A pipeline sensor arrangement method based on modal feature extraction and inversion

Abstract

This invention discloses a pipeline sensor placement method based on modal feature extraction and inversion, comprising: establishing a three-dimensional model of the pipeline and importing it into a numerical simulation system to obtain the pipeline model; performing modal analysis to obtain the vibration modal responses, mode shapes, and preset positions of the sensors for each order of the pipeline model, and obtaining the feature matrix of the full-order vibration response function through the preset positions of the sensors; placing sensors according to the preset positions, measuring the vibration response vector, and inverting the vibration feature values ​​based on the feature matrix; calculating the vibration value at any point in the pipeline using the vibration feature values ​​and the mode shapes of the vibration modal responses for each order of the pipeline model; comparing the calculated vibration value with the measured value to obtain the inversion error; if the inversion error exceeds 10%, increasing the modal analysis order and repeating the above process; otherwise, outputting the sensor placement positions. This invention uses fewer sensors to obtain comprehensive vibration characteristics of the pipeline, reducing costs.

Description

Technical Field

[0001] This invention relates to the field of pipeline sensor placement technology, and more specifically, to a pipeline sensor placement method based on modal feature extraction and inversion. Background Technology

[0002] Based on existing hardware, software, and monitoring methods, various mechanical equipment and their associated pipeline structures can be monitored in real time. The monitored data allows for intelligent diagnosis of equipment operation, and in the event of mechanical failure, emergency measures can be taken to prevent serious safety accidents.

[0003] As a key supporting structure for maintaining mechanical equipment, the vibration of pipelines directly affects the working condition and service life of related machinery. Real-time monitoring of pipeline operation requires the coordinated operation of various signal sensors, which must be strategically positioned to obtain effective monitoring data. A central data processing system and specific data calculation methods are then used to uniformly process the data recorded by each sensor, yielding the vibration characteristics of the entire pipeline.

[0004] Current methods for selecting the placement of pipeline vibration sensors are often based on the pipeline structure itself, typically choosing locations such as: 1) bends with high flexibility; 2) locations near the vibration source; 3) suspended positions between two fixed points; 4) interfaces between the pipeline and important equipment; 5) sections of the pipeline with large concentrated masses such as valves and tees; and 6) areas with large responses in pipeline modal analysis. However, these principles for selecting pipeline vibration sensor locations rely too heavily on engineers' experience and lack a unified standardized method. This often results in redundant and ineffective sensor placements, or the avoidance of critical locations and the neglect of important information. Summary of the Invention

[0005] To address the problems existing in the prior art, this invention provides a pipeline sensor placement method based on modal feature extraction and inversion. This method can obtain comprehensive vibration characteristics of the pipeline using fewer sensors, reduce the cost of pipeline monitoring equipment, provide effective monitoring data, and, in conjunction with relevant calculation methods, effectively avoid abnormal vibrations in the pipeline.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a pipeline sensor placement method based on modal feature extraction and inversion, specifically including the following steps:

[0007] Step 1: Create a 3D model of the pipeline based on the engineering drawings, import it into the ANSYS Workbench numerical simulation system, perform mesh generation, set the pipeline operating parameters according to the application scenario of the pipeline, set the initial order of modal analysis, and obtain the pipeline model.

[0008] Step 2: Perform modal analysis on the obtained pipeline model to obtain the vibration modal response, mode shape and preset position of the sensor for each order of the pipeline model, and obtain the characteristic matrix of the full-order vibration response function through the preset position of the sensor;

[0009] Step 3: Arrange the sensors according to their preset positions, measure the vibration response vector, and invert the vibration characteristic values ​​using the characteristic matrix;

[0010] Step 4: Calculate the vibration value of any point in the pipeline using the vibration characteristic values ​​and the mode shapes of the vibration modal responses of each order of the pipeline model; compare the vibration calculation value with the measured value to obtain the inversion error. If the inversion error exceeds 10%, increase the modal analysis order and repeat steps 2-4; otherwise, output the sensor placement position.

[0011] Furthermore, step 2 includes the following sub-steps:

[0012] Step 21: Perform modal analysis on the pipeline model to obtain the vibration displacement response and mode shape of the pipeline model in each mode. Use the maximum response position of each mode of the pipeline model as the predetermined sensor placement position and number it.

[0013] Step 22: Obtain the vibration response function based on the displacement characteristics B1 and frequency characteristics B2 of the pipeline vibration at the sensor placement location, and combine the displacement characteristics B1 and frequency characteristics B2 to form the feature matrix B of the full-order vibration response function.

[0014] Furthermore, the feature matrix B = B1·B2.

[0015] Furthermore, the process of inverting the vibration eigenvalues ​​in step 3 is as follows:

[0016] B×C=D

[0017] Where B is the characteristic matrix, C is the vibration characteristic value, and D is the measured vibration response vector.

