Weld health monitoring system and method based on digital twin and ultrasonic lamb wave

Abstract

The present invention provides a weld health monitoring system and method based on digital twins and ultrasonic lamb waves. The weld health monitoring system and method based on digital twins and ultrasonic lamb waves include: a rail weld, a state acquisition module, a circuit conversion module, a random code generator, a transmission unit, a control unit, a rail weld digital twin, and a visualization management and control platform; the rail weld digital twin is a dynamically operable three-dimensional digital model of the rail, and each component in the three-dimensional digital model is mapped one-to-one with the physical entity components of the rail; the visualization platform can display the health status of the rail weld in real time. The present invention adopts random code frequency hopping technology, which can realize the unification of the operating status and data of the rail weld in reality and the virtual equipment in the three-dimensional program, so as to achieve a more intuitive conclusion of the health status of the rail weld and a clearer discovery of the damage location of the rail.

Description

Technical Field

[0001] The present invention relates to the technical field of rail weld detection, and in particular to a weld health monitoring system and method based on digital twins and ultrasonic lamb waves. Background Art

[0002] With the rapid development of railway operating mileage and passenger and freight volume, the number of rails in service is increasing, and the number of rails working under heavy load is also increasing. During their service, rails need to withstand the extrusion and impact of train wheels for a long time. The working conditions are harsh and the stress on the rails is relatively concentrated. Rails in long-term service will inevitably suffer from various damages such as bending, deformation, wear, corrosion, surface defects, cracks, crushing, and peeling. As ground basic equipment, the working condition of rails directly affects the safety of train transportation. Trains passing through damaged rails will bring huge safety hazards and may even lead to major traffic accidents such as derailment and overturning. Major traffic accidents such as train derailment and overturning due to damage tell us that strengthening real-time monitoring of rail damage is of great significance to ensuring the safe operation of trains.

[0003] At present, the methods that are widely studied in rail damage detection at home and abroad mainly include magnetic flux leakage detection, eddy current detection, X-ray detection, electromagnetic detection, manual patrol detection, optical fiber detection, structural health monitoring technology, etc. Among them, magnetic flux leakage detection technology is limited to detecting damage defects on the surface and near the surface of the rail, and is not suitable for complex structures with cross-sectional shapes, nor is it suitable for detecting narrow cracks; eddy current detection is only applicable to metal surface detection, and it is difficult to detect internal damage and cannot locate the damage; X-ray detection is difficult to use in rail damage detection due to the large size of the equipment, large investment, and certain hazards; electromagnetic detection has low energy conversion efficiency, and has good detection effects on parts such as rail heads, but cannot detect defects at the bottom of the rail waist; manual patrol detection is greatly affected by factors such as weather, and cannot detect internal conditions; in optical fiber detection, the mechanical properties of optical fiber are poor, and large external forces and bending will cause optical fiber breakage, and the maintenance and installation costs are high. In the current rail damage detection, large rail flaw detection vehicles are generally used to carry out rapid periodic inspections of in-service lines. However, the detection capability of rail flaw detection vehicles depends largely on the alignment of ultrasonic wheel probes (hereinafter referred to as probe wheels) during rapid operation. In addition, because ultrasonic waves decay quickly in the air, ultrasonic signals cannot effectively enter the designated area inside the rails, and there is a possibility of missed damage detection. At the same time, due to the continuous increase in railway lines, the human resources required for rail detection have increased, so only periodic inspections can be carried out using flaw detection vehicles, and real-time detection cannot be achieved. There may be damaged rails that have not been detected for a period of time, increasing the risk of train operation.

[0004] In general, there is a certain window period when using traditional detection methods, and they are costly and inefficient. Therefore, new non-destructive testing methods with low cost and high efficiency need to be studied. At present, the structural health monitoring method based on Lamb waves is a hot research area. Ultrasonic lamb waves have the characteristics of long propagation distance, large detection area, high speed, low cost and no radiation. The structural health monitoring technology (SHM) uses on-site non-destructive sensing technology to collect data, and monitors rail damage or warns of early structural degradation by analyzing the signal.

