered best mode of implementing the present invention or those features that are not relevant to implementing the present invention).

[0114] It should be understood that in the development of any actual implementation, as in any engineering or design project, a large number of specific implementation decisions may be made. Such development efforts may be complex and time-consuming, but for those of ordinary skill in the art who benefit from this disclosure, without excessive experimentation, such development efforts will be a routine task of design, manufacturing and production.

[0115] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered within the scope of the claims of the present invention.

Claims

1. A vibration monitoring system based on transfer function, characterized in that: include: A data acquisition module, used to acquire signals of monitoring targets located in a high-temperature working area; A model building module, used to establish a transfer function between the data acquisition module and the monitoring target based on a product simulation model and a method based on fault physics; The data processing module is used to convert the signal of the monitoring target into vibration information of the monitoring target through the transfer function.

2. A vibration monitoring system based on transfer function according to claim 1, characterized in that: The data acquisition module detects the product through a sensor located in a low-temperature working area.

3. A vibration monitoring system based on transfer function according to claim 2, characterized in that: The model building module includes: Select the initial number and layout of sensors; Calculating position data of the sensor by using the product simulation model, and dividing the position data into a training data set and an evaluation data set; Selecting the transfer function type and completing the modeling of the transfer function based on the training data set; performing accuracy evaluation on the transfer function in combination with the evaluation data set; If the evaluation accuracy reaches the expected target, the data processing procedure is completed; otherwise, the number and layout of the sensors and the type of transfer function are adjusted to perform a new round of iteration.

Images