[0018] Furthermore, the initial order of the modal analysis is set to 6.

[0019] Compared with existing technologies, this invention has the following advantages: The pipeline sensor placement method of this invention is based on modal simulation data, fully extracting the simulation results of the pipeline. Based on the vibration mode characteristics of the pipeline, it proposes a method for establishing the characteristic matrix of the full-order vibration response function and a method for solving the inverse vibration eigenvalues. Combined with measured data, it can achieve a combination of simulation and measured data. This method can use fewer sensors to obtain comprehensive vibration characteristics of the pipeline, reducing the cost of pipeline monitoring equipment, providing effective monitoring data, and, in conjunction with relevant calculation methods, effectively preventing abnormal vibrations in pipelines. Attached Figure Description

[0020] Figure 1 The flowchart is a process for the pipeline sensor placement method based on modal feature extraction and inversion according to the present invention.

[0021] Figure 2 A schematic diagram of the lubricating oil supply pipeline for the synchronous condenser rotor;

[0022] Figure 3 A schematic diagram showing the pre-defined sensor placement positions for the lubricating oil supply pipeline model of the camera rotor, based on modal analysis. Detailed Implementation

[0023] The technical solution of the present invention will be further explained and described below with reference to the accompanying drawings.

[0024] like Figure 1 This is a flowchart of the pipeline sensor placement method based on modal feature extraction and inversion according to the present invention. The pipeline sensor placement method specifically includes the following steps:

[0025] Step 1: Establish a 3D model of the pipeline based on the engineering drawings, import it into the ANSYS Workbench numerical simulation system, perform mesh generation, and set the pipeline operation parameters according to the application scenario of the pipeline. Based on engineering experience, set the initial order of modal analysis to 6 to obtain the pipeline model. In this invention, if the initial order is set too small, it will increase the data calculation error; if the initial order is set too large, it will increase the operating cost.

[0026] Step 2: Perform modal analysis on the obtained pipe model to obtain the mode shapes of each vibration mode response of the pipe model and the preset positions of the sensors, and obtain the characteristic matrix of the full-order vibration response function through the preset positions of the sensors; specifically including the following sub-steps:

[0027] Step 21: Perform modal analysis on the pipeline model to obtain the vibration displacement response and mode shape of the pipeline model in each mode. Use the maximum response position of each mode of the pipeline model as the predetermined sensor placement position and number it. Using the maximum response position as the sensor placement position can obtain the vibration state of the pipeline with fewer sensors.

[0028] Step 22: Obtain the vibration response function based on the displacement feature B1 and frequency feature B2 of the pipe vibration at the sensor placement location, and combine the displacement feature B1 and frequency feature B2 to form the feature matrix B = B1·B2 of the full-order vibration response function; In this invention, the vibration response function is in the form of a sine function Asin(ωt), and each element in B conforms to this form, ignoring the time term.

[0029] Step 3: Arrange the sensor according to the preset position of the sensor, measure the vibration response vector C, and invert the vibration characteristic value D by combining the characteristic matrix. The specific inversion process is: B×C=D.

[0030] Step 4: Calculate the vibration value at any point in the pipeline using the vibration characteristic values ​​and the mode shapes of the vibration modal responses of each order of the pipeline model; compare the vibration calculation value with the measured value to obtain the inversion error. If the inversion error exceeds 10%, it indicates that the initial order is too small, resulting in a large data calculation error. It is necessary to increase the modal analysis order and repeat steps 2-4; otherwise, output the sensor placement position.

[0031] The present invention provides a pipeline sensor placement method based on modal feature extraction and inversion, which can quickly and accurately obtain comprehensive vibration characteristics on pipelines by using fewer sensors and modal analysis data, thereby reducing sensor placement costs.

[0032] Example

[0033] Taking the lubricating oil supply pipeline of the rotor bearing of a synchronous condenser as an example, such as Figure 2 As shown in the attached figure, the positions marked 1-11 represent locations where vibrations are likely to be significant. Using existing methods, at least 11 sets of sensors would be required. However, by using the pipe sensor deployment method based on modal feature extraction and inversion of this invention, the number of sensors is reduced to 6, and the cost is significantly reduced. The specific analysis process is as follows:

[0034] (1) Establish a three-dimensional model of the pipeline based on the engineering drawings, import it into the ANSYS Workbench numerical simulation system, perform mesh generation, set the pipeline operation parameters according to the application scenario of the pipeline, set the initial order of modal analysis to 6, and obtain the pipeline model.

