Method and system for evaluating performance of wide and narrow joints of ballastless track based on multi-data fusion

Abstract

The application provides a kind of based on multi-sensor data fusion's performance evaluation method and system of wide and narrow joint of ballastless track, it is related to track joint performance evaluation technical field, comprising: based on design standard through finite element simulation to establish the theoretical benchmark model of healthy joint, and extract benchmark feature vector;When there is no train operation, apply preset excitation and synchronously collect input force and output response signal;By frequency domain analysis, obtain measured frequency response function, and extract resonance frequency, bandwidth, damping ratio and dynamic stiffness and other multi-dimensional characteristic parameters to form state feature vector;Then compare the current state vector with the theoretical benchmark to obtain the performance deviation coefficient, and calculate the performance degradation coefficient combined with historical data;Finally, based on the two coefficients, an evaluation model is constructed to calculate the health index, and by matching with the preset threshold, the precise grading evaluation of joint performance is realized. This method realizes the precise quantitative evaluation and predictive maintenance of joint state.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of track joint performance evaluation, in particular to a wide-narrow joint performance evaluation method and system for ballastless track based on multi-sensor data fusion. BACKGROUND

[0002] As a weak link of track structure, the performance state of track joint is directly related to the safety, stability and wheel-rail noise level of train operation. Traditional track joint performance evaluation relies on manual inspection, static geometric dimension measurement or simple vibration amplitude judgment. These methods are highly subjective and inefficient, and it is difficult to find potential damage inside the joint, so it is impossible to achieve precise and quantitative evaluation of the structure health state. With the rapid development of high-speed railway and urban rail transit, intelligent and refined operation and maintenance of track infrastructure are required, and it is particularly important to develop a method that can objectively and comprehensively evaluate the performance state of track joint.

[0003] In the prior art, a vibration control performance evaluation method for the whole life cycle of a subway vibration reduction track (CN116245278A) is disclosed. By long-term monitoring and comparative analysis of the vibration response of ordinary tracks and vibration reduction tracks, Z vibration level and insertion loss are used to evaluate the overall vibration reduction performance and service state of the vibration reduction track. This method focuses on evaluating the vibration transmission characteristics of the overall track system and the efficiency of the vibration reduction element, but the evaluation object is the macroscopic "track-tunnel" system, and the evaluation index reflects the overall energy level of vibration. However, this method does not involve fine performance diagnosis of specific local components (such as joints) in the track structure, and lacks direct perception and evaluation of changes in the structural dynamics of the joint itself. Therefore, it cannot effectively identify and locate damage and performance degradation of the joint itself, and cannot meet the needs of targeted and preventive maintenance of key local structures of the track.

[0004] The above information disclosed in the background section is only intended to enhance the understanding of the background of the present disclosure, and therefore it can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0005] The present application aims to provide a wide-narrow joint performance evaluation method and system for ballastless track based on multi-sensor data fusion to solve the problems raised in the background.

[0006] To achieve the above-mentioned purpose, the present application provides the following technical solutions:

[0007] A wide-narrow joint performance evaluation method and system for ballastless track based on multi-sensor data fusion, the specific steps comprising:

[0008] S1: based on the joint design standard, a theoretical benchmark model of the joint to be evaluated in a healthy state is established by a finite element method, a benchmark feature vector is obtained according to the theoretical benchmark model, and the benchmark feature vector includes a theoretical resonance frequency, a theoretical half-power bandwidth, a theoretical damping ratio and a theoretical dynamic stiffness;

[0009] S2: when no train is running, a preset excitation signal is applied to the joint to be evaluated, and during the entire excitation action and the excitation decay process, a time-domain input force signal and a time-domain output response signal caused by the excitation signal are synchronously collected by using a force sensor and an acceleration sensor;

[0010] S3: the time-domain input force signal and the time-domain output response signal collected are analyzed in the frequency domain, and a measured frequency response function of the joint is solved; multi-dimensional feature parameters including a resonance frequency, a half-power bandwidth, a damping ratio and a dynamic stiffness are extracted from the measured frequency response function, and together constitute a state feature vector for representing the current state of the joint to be evaluated;

[0011] S4: the state feature vector is compared with the benchmark feature vector stored in the theoretical benchmark model to determine a comprehensive performance deviation coefficient; a historical state feature vector corresponding to the latest historical detection record of the joint to be evaluated is called, and a comprehensive performance degradation coefficient is determined based on the historical state feature vector;

[0012] S5: a joint performance evaluation model is constructed based on the comprehensive performance deviation coefficient and the comprehensive performance degradation coefficient, the comprehensive health index of the joint to be evaluated is determined by using the model, and the grading evaluation result is output by matching with a preset grading threshold.