[0005] However, the use of ultrasonic lamb waves for structural health detection also faces challenges. Ultrasonic lamb waves propagate in all directions in the material. As the propagation distance increases, dispersion will occur and the energy attenuation will be large, especially when encountering structures such as welds, the energy loss is more serious. At present, most scholars use the reflection characteristics, refraction characteristics and attenuation characteristics of ultrasonic lamb waves after vertical incidence on the weld to detect defects in the weld. The reflected wave signal obtained in this way is very weak, and it is difficult to accurately determine the location of the damage. At the same time, due to the use of five peaks and the lamb wave has multiple modes and is converted by mode, when the pulse signal band is narrower, the time required for the time signal is longer, resulting in the overlap of echo signals, multi-modal signals and direct signals. In addition, the five peaks are non-equal amplitude waves, the signal peak power is small, and the reflected wave peak amplitude is small, which makes it difficult to detect minor damage.

[0006] In addition, it is also of great significance to realize real-time detection and synchronous analysis of data. The existing traditional rail damage detection uses rail flaw detection vehicles for periodic detection, but this method obviously cannot analyze and process the data in real time, and no trains can run during the operation of the flaw detection vehicle. The present invention can realize real-time detection and synchronous analysis of data by constructing a digital twin model for synchronous transmission of data. Moreover, it is only necessary to analyze the digital twin model, without having to go to the site in person, which greatly reduces manpower consumption, saves time, and provides safety protection for train operation.

[0007] Digital twin technology. Digital twin, also known as digital mapping and digital mirroring, is officially explained as a digital twin that makes full use of data such as physical models, sensor updates, and operation history, integrates multi-disciplinary, multi-physical, multi-scale, and multi-probability simulation processes, and completes mapping in virtual space, thereby reflecting the entire life cycle of the corresponding physical equipment. Digital twin is a concept that transcends reality and can be regarded as a digital mapping system of one or more important, interdependent equipment systems. In simple terms, digital twin is to create a digital version of a "clone" based on a device or system. This "clone" is also called a "digital twin". It is created on an information platform and is virtual. Unlike design drawings and modeling software such as CAD, the digital twin model is a dynamic simulation of physical objects. Changes to the digital twin model are based on the physical design model of the entity, the data fed back by the sensors on the entity, and the historical data of the entity's operation. If system design changes are required or if you want to know how the system responds under special external conditions, engineers can conduct "experiments" on the twin. In this way, it can avoid the impact on the main body and improve efficiency and save costs.

[0008] In addition, digital twins also have the characteristics of a full life cycle, that is, they can run through the entire cycle of the product including design, development, service, maintenance and even scrapping and recycling; they also have real-time / quasi-real-time characteristics, and a comprehensive real-time and quasi-real-time connection can be established between the entity and the twin; there is also a two-way relationship, not only the entity outputs data to the twin, but the twin can also feedback data to the entity.

[0009] The present invention is to establish a digital twin model of the rail structure, collect data through ultrasonic lamb waves, process data through sampling and random code frequency modulation, locate damage through elliptical positioning technology, and synchronize information and interact with the established digital twin model. By real-time analysis of the collected continuous data, when the rail damage reaches the threshold level, the digital twin model can be used to timely discover and take countermeasures. Summary of the invention

[0010] The technical problem solved by the present invention is to provide a weld health monitoring system and method based on digital twins and ultrasonic lamb waves, which realize data transmission and conversion through a state acquisition module, a circuit conversion module, a transmission unit, and a control unit. By means of ultrasonic lamb waves, continuous data is obtained, and the health of the rail can be continuously monitored, thus solving the problem of window period in existing detection methods.