[0035] (2) Modal analysis is performed on the obtained pipeline model to obtain the vibration modal responses, mode shapes, and preset positions of the sensors for each order of the pipeline model. The characteristic matrix of the full-order vibration response function is obtained through the preset positions of the sensors. Specifically,

[0036] Step 21: Perform modal analysis on the pipeline model to obtain the vibration displacement response and mode shapes of the pipeline model in each mode, as shown in Table 1. Use the maximum response location of each mode of the pipeline model as the predetermined sensor placement location, as shown in Table 21. Figure 3 and numbered (k) j (j = 1, 2, ..., 6), the frequencies of each mode and the amplitudes of each monitoring point are shown in Table 2;

[0037] Table 1 Modal vibration displacement response of the pipeline model

[0038]

[0039]

[0040] Table 2. Frequency of each mode and amplitude at each monitoring point

[0041]

[0042] Step 22: Obtain the vibration response function based on the displacement feature B1 and frequency feature B2 of the pipeline vibration at the sensor placement location, and combine the displacement feature B1 and frequency feature B2 to form the feature matrix B of the full-order vibration response function, where the displacement feature B1 is represented as follows:

[0043]

[0044] Frequency characteristic B2 is represented as follows:

[0045]

[0046] Step 3: Arrange the sensor according to the preset position, measure the vibration response vector D, and invert the vibration characteristic value C by combining the characteristic matrix;

[0047] B×C=D

[0048] C = (c1, c2, c3, c4, c5, c6) T

[0049] D = (d1, d2, d3, d4, d5, d6) T

[0050] Among them, c i , i = 1 to 6, represent the 6 eigenvalues ​​of the vibration vectors that need to be solved; d i , i = 1 to 6, represents the vibration response vector measured by the sensor.

[0051] Step 4: Combine the vibration mode response modes of the pipeline model in Table 1 to obtain the calculated global pipeline vibration value E after inversion. The vibration information of any point in the pipeline can be obtained from E.

[0052] (M.1,M.2,M.3,M.4,M.5,M.6)×C=E

[0053] The inversion error was obtained by comparing the calculated and measured values ​​of vibration. The calculated inversion error was 8.2%, which is less than 10%, indicating that the initial modal order is reasonable.

[0054] Using the modal feature extraction and inversion methods described above, the placement location of the pipeline sensor in the pipeline was obtained. After testing, its vibration response performance was found to be... Figure 2The vibration response measured by the 11 sets of sensors in the invention is similar, indicating that the number of sensors deployed based on the pipeline sensor arrangement method of the present invention is reduced, the cost is reduced, but the measurement effect will not change. At the same time, it provides a foundation for subsequent pipeline vibration digital visualization and digital twin technology upgrades.

[0055] The above are merely preferred embodiments of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should be considered within the scope of protection of the present invention.

Claims

1. A method for arranging pipeline sensors based on modal feature extraction and inversion, characterized in that, Specifically, the steps include the following: Step 1: Create a 3D model of the pipeline based on the engineering drawings, import it into the ANSYS Workbench numerical simulation system, perform mesh generation, set the pipeline operating parameters according to the application scenario of the pipeline, set the initial order of modal analysis, and obtain the pipeline model. Step 2: Perform modal analysis on the obtained pipe model to obtain the vibration modal responses, mode shapes, and preset sensor positions for each order of the pipe model. Then, obtain the characteristic matrix of the full-order vibration response function using the preset sensor positions. This includes the following sub-steps: Step 21: Perform modal analysis on the pipeline model to obtain the vibration displacement response and mode shape of the pipeline model in each mode. Use the maximum response position of each mode of the pipeline model as the predetermined sensor placement position and number it. Step 22: Based on the displacement characteristics of pipeline vibration at the sensor placement location. B 1 and frequency characteristics B 2. Obtain the vibration response function and the displacement characteristics. B 1 and frequency characteristics B 2. Characteristic matrices constituting the full-order vibration response function B ; Step 3: Arrange the sensors according to their preset positions, measure the vibration response vector, and invert the vibration characteristic values ​​using the characteristic matrix; Step 4: Calculate the vibration value of any point in the pipeline using the vibration characteristic values ​​and the mode shapes of the vibration modal responses of each order of the pipeline model; compare the vibration calculation value with the measured value to obtain the inversion error. If the inversion error exceeds 10%, increase the modal analysis order and repeat steps 2-4; otherwise, output the sensor placement position.

2. The pipeline sensor placement method based on modal feature extraction and inversion according to claim 1, characterized in that, The feature matrix .

3. The pipeline sensor placement method based on modal feature extraction and inversion according to claim 1, characterized in that, The process of inverting vibration eigenvalues ​​in step 3 is as follows: in, B The characteristic matrix, C For vibration characteristic values, D This is the measured vibration response vector.

4. The pipeline sensor placement method based on modal feature extraction and inversion according to claim 1, characterized in that, The initial order of the modal analysis was set to 6.

Images