[0013] Further, based on the joint design standard, a theoretical benchmark model is established by a finite element method, and the specific logic is as follows:

[0014] According to the design drawing of the joint to be evaluated, the geometric parameters are obtained, including the size of the ballastless track slab, the shape of the ballastless track slab, the concrete model and the ballastless track structure; based on the geometric parameters, a three-dimensional parametric geometric model of the joint area is established;

[0015] Material properties are given to each component in the geometric model, including the elastic modulus, density and Poisson's ratio of the ballastless track slab; then the finite element mesh is divided, hexahedral dominant elements are used for division, and local seeds are set around the weld area and bolt holes, and the element size of the area is controlled to be 1 / 10 of the global element size. to to realize mesh refinement;

[0016] The established finite element model is subjected to modal analysis, and the first several natural frequencies and mode shapes of the joint to be evaluated in a healthy state are calculated; ​Natural frequency and mode shape; based on the modal analysis results, the nodes in the model corresponding to the actual positions of the field excitation device are defined as input points, a unit harmonic force excitation is applied to the input points, and the nodes in the model corresponding to the actual installation positions of the field acceleration sensors are defined as output points, the acceleration responses of the output points are read, and the theoretical frequency response function between the input points and the output points is calculated;

[0017] From the theoretical frequency response function curve, the theoretical resonance frequency, the theoretical half-power bandwidth and the theoretical damping ratio corresponding to the first order mode are extracted, and the theoretical dynamic stiffness is obtained by calculating the reciprocal of the original dynamic flexibility of the theoretical frequency response function, which together constitute the reference feature vector.

[0018] Further, the specific execution process of S2 is as follows:

[0019] During the window period without train operation, a pre-set transient excitation signal is applied to the ballastless track slab by using an exciter as an excitation device; the time-domain input force signal generated by the force sensor installed on the excitation device and the time-domain output response signal generated by the acceleration sensor during the entire excitation action and the excitation attenuation process are synchronously collected; the starting point of the excitation signal is taken as the synchronous trigger reference in the collection process;

[0020] Among them, for the longitudinal joint between the track slabs, the acceleration sensor is fixed to the side surface of the rail support of the track slab on both sides of the joint, 0.3-0.5m away from the edge of the joint; for the transverse joint, the acceleration sensor is installed on the side surface of the track slab on both sides of the joint; the acceleration sensor is connected with the basic structure through a stainless steel support.

[0021] Further, the collected time-domain input force signal and time-domain output response signal are respectively subjected to Fourier transform to obtain corresponding frequency-domain signals and , and the self-power spectrum of the force signal and the cross-power spectrum of the force signal and the response signal are calculated according to and :

[0022]

[0023] In the formula, is the complex conjugate of ;

[0024] The measured frequency response function is calculated based on the following formula :

[0025] In the formula, (w) is the measured frequency response function, is the cross-power spectrum of the force signal and the response signal, The power spectrum of the force signal;

[0026] From the measured frequency response function Amplitude spectrum | The peak value is identified in the equation, and the frequency corresponding to the peak value is the resonant frequency of that order, denoted as: ,in, Index of order;

[0027] For the Calculate the two frequencies corresponding to the 3dB drop in amplitude at the first resonance peak. and The half-power bandwidth is calculated using the following formula:

[0028]

[0029] In the formula, The bandwidth is the half-power bandwidth of the i-th order;

[0030] According to the The first resonant frequency and its half-power bandwidth are calculated using the following formula: Step damping ratio :

[0031]

[0032] Take amplitude spectrum The average value in the 0-50Hz low-frequency range is used to calculate the dynamic stiffness, and its reciprocal is calculated.

[0033]

[0034] In the formula, For dynamic stiffness, Amplitude spectrum The average value in the low-frequency range of 0-50Hz;

[0035] The extracted front The resonance frequency, half-power bandwidth, damping ratio, and dynamic stiffness are combined to form a state feature vector that characterizes the current state of the joint to be evaluated.

[0036] Furthermore, the state feature vector is compared with the benchmark feature vector stored in the theoretical benchmark model to determine the overall performance deviation coefficient, based on the following formula:

[0037]

[0038] In the formula, The comprehensive performance deviation coefficient is used to characterize the degree of deviation between the current overall performance status of the joint being evaluated and the health benchmark. an index of a characteristic parameter participating in comparison, a number of characteristic parameters participating in comparison, and , an order number; a characteristic parameter value of an order participating in comparison, a reference value of an order characteristic parameter, a reference characteristic vector obtained from S1; the characteristic parameters include resonance frequency, half-power bandwidth, damping ratio, and dynamic stiffness;

[0039] retrieve the historical state characteristic vector and the corresponding comprehensive performance deviation coefficient corresponding to the latest times of historical detection records of the joint to be evaluated, and arrange them in chronological order; calculate the change rate of the comprehensive performance deviation coefficient of two adjacent historical detection records, and calculate the arithmetic mean value to obtain the comprehensive performance degradation coefficient;

[0040] The formula for calculating the comprehensive performance degradation coefficient is as follows:

[0041]

[0042] In the formula, is a comprehensive performance degradation index, used to represent the average rate of deterioration of the joint performance state over time; is the total number of historical detection records called, is the index of the historical detection record, is the comprehensive performance deviation coefficient calculated according to the order historical detection record, is the comprehensive performance deviation coefficient calculated according to the order historical detection record; represents the time interval between the order historical detection record and the order historical detection record.

[0043] Further, a joint performance evaluation model is constructed with the comprehensive performance deviation coefficient and the comprehensive performance degradation coefficient as input variables, and the comprehensive health index of the joint to be evaluated is determined by using the joint performance evaluation model; the expression of the joint performance evaluation model is as follows:

[0044]

[0045] In the formula, is the comprehensive health index of the joint to be evaluated, used to represent the overall health condition of the joint to be evaluated; is a preset upper limit of the health index; is a natural constant; and The preset weight is , .