[0011] In order to solve the above technical problems, the weld health monitoring system based on digital twin and ultrasonic lamb wave provided by the present invention comprises: a rail weld to be tested, a random code generator, a state acquisition module, a circuit conversion module, a transmission unit, a control unit, a data processing module, and a digital twin system;

[0012] The state acquisition module uses one or more piezoelectric sensor arrays, which are symmetrically arranged on both sides of a preset reference line, to generate an excitation signal and a response signal, and to acquire lambda wave response signals on the symmetrical propagation paths on both sides of the reference line;

[0013] The random code generator is used to define the frequency band of the transmission signal, and the state acquisition module collects data under the control of the random code generator;

[0014] The circuit conversion module is used to convert the charge signal generated by the state acquisition module into an electrical signal;

[0015] The transmission unit is used to transmit the electrical signal generated by the circuit conversion module;

[0016] The control unit is used to convert the electrical signal into a visual curve;

[0017] The data processing module is used to process the collected continuous ultrasonic lamb wave signal;

[0018] The digital twin system includes a rail weld digital twin, mechanism model software and a visualization management and control platform. The rail weld digital twin is a dynamically operable three-dimensional digital model of the rail weld. The visualization management and control platform provides a virtual platform for the rail weld three-dimensional digital twin to display the real-time status of the rail weld three-dimensional digital twin.

[0019] Preferably, the piezoelectric sensor includes a shell and a sensitive element, the sensitive element is used to convert the collected waveform data into a charge signal, the sensitive element is embedded in a sensitive area on a surface of one side of the shell, and the piezoelectric sensor emits lamb wave signals to all directions.

[0020] Preferably, the control unit is connected to one or more transmission units, each transmission unit is connected to one or more circuit conversion modules, and the one or more transmission units and circuit conversion modules are used to form a sensor network.

[0021] Preferably, any group of piezoelectric sensor arrays includes a piezoelectric sensor for generating an excitation signal and a receiving sensor for receiving the excitation signal.

[0022] Preferably, the rail weld digital twin is a dynamically operable three-dimensional digital model of the rail weld, the model is a digital mapping of the rail weld, and the three-dimensional digital model corresponds to the health status of the rail intuitively and in real time.

[0023] A weld health monitoring method based on digital twin and ultrasonic lamb wave comprises the following steps:

[0024] Step 1: For the rail weld, establish a corresponding three-dimensional mechanism model according to its composition structure;

[0025] Step 2: including at least one group of piezoelectric sensor arrays, symmetrically arranged on both sides of a preset reference line and installed on the surface of the rail, for collecting lamb wave response signals, the lamb wave signals collected by the piezoelectric sensors are transmitted to the data circuit conversion module, and transmitted to the control unit through the transmission unit;

[0026] Step 3: After completing step 2, the control unit sends the data to the data processing module, and after wavelet reconstruction, denoising and correlation operation, the arrival time of the reflected wave is selected; if there is no reflected wave, no processing is performed; if there is a reflected wave, the propagation distance of the reflected wave and the piezoelectric sensor number are transmitted to the digital twin system;

[0027] Step 4: The digital twin system locates the damage position through elliptical positioning, fits the data with the digital twin model and performs visualization.

[0028] Preferably, the 'db4' wavelet is used, and the received signal is reconstructed and denoised using the minimum maximum variance threshold with a soft threshold and 7 decomposition levels.

[0029] Preferably, the autocorrelation characteristics of the random code are utilized to perform a cross-correlation operation on the collected continuous lamb wave signal and the transmitted pulse signal to generate a pulse signal narrower than the received signal.

[0030] Preferably, the ultrasonic lamb wave velocities at different temperatures are measured in advance in the laboratory, and the reflected wave propagation distance is calculated by detecting the time when the reflected wave appears and selecting the lamb wave propagation velocity according to the temperature gradient.

[0031] In the step 1, the model is established using software such as MATLAB, 3DMax, SolidWorks or Ansys to establish its geometric model, and add contact forms and corresponding material properties to the rail weld. At the same time, finite element analysis is used to predict the various physical properties and overall performance of the model, and it is constantly adjusted and optimized in the virtual environment.

[0032] The visual management and control platform renders the constructed model in real time; if there is warning information, the unacceptable differences between the digital twin model and the physical world entity are displayed through one or more dimensions, and the abnormal area is marked with a specific color.

[0033] The abnormal warning module determines the damage of a certain rail based on the signals collected on both sides of the rail reference line. If there is damage, the information that there is damage in the rail monitoring section is sent to the visualization platform.

[0034] The present invention uses ultrasonic lamb waves to collect data, uses electromagnetic energy or solar energy to power it, and realizes data transmission and conversion through a state acquisition module, a circuit conversion module, a transmission unit, and a control unit. By using ultrasonic lamb waves to obtain continuous data, the health of the rail can be continuously monitored, which solves the problem of a window period in existing detection methods.

[0035] The present invention uses digital twin technology to construct a digital twin 3D model, uses wavelet and frequency hopping technology to analyze and organize data, and realizes data transmission through wireless or wired methods to achieve data interaction and synchronous analysis, which well solves the problem of poor data real-time performance of existing detection methods; in addition, remote monitoring can be achieved through the constructed digital twin model, and when problems arise, the best solutions can be discovered in time and proposed, greatly improving efficiency; a series of experiments can also be carried out on the constructed digital twin model, which not only avoids the impact on the entity, but also improves efficiency and saves costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 A diagram of a weld health monitoring system based on digital twins and ultrasonic lamb waves provided by the present invention;

[0037] Figure 2 for Figure 1 The schematic diagram of ellipse positioning in the weld health monitoring system based on digital twin and ultrasonic lamb wave is shown;

[0038] Figure 3 A schematic diagram of a group velocity dispersion curve of a weld health monitoring system based on digital twins and ultrasonic lamb waves provided by the present invention;

[0039] Figure 4 Schematic diagram of the phase velocity dispersion curve of the weld health monitoring system based on digital twin and ultrasonic lamb wave provided by the present invention. DETAILED DESCRIPTION

[0040] The present invention will be further described below in conjunction with the accompanying drawings and implementation modes.

[0041] like Figure 1-4As shown, at least one group of piezoelectric sensor arrays is selected, symmetrically arranged on both sides of a preset reference line, and adsorbed on the surface of the rail by magnets. Each group of piezoelectric sensors includes a piezoelectric sensor for generating an excitation signal and a receiving sensor for receiving the excitation signal. The sensor is connected to the circuit conversion module via a cable. The state acquisition module, the random code generator, and the control unit can be integrated via a circuit board.

[0042] It should be noted that the present invention does not limit the selection of sensors. A circular thin-sheet piezoelectric ceramic sensor (PZT) is used here, which is an artificial ferroelectric with electric domains inside after polarization treatment and has piezoelectric properties. The positive piezoelectric effect refers to the deformation of the piezoelectric ceramic itself when it is subjected to an external load in the polarization direction. This deformation will change the position of the electric domain and cause the internal electric polarization phenomenon. At this time, charges with opposite signs will be generated on both sides of the polarization direction of the piezoelectric ceramic, that is, a potential difference will be formed on the two sides; when the external load disappears, the deformation also disappears immediately, and the potential difference between the two sides of the piezoelectric ceramic also disappears, returning to the original zero potential state: when the external load changes, that is, the deformation changes, the electric polarization phenomenon also changes accordingly, and the potential difference on the two sides of the piezoelectric ceramic also changes accordingly. This change is reflected in the change in the output signal voltage amplitude, and the amplitude change is proportional to the change in the external load. This positive piezoelectric effect can make piezoelectric ceramics as signal receivers. The inverse piezoelectric effect is that when different voltages are applied to the positive and negative electrodes of a piezoelectric ceramic, that is, the piezoelectric ceramic is subjected to an electric field, the electric domain position changes, causing the piezoelectric ceramic to deform macroscopically; when the applied voltage value is different, that is, when the electric field is different, the degree of deformation of the piezoelectric ceramic is different, and the deformation size is proportional to the voltage; when the voltage disappears, the electric field effect disappears immediately, and the piezoelectric ceramic returns to its original deformation state. This inverse piezoelectric effect can make piezoelectric ceramics act as an exciter of stress wave signals, thus reflecting that piezoelectric ceramics can convert electrical energy into mechanical energy.