[0046] Further, based on the matching result of the comprehensive health index and the preset joint performance grading threshold, the grading evaluation result of the joint to be evaluated is output, and the specific logic is as follows:

[0047] The first threshold value is preset And the second threshold value , and satisfy: ;

[0048] If , it is determined that the performance level of the joint to be evaluated is "healthy";

[0049] If , it is determined that the performance level of the joint to be evaluated is "attention";

[0050] If , it is determined that the performance level of the joint to be evaluated is "abnormal".

[0051] The application also provides a ballastless track wide and narrow joint performance evaluation system based on multi-sensor data fusion, which is used to execute the above-mentioned ballastless track wide and narrow joint performance evaluation method based on multi-sensor data fusion, comprising:

[0052] The theoretical benchmark modeling module is used to establish a theoretical benchmark model of the joint to be evaluated in a healthy state based on the joint design standard through a finite element method, and a benchmark feature vector is obtained according to the theoretical benchmark model, wherein the benchmark feature vector includes a theoretical resonance frequency, a theoretical half-power bandwidth, a theoretical damping ratio and a theoretical dynamic stiffness.

[0053] The signal acquisition module is used to apply a preset excitation signal to the joint to be evaluated when there is no train operation, and to synchronously collect time-domain input force signals and time-domain output response signals caused by the excitation signal during the entire excitation period and the excitation decay process using force sensors and acceleration sensors.

[0054] The feature vector construction module is used to perform frequency domain analysis on the collected time-domain input force signals and time-domain output response signals, solve the measured frequency response function of the joint, and extract multi-dimensional feature parameters including resonance frequency, half-power bandwidth, damping ratio and dynamic stiffness from the measured frequency response function to form a state feature vector for representing the current state of the joint to be evaluated.

[0055] The coefficient calculation module is used to compare the state feature vector with the benchmark feature vector stored in the theoretical benchmark model to determine a comprehensive performance deviation coefficient, and to retrieve the latest The secondary historical detection record corresponds to a historical state feature vector, and a comprehensive performance degradation coefficient is determined based on the historical state feature vector;

[0056] The health grading evaluation module is used for constructing a joint performance evaluation model based on the comprehensive performance deviation coefficient and the comprehensive performance degradation coefficient, determining the comprehensive health degree index of the joint to be evaluated by using the model, and outputting a grading evaluation result by matching with a preset grading threshold.

[0057] Compared with the prior art, the beneficial effects of the present application are:

[0058] The present application realizes multi-dimensional accurate perception and quantitative evaluation of joint performance from local damage to overall state by combining the theoretical benchmark model with the measured dynamic response analysis, effectively overcomes the limitations of strong subjectivity and inability to detect internal damage of the traditional method; by introducing the comprehensive health degree index and the grading evaluation mechanism, not only can the deviation degree of the current performance of the joint from the health benchmark be objectively reflected, but also the performance degradation trend can be sharply captured, thereby providing a scientific basis and decision support for preventive maintenance and accurate control of the track joint, and significantly improving the intelligent level and safety and reliability of the track infrastructure operation and maintenance. BRIEF DESCRIPTION OF DRAWINGS

[0059] Figure 1 It is a whole method flowchart of the present application;

[0060] Figure 2 It is a 3D scatter plot image of the comprehensive performance deviation coefficient and the comprehensive health degree index;

[0061] Figure 3 It is a parallel coordinate image of the comprehensive performance degradation coefficient and the comprehensive health degree index;

[0062] Figure 4 It is a whole system module schematic diagram of the present application. DETAILED DESCRIPTION

[0063] In order to make the purpose, technical scheme and advantages of the present application more clear and obvious, the present application is further described in detail below in combination with specific embodiments.

[0064] It should be noted that, unless otherwise defined, technical terms or scientific terms used in the present application shall have the common meaning as understood by one of ordinary skill in the art to which this application pertains. The terms "first", "second", and similar terms used in the present application do not denote any order, quantity, or importance, but are only used to distinguish different components. The terms "include", "contain", and similar terms mean that the elements or objects before the terms encompass the elements or objects listed after the terms and their equivalents, and do not exclude other elements or objects. The terms "connect" or "connected" and similar terms are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. The terms "upper", "lower", "left", "right", and the like only represent relative positional relationships, which can change when the absolute positions of the described objects change.

[0065] Embodiment:

[0066] Referring to Figures 1-3 The present application provides a technical solution:

[0067] The performance evaluation method for wide and narrow joints of ballastless track based on multi-sensor data fusion includes the following specific steps:

[0068] S1: Based on the joint design standard, a theoretical benchmark model of the joint to be evaluated in a healthy state is established by a finite element method, and a benchmark feature vector is obtained according to the theoretical benchmark model, the benchmark feature vector including a theoretical resonance frequency, a theoretical half-power bandwidth, a theoretical damping ratio, and a theoretical dynamic stiffness;

[0069] In this embodiment, the theoretical benchmark model is established based on the joint design standard by a finite element method, and the specific logic followed is as follows:

[0070] According to the design drawings of the joint to be evaluated, its geometric parameters are obtained, including the size of the ballastless track slab, the shape of the ballastless track slab, the type of concrete, and the structure of the ballastless track; based on the geometric parameters, a three-dimensional parametric geometric model of the joint area is established;

[0071] Material properties are assigned to each component in the geometric model, including the elastic modulus, density, and Poisson's ratio of the ballastless track slab; then the geometric model is subjected to finite element meshing, and hexahedral dominant elements are used for meshing, and local seeds are set around the weld area and bolt holes, with the element size in this area controlled to be to to achieve mesh refinement;

[0072] Modal analysis is performed on the established finite element model to calculate the first six order natural frequencies and the corresponding mode shapes in the healthy state; natural frequencies and vibration modes; based on the modal analysis results, define the nodes in the model corresponding to the actual positions of the field excitation devices as input points, apply a unit harmonic force excitation to the input points, and define the nodes in the model corresponding to the actual installation positions of the field acceleration sensors as output points, read the acceleration responses of the output points, and calculate the theoretical frequency response functions between the input points and the output points;

[0073] From the theoretical frequency response function curve, extract the theoretical resonance frequencies, theoretical half-power bandwidths, and theoretical damping ratios of the first M modes, and obtain the theoretical dynamic stiffness by calculating the reciprocal of the original point dynamic flexibility of the theoretical frequency response function, to jointly constitute the reference feature vector.

[0074] The theoretical reference model established in step S1 is the cornerstone and core innovation point for realizing precise evaluation. Unlike the prior art which mainly relies on long-term monitoring of vibration responses to evaluate the performance of a macroscopic system, the present application defines an absolute and objective health state quantitative standard, i.e., a reference feature vector, for each joint to be evaluated by constructing a "digital twin" in a virtual space based on the precise design parameters and material properties of the joint. This model simulates the dynamic characteristics (such as the first M natural frequencies, damping ratios, and dynamic stiffness) of the joint in an ideal healthy state, which enables subsequent comparison of measured data to be no longer a simple historical data trend analysis or a vague comparison with general standards, but a comparison with an accurate theoretical ideal value derived from its own design genes.

[0075] The significance of this step is that it upgrades the evaluation method from a relative state evaluation relying on historical data accumulation to an absolute health diagnosis based on physical principles. The advantage is that it greatly improves the accuracy and early warning capability of the evaluation. Even a newly laid joint can be effectively identified if its measured dynamic characteristics deviate from its own theoretical reference model, thereby discovering early defects that have not yet caused significant vibration abnormalities but have internal potential damage, achieving true preventive maintenance and overcoming the limitations of the prior art that cannot perform fine performance diagnosis on specific local components.

[0076] S2: When there is no train operation, apply a predetermined excitation signal to the joint to be evaluated, and use force sensors and acceleration sensors to simultaneously collect time-domain input force signals and time-domain output response signals induced by the excitation signal during the entire excitation period and the excitation decay process;

[0077] In this embodiment, the specific execution process of S2 is as follows:

[0078] ​During the non-train operation window, a shaker is used as an excitation device to apply a preset transient excitation signal to the ballastless track slab; time-domain input force signals generated by force sensors installed on the excitation device and time-domain output response signals generated by acceleration sensors during the entire excitation action and in the excitation decay process are synchronously collected; the collection process takes the starting point of the excitation signal as a synchronous trigger reference;

[0079] Among them, for the longitudinal joint between the track slabs, the acceleration sensors are fixed to the side surfaces of the rail seats of the track slabs on both sides of the joint, 0.3-0.5 m away from the edges of the joint; for the transverse joint, the acceleration sensors are installed on the side surfaces of the track slabs on both sides of the joint; the acceleration sensors are connected with the basic structure through stainless steel supports.

[0080] Step S2 is a key data acquisition link for realizing precise and active diagnosis of the scheme, and the core significance thereof lies in that pure input and output signals reflecting only the structural dynamic characteristics of the joint are obtained through the self-controllable excitation-response test. Unlike the prior art which passively relies on random excitation of trains to monitor the vibration of the tunnel wall, the preset transient excitation signal is used to actively excite the vibration response of the joint during the "window period", and the input force and output acceleration signals are synchronously collected. This method completely gets rid of the dependence on the operating train, avoids the signal pollution caused by random factors such as train load, running speed and marshalling, and makes the collected data have high repeatability and comparability.

[0081] The advantages are: first, the signal quality is high, the active excitation can effectively cover the main working frequency band of the joint, and the extracted feature parameters can accurately reflect the structural state of the joint; second, the measured signal is directly derived from the excitation and response at the joint, avoiding the attenuation and interference of the vibration wave on the propagation path in the track bed and tunnel structure, thereby greatly improving the sensing sensitivity and signal-to-noise ratio of the state change of the joint itself, and laying a reliable data foundation for subsequent precise state recognition and early fault diagnosis.

[0082] S3: performing frequency domain analysis on the collected time-domain input force signals and time-domain output response signals, solving the measured frequency response function of the joint; extracting multi-dimensional feature parameters including resonance frequency, half-power bandwidth, damping ratio and dynamic stiffness from the measured frequency response function, which together constitute a state feature vector for characterizing the current state of the joint to be evaluated;

[0083] In this embodiment, the collected time-domain input force signals and time-domain output response signals are respectively subjected to Fourier transform to obtain corresponding frequency domain signals and , and the self-power spectrum of the force signal and the cross-power spectrum of the force signal and the response signal are calculated according to and . :

[0084]

[0085] In the formula, yes The complex conjugate;

[0086] The measured frequency response function is calculated based on the following formula. :

[0087] In the formula, (w) is the measured frequency response function. The cross-power spectrum of the force signal and the response signal. The power spectrum of the force signal;

[0088] From the measured frequency response function Amplitude spectrum | The peak value is identified in the equation, and the frequency corresponding to the peak value is the resonant frequency of that order, denoted as: ,in, Index of order;

[0089] Among them, the order refers to the inherent vibration mode levels exhibited by the track joint system in different vibration frequency ranges, and their corresponding relationship is as follows: from the measured frequency response function amplitude spectrum | From the low-frequency band to the high-frequency band, the peaks are divided in the order of their appearance. The first peak corresponds to the first-order natural vibration mode of the system, and its corresponding frequency is the first-order resonant frequency. The second peak corresponds to the second-order natural vibration mode, and its corresponding frequency is the second-order resonant frequency. And so on, each subsequent peak corresponds to a higher-order natural vibration mode in order of frequency from low to high.