[0043] After arranging the excitation device and the receiving device, the excitation frequency of the lamb wave should be selected. In this example, the piezoelectric ceramic sensor is arranged at the waist of the rail. Since the rail flange has a plate-like structure, the stress wave excited on the flange will be continuously reflected at the edge, so its dispersion curve is studied. The dispersion equation of the lamb wave is divided into symmetric mode (S mode) and antisymmetric mode (A mode):

[0044]

[0045]

[0046]

[0047]

[0048]

[0049] Among them is C L Longitudinal wave velocity, C T is the shear wave velocity, C P is the phase velocity, h is the rail thickness, k is the wave number, and ω is the angular frequency. Using the above formula, p and q can be expressed as follows:

[0050]

[0051]

[0052] Using the above equation, the phase velocity dispersion curves of each mode of lamb wave can be solved, and the group velocity can be obtained from the phase velocity:

[0053]

[0054] Normally, the shear wave velocity C T Less than the longitudinal wave velocity C L , using the graphical method and the dichotomy method, the solution interval is divided into three parts, namely (0,C T ],(C T ,C L ] and (C L ,+∞), when C P The dispersion equation is simplified in different intervals, and solved with the help of MATLAB programming and drawing functions to obtain the group velocity and phase velocity dispersion curves.

[0055] Regardless of whether it is a symmetric mode or an antisymmetric mode, any frequency-thickness product corresponds to at least two modes. As the frequency-thickness product increases, the types of modes increase. If a larger frequency-thickness product is selected, the complexity of lamb wave damage detection will increase. Therefore, in this example, the frequency-thickness product is selected to be 1 MHz·mm, including the A0 and S0 modes.

[0056] The signals used to excite lamb waves are mainly divided into narrowband and broadband excitation signals. Due to the dispersion characteristics and multi-mode characteristics of lamb waves, broadband signals make signal analysis very difficult, causing a series of problems such as signal aliasing. Therefore, in this example, narrowband signals are selected as excitation signals, which are more likely to form wave packet aliasing. The five-peak signal is as follows:

[0057]

[0058] Where A is the signal amplitude, f C is the signal center frequency, n is the number of peaks, and H is the Heaviside step function. According to the frequency-thickness product, a five-peak signal excitation with a center frequency of 250kHz and an amplitude of 10 is selected in this example.

[0059] After the control unit completes part of the data processing, it is transmitted to the data processing module wirelessly or wiredly. The module first denoises the signal through wavelet reconstruction, then performs cross-correlation operation on the transmitted pulse signal and the received signal to generate a pulse signal narrower than the received signal, and detects whether there is a reflected wave. If there is no reflected wave, no processing is performed. If there is a reflected wave, the arrival time of the reflected wave is detected, and the propagation distance of the reflected wave is calculated, and the sensor number and the propagation distance are transmitted to the digital twin system.

[0060] The digital twin system uses input parameters and an elliptical positioning method to locate the damage, and displays the damage location on the monitor through a visualization platform.