[0090] For the Calculate the two frequencies corresponding to the 3dB drop in amplitude at the first resonance peak. and The half-power bandwidth is calculated using the following formula:

[0091]

[0092] In the formula, The bandwidth is the half-power bandwidth of the i-th order;

[0093] According to the The first resonant frequency and its half-power bandwidth are calculated using the following formula: Step damping ratio :

[0094]

[0095] Take amplitude spectrum The average value in the low frequency band of 0-50Hz is calculated, and the inverse value is obtained as the dynamic stiffness, i.e.:

[0096]

[0097] In the formula, is the dynamic stiffness, is the amplitude spectrum The average value in the low frequency band of 0-50Hz is calculated;

[0098] The extracted first-order resonance frequency, the half-power bandwidth of each order, the damping ratio of each order and the dynamic stiffness are combined to form a state feature vector for representing the current state of the joint to be evaluated.

[0099] Step S3 is the core technical link for converting the original data to state information, and its significance lies in that through accurate frequency domain analysis and feature extraction, the collected time domain vibration signal is converted into multi-dimensional feature parameters which can directly and sensitively represent the dynamics state of the joint structure. Unlike the prior art which mainly focuses on macroscopic and overall energy indicators such as Z vibration level and insertion loss, the present application extracts a series of characteristic parameters such as resonance frequency, half-power bandwidth, damping ratio and dynamic stiffness through in-depth analysis of the measured frequency response function. These parameters reveal the structural integrity, energy dissipation mechanism and bearing stiffness of the joint from different dimensions, providing a rich and accurate information source for subsequent performance evaluation.

[0100] The advantages are: first, the diagnosis dimension is more fine and deep, not only can judge whether the joint is abnormal, but also can preliminarily judge where the abnormality is and what kind of abnormality is through the change mode of each parameter, realizing the preliminary positioning and identification of the fault; second, the early warning ability is stronger, these structural dynamics parameters are extremely sensitive to small damage and performance degradation, and often change before the macroscopic vibration response, so that potential faults can be found earlier, overcoming the limitation of the prior art that only relies on vibration energy indicators and leads to lagging in early warning.

[0101] S4: comparing the state feature vector with the reference feature vector stored in the theoretical reference model to determine the comprehensive performance deviation coefficient; retrieving the historical state feature vector corresponding to the latest historical detection record of the joint to be evaluated, and determining the comprehensive performance degradation coefficient based on the historical state feature vector; S4: comparing the state feature vector with the reference feature vector stored in the theoretical reference model to determine the comprehensive performance deviation coefficient; retrieving the historical state feature vector corresponding to the latest historical detection record of the joint to be evaluated, and determining the comprehensive performance degradation coefficient based on the historical state feature vector;

[0102] In this embodiment, the state feature vector is compared with the reference feature vector stored in the theoretical reference model to determine the comprehensive performance deviation coefficient, and the formula is as follows:

[0103]

[0104] In the formula,​ is a comprehensive performance deviation coefficient, used to represent the deviation degree of the current overall performance state of the wide-narrow joint to be evaluated from the health benchmark; is an index of the characteristic parameters participating in comparison, is the number of characteristic parameters participating in comparison, and , is an order; is the value of the characteristic parameter participating in comparison, is the value of the characteristic parameter, and the benchmark value of the characteristic parameter is obtained from the benchmark characteristic vector S1; the characteristic parameters include resonance frequency, half-power bandwidth, damping ratio, and dynamic stiffness;

[0105] is a dependent variable comprehensive performance deviation coefficient Specifically reflects the overall deviation degree of the current measured state of the wide-narrow joint to be evaluated from the health theoretical benchmark, and the greater the value, the greater the deviation degree. The essence is the weighted root mean square value of the relative deviation of each characteristic parameter. By quantifying the deviation level of all key dynamic parameters, the multi-dimensional feature difference is fused into a single scalar. The technical effect is to convert the complex structural state difference into an intuitive numerical index, to provide a unified and quantifiable decision basis for subsequent performance grade determination, and to significantly improve the objectivity and efficiency of state evaluation.

[0106] is associated with the dependent variable The relevance comes from its direct participation in the calculation of the deviation degree: the relative deviation of each parameter is used as an input component of The deviation of any parameter relative to the benchmark value will increase the deviation square term, and in turn directly push up the value of The deviation behavior of all independent variables jointly determines the final size of the dependent variable.

[0107] In the formula, the dependent variable is strictly positively correlated with the relative deviation of each independent variable. The positive deviation or negative deviation of the measured value of any characteristic parameter relative to the benchmark value will be converted into a positive contribution due to its square term, resulting in the increase of This design ensures that only pays attention to the magnitude of deviation but not the direction, so as to fully capture any form of performance degradation.

[0108] This formula possesses significant formal rationality: it employs a structure of taking the square root of the sum of squares of relative deviations. First, using relative deviations eliminates the influence of differences in dimensions and orders of magnitude among different characteristic parameters, allowing each parameter to participate equally in the comprehensive calculation without dimensionality. Second, squaring the deviations amplifies the contribution of significant deviations to attract sufficient attention while preventing positive and negative deviations from canceling each other out, ensuring that deviations in any direction are effectively captured. Finally, by taking the square root after summing, the overall deviation level is restored to a level close to the original deviation, resulting in the final formula... It is a scalar value that intuitively reflects the degree of relative deviation of the whole. This structure is mathematically rigorous and easy to understand and apply in engineering.

[0109] Retrieve the most recent seam to be evaluated The historical state feature vectors and corresponding comprehensive performance deviation coefficients corresponding to each historical detection record are arranged in chronological order; the rate of change of the comprehensive performance deviation coefficients of each pair of adjacent historical detection records is calculated, and their arithmetic mean is calculated to obtain the comprehensive performance degradation coefficient;

[0110] The formula used to calculate the overall performance degradation factor is as follows:

[0111]

[0112] In the formula, It is a comprehensive performance degradation index, used to characterize the average rate at which the performance status of a joint deteriorates over time. The total number of historical detection records invoked. An index for historical detection records. According to the first The comprehensive performance deviation coefficient is calculated from the historical test records. According to the first The comprehensive performance deviation coefficient calculated from the historical test records; Indicates the first The first historical detection record and the first The time interval between each historical detection record.

[0113] Dependent variable comprehensive performance degradation coefficient Specifically, it reflects the average rate of deterioration of the joint's performance over time; a higher value indicates a faster rate of deterioration. Its meaning is the overall performance deviation coefficient per unit time. The average change. The technical effect lies in quantifying the dynamic trend of performance status changes, realizing a leap from static point-based assessment to dynamic process prediction, providing a key time dimension indicator for predictive maintenance, and enabling early warning of accelerated degradation risks when absolute deviations are not yet significant.

[0114] The independent variables in the formula include the first... the second detection and the first The three have an essential correlation with the dependent variable and jointly determine the value of the dependent variable: The difference directly reflects the change in the performance deviation from the healthy benchmark within the adjacent detection period, and is the core basis for calculating the performance degradation rate. The larger the difference, the more significant the change in the performance deviation within the period. Then it is used to convert the change amplitude into a change rate per unit time, eliminating the interference of different detection intervals (such as the same change amplitude, a short time interval represents a faster degradation rate), ensuring the comparability of the degradation degree in different periods in the time dimension. If any independent variable is missing, either the performance change amplitude cannot be obtained, or the time dimension rate cannot be quantified, ultimately leading to the inability of the dependent variable to accurately represent the average rate of joint performance degradation.

[0115] The dependent variable is positively correlated with the independent variable : When is fixed, the larger the performance deviation change rate in the adjacent period, the higher the calculated , indicating that the average rate of joint performance degradation is faster; while the dependent variable is negatively correlated with the independent variable : When is fixed, the longer the performance deviation change rate in the adjacent period, the lower the calculated , indicating that the average rate of joint performance degradation is slower. This relationship accurately quantifies the impact of performance change amplitude and time interval on the degradation rate.

[0116] This formula has clear formal rationality: its essence is to calculate the arithmetic mean of the first-order difference derivative of the comprehensive performance deviation coefficient at consecutive time points. The formula accurately captures the instantaneous change rate of performance deviation by calculating the change rate between the adjacent two detections; then it calculates the arithmetic mean of the instantaneous change rate of all consecutive time intervals, thereby smoothing out short-term fluctuations and refining the stable average change trend of performance deviation in a longer observation period. This processing enables to effectively represent the persistence and directionality of performance state deterioration, avoiding misjudgment caused by single mutation, and ensuring the robustness and forward-looking nature of the evaluation results.

[0117] Step 4 is the core data processing link of the present scheme to realize the leap from state monitoring to performance evaluation, and its significance lies in that by constructing the comprehensive performance deviation coefficient and the degradation coefficient, the comparison results of multi-dimensional characteristic parameters and the historical change trend are fused into two quantitative indexes with clear physical meaning, thereby realizing the comprehensive and dynamic evaluation of the joint performance. Unlike the prior art which mainly relies on the absolute size of a single measurement value or simple statistics for judgment, the present scheme not only quantifies the overall deviation degree of the current state from the theoretical health benchmark through the weighted root mean square algorithm, reflecting the "absolute degradation amount" of the performance, but also captures the "relative degradation speed" of the performance by calculating the average change rate of the historical data, thereby revealing the performance degradation trend.

[0118] The advantages of the present scheme are as follows: first, the evaluation dimension is more comprehensive, and both the current state and the change trend of the performance are considered, thereby overcoming the one-sidedness of the prior art which only focuses on the instantaneous state or only counts the number of over-standard times while ignoring the deterioration rate; second, the early warning capability is highly forward-looking, and the comprehensive performance degradation coefficient can sensitively capture the inflection point of accelerated performance deterioration, thereby issuing an early warning according to the deterioration trend even if the current state has not yet exceeded the standard, realizing truly predictive maintenance, and being significantly superior to the lagging evaluation mode of the prior art which can only alarm after the over-standard occurs.

[0119] S5: constructing a joint performance evaluation model based on the comprehensive performance deviation coefficient and the comprehensive performance degradation coefficient, determining the comprehensive health degree index of the joint to be evaluated by using the model, and outputting the grading evaluation result by matching with the preset grading threshold value;

[0120] In the present embodiment, the comprehensive performance deviation coefficient and the comprehensive performance degradation coefficient are taken as input variables to construct a joint performance evaluation model, and the comprehensive health degree index of the joint to be evaluated is determined by using the joint performance evaluation model; the expression of the joint performance evaluation model is as follows:

[0121]

[0122] In the formula, is the comprehensive health degree index of the joint to be evaluated, and is used to represent the overall health condition of the joint to be evaluated; is the preset upper limit of the health degree index; is a natural constant; and is a preset weight, and , .

[0123] wherein, The determination method of is as follows: when the joint is in an ideal healthy state, both the comprehensive performance deviation coefficient and the comprehensive performance degradation coefficient are zero; by substituting this ideal state value into the joint performance evaluation model, can be solved; and accordingly, ​The preset upper limit of the health index, usually set to 100, represents the perfect state of the joint performance, serving as the upper limit of the evaluation quantization.

[0124] Set The core reason lies in the realistic demand and safety priority of track joint maintenance: from the perspective of engineering practice, the comprehensive performance deviation coefficient directly reflects the deviation of the current state of the joint from the health benchmark, which is the core basis for judging whether the current joint meets the safety service requirements and whether there is an immediate risk; if It means that the current performance of the joint has deviated significantly from the healthy state, which may directly affect the stability of the track structure and the safety of driving, and needs to be prioritized and decided; while the comprehensive performance deterioration coefficient Although it can represent the performance deterioration trend, it more reflects the long-term change rule, and the risk reflected has a time buffer, which can be further used to develop subsequent monitoring or maintenance plans based on the current state. At the same time, the core goal of track maintenance is to ensure "current safety" first, and then to ensure "long-term stability" through trend analysis, so It is necessary to give higher weight to the performance deviation coefficient that is directly related to the current safety state, so as to avoid ignoring the existing performance hidden danger due to excessive emphasis on trends, ensure that the evaluation results are consistent with the realistic logic of "prioritizing immediate risks and planning long-term maintenance" in field maintenance, and improve the guidance value of the model to actual maintenance decisions.

[0125] Dependent variable comprehensive health index Specifically reflects the overall health level of the comprehensive current state and deterioration trend of the joint to be evaluated. Its meaning is a normalized score, and the higher the value, the better the health condition. The technical effect lies in integrating the two-dimensional indicators of "performance deviation degree" and "deterioration rate" into a intuitive and continuous health score through a mathematical model, realizing the quantitative integration and intuitive expression of complex performance state, and providing a core basis for automated grading decisions.

[0126] Independent variables and The relevance of the dependent variable comes from the fact that they jointly determine the health risk of the joint: Characterizes the "severity of the disease", and Characterizes the "speed of disease development". Both of them form the total "penalty" of health through linear combination The increase of any independent variable will increase the total penalty, resulting in a decrease in the value of the exponential function, i.e. the health decreases.

[0127] In the formula, the dependent variable is in strict negative correlation with the independent variables and . and An increase in will result in an increase in the exponential part of the exponential term, which in turn will cause the value of the natural exponential function to decrease, ultimately leading to a decrease in the health index This negative correlation is consistent with engineering logic: the greater the performance deviation or the faster the performance degradation, the lower the health should be. The model amplifies the negative impact of high-risk states on health through the exponential function.

[0128] The rationality of the form of this formula lies in its complete fit with the core needs of the overall health assessment of track joint, as well as its mathematical rigor and engineering practicality: from the function structure, the use of the exponential form with a natural constant as the base ensures that the overall health index decreases reasonably when the input variables increase, consistent with the objective law that "the greater the performance deviation or the faster the performance degradation, the lower the health". The introduction of processing can filter out performance non-degradation, avoiding interference from non-degradation trends in health assessment, consistent with the practical needs in engineering that "only negative degradation trends need to be considered". The setting of weights not only clearly prioritizes "current performance deviation over long-term degradation trend" through weight allocation, but also ensures that the impact of input variables on health is within a reasonable quantitative range. At the same time, setting as the upper limit ensures that the value of is always anchored to the "health benchmark", making it easy to match with pre-set classification thresholds. The overall form, from variable processing, function selection to weight design, is closely centered around the goal of "accurately quantifying the overall health status of the joint", achieving a high degree of adaptation between mathematical models and engineering reality.

[0129] Table 1: Overall Health Index Statistics

[0130]

[0131] In this data analysis, through the analysis of 15 groups of sample data, it can be clearly observed that the overall health index is in a significant non-linear negative correlation with the performance deviation coefficient and the performance degradation coefficient When the current state of the joint deviates slightly from the health benchmark and the performance is stable or showing an improvement trend, the health index maintains a high level, indicating that the joint is in good condition. Conversely, when and either significantly increase, especially when both are at high levels, the health index will decrease sharply, indicating that not only is the current state of the joint poor, but the performance is also deteriorating rapidly, posing a higher risk.