[0061] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A weld health monitoring system based on digital twin and ultrasonic lamb wave, characterized in that: include: Rail weld to be tested, random code generator, state acquisition module, circuit conversion module, transmission unit, control unit, data processing module, digital twin system; The state acquisition module uses one or more piezoelectric sensor arrays, which are symmetrically arranged on both sides of a preset reference line, to generate an excitation signal and a response signal, and to acquire lambda wave response signals on the symmetrical propagation paths on both sides of the reference line; The random code generator is used to define the frequency band of the transmission signal, and the state acquisition module collects data under the control of the random code generator; The circuit conversion module is used to convert the charge signal generated by the state acquisition module into an electrical signal; The transmission unit is used to transmit the electrical signal generated by the circuit conversion module; The control unit is used to convert the electrical signal into a visual curve; The data processing module is used to process the collected continuous ultrasonic lamb wave signal; The digital twin system includes a rail weld digital twin, mechanism model software and a visualization management and control platform. The rail weld digital twin is a dynamically operable three-dimensional digital model of the rail weld. The visualization management and control platform provides a virtual platform for the rail weld three-dimensional digital twin to display the real-time status of the rail weld three-dimensional digital twin.

2. The weld health monitoring system based on digital twin and ultrasonic lamb wave according to claim 1 is characterized in that: The piezoelectric sensor comprises a shell and a sensitive element. The sensitive element is used to convert the collected waveform data into a charge signal. The sensitive element is embedded in a sensitive area on one side surface of the shell, and the piezoelectric sensor emits lamb wave signals to all around.

3. The weld health monitoring system based on digital twin and ultrasonic lamb wave according to claim 1 is characterized in that: The control unit is connected to one or more transmission units, each transmission unit is connected to one or more circuit conversion modules, and the one or more transmission units and circuit conversion modules are used to form a sensor network.

4. The weld health monitoring system based on digital twin and ultrasonic lamb wave according to claim 1 is characterized in that: Any set of piezoelectric sensor arrays includes a piezoelectric sensor for generating an excitation signal and a receiving sensor for receiving the excitation signal.

5. The weld health monitoring system based on digital twin and ultrasonic lamb wave according to claim 1 is characterized in that: The rail weld digital twin is a dynamically operable three-dimensional digital model of the rail weld, which is a digital mapping of the rail weld, and the three-dimensional digital model corresponds to the health status of the rail intuitively and in real time.

6. The weld health monitoring method based on digital twin and ultrasonic lamb wave is applied to the rail weld health monitoring system based on digital twin and ultrasonic lamb wave as described in any one of claims 1 to 5, characterized in that: The steps include: Step 1: For the rail weld, establish a corresponding three-dimensional mechanism model according to its composition structure; Step 2: including at least one group of piezoelectric sensor arrays, symmetrically arranged on both sides of a preset reference line and installed on the surface of the rail, for collecting lamb wave response signals, the lamb wave signals collected by the piezoelectric sensors are transmitted to the data circuit conversion module, and transmitted to the control unit through the transmission unit; Step 3: After completing step 2, the control unit sends the data to the data processing module, and after wavelet reconstruction, denoising and correlation operation, the arrival time of the reflected wave is selected; if there is no reflected wave, no processing is performed; if there is a reflected wave, the propagation distance of the reflected wave and the piezoelectric sensor number are transmitted to the digital twin system; Step 4: The digital twin system locates the damage position through elliptical positioning, fits the data with the digital twin model and performs visualization.

7. The method for weld health monitoring based on digital twin and ultrasonic lamb wave according to claim 6 is characterized in that: The 'db4' wavelet is used, and the minimum maximum variance threshold is used to reconstruct and denoise the received signal under the parameters of soft threshold and decomposition level 7.

8. The method for weld health monitoring based on digital twin and ultrasonic lamb wave according to claim 6 is characterized in that: By utilizing the autocorrelation characteristics of random codes, the collected continuous lamb wave signal is cross-correlated with the transmitted pulse signal to generate a pulse signal narrower than the received signal.

9. The method for weld health monitoring based on digital twin and ultrasonic lamb wave according to claim 6, characterized in that: The ultrasonic lamb wave velocities at different temperatures are measured in advance in the laboratory. The reflected wave propagation distance is calculated by detecting the time when the reflected wave appears and selecting the lamb wave propagation rate according to the temperature gradient.

Images