[0132] The effectiveness of the evaluation model is reflected in its ability to intelligently balance static and dynamic indicators: performance deviation coefficient The dominant determines the basic level of health, and the performance degradation coefficient plays an important role in adjustment and early warning. Even if the current absolute performance deviation is acceptable, if the degradation trend is obvious, the health will be significantly lowered, thus achieving early risk warning. This fusion mechanism makes the evaluation results reflect not only the immediate health status of the joint, but also predict the future performance change trend, providing a reliable quantitative basis for precision and predictive maintenance.

[0133] Based on the matching results of the comprehensive health index and the preset joint performance classification threshold, the classification evaluation results of the joint to be evaluated are output, and the specific logic is as follows:

[0134] The first threshold and the second threshold are set, and satisfy: ; wherein the first threshold and the second threshold are set based on a large amount of historical detection data and industry specifications, combined with joint performance degradation rules and maintenance experience.

[0135] If , it is determined that the performance level of the joint to be evaluated is “healthy”, and the evaluation results are output simultaneously: the performance is good, and no maintenance is needed;

[0136] If , it is determined that the performance level of the joint to be evaluated is “attention”, and the evaluation results are output simultaneously: the performance is declining, and it is recommended to strengthen monitoring;

[0137] If , it is determined that the performance level of the joint to be evaluated is “abnormal”, and the evaluation results are output simultaneously: the performance is deteriorating, and it is recommended to plan maintenance.

[0138] Step 5 is the final link of the intelligent decision-making of the present scheme, and its core significance lies in that by constructing a nonlinear fusion evaluation model, the comprehensive performance deviation coefficient and the comprehensive performance degradation coefficient, two key indicators, are intelligently fused into a comprehensive health index, and based on this, the performance of the joint is accurately classified and the operation and maintenance decision is directly output. Unlike the existing technology which only provides isolated indicators or simple threshold alarms, the health index introduced in the present application is a 0 to continuous and normalized score, which amplifies the negative impact of high-risk state through an exponential function model, and is more consistent with the performance acceleration degradation law in engineering practice.

[0139] The advantages are: first, the decision-making intelligence degree is high, the model automatically fuses static and dynamic information, outputs a single, clear and easy-to-understand health score, greatly reduces the dependence on artificial experience judgment, supports the standardization and automation of operation and maintenance decision-making; second, the operation guidance is strong, the final output is not only a grade label, but also a suggestion directly related to specific maintenance actions, forming a complete closed loop from "data collection" to "operation instruction", which directly implements the value of the evaluation result to the guidance of the field operation level, overcomes the deficiency that the existing technology evaluation result is disconnected with the specific maintenance measures, and significantly improves the accuracy and efficiency of the maintenance work.

[0140] Please refer to Figure 4 , a ballastless track wide-narrow joint performance evaluation system based on multi-sensor data fusion, comprising:

[0141] a theoretical benchmark modeling module, which establishes a theoretical benchmark model of the joint to be evaluated in a healthy state based on joint design standards through a finite element method, and acquires a benchmark feature vector from the theoretical benchmark model, the benchmark feature vector including a theoretical resonance frequency, a theoretical half-power bandwidth, a theoretical damping ratio and a theoretical dynamic stiffness;

[0142] a signal acquisition module, which is used to apply a preset excitation signal to the joint to be evaluated when there is no train operation, and synchronously acquires time-domain input force signals and time-domain output response signals caused by the excitation signal using force sensors and acceleration sensors during the entire excitation action and the excitation decay process;

[0143] a feature vector construction module, which is used to perform frequency domain analysis on the acquired time-domain input force signals and time-domain output response signals, solve the measured frequency response function of the joint, and extract multi-dimensional feature parameters including resonance frequency, half-power bandwidth, damping ratio and dynamic stiffness from the measured frequency response function to jointly constitute a state feature vector for representing the current state of the joint to be evaluated;

[0144] a coefficient calculation module, which is used to compare the state feature vector with the benchmark feature vector stored in the theoretical benchmark model, determine a comprehensive performance deviation coefficient, and retrieve a historical state feature vector corresponding to the latest historical detection record of the joint to be evaluated, determine a comprehensive performance degradation coefficient based on the historical state feature vector;

[0145] a health grading evaluation module, which is used to construct a joint performance evaluation model based on the comprehensive performance deviation coefficient and the comprehensive performance degradation coefficient, determine a comprehensive health degree index of the joint to be evaluated by using the model, and output a grading evaluation result by matching with a preset grading threshold.

[0146] ​The above formulas are all dimensionless values calculated, the formula is obtained by collecting a large amount of data to simulate the most recent real situation, and the preset parameters in the formula are set by a person skilled in the art according to the actual situation.

[0147] The above embodiments can be implemented wholly or partially by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented wholly or partially in the form of a computer program product. Those skilled in the art can realize that the units and algorithm steps of the examples described in connection with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are executed by hardware or software methods depends on the specific application and design constraints of the technical solutions.

[0148] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, which can be located in one place or distributed on multiple network units. Part or all of the units can be selected to achieve the purpose of the embodiments according to actual needs.

[0149] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can easily think of changes or replacements within the technical range disclosed in the present application, which should be covered within the protection scope of the present application.

Claims

1. A method for evaluating the performance of wide and narrow joints in ballastless track based on multi-sensor data fusion, characterized in that, The specific steps include: S1: Based on the joint design standard, a theoretical benchmark model of the joint to be evaluated under a healthy state is established using the finite element method. The benchmark feature vector is obtained based on the theoretical benchmark model. The benchmark feature vector includes the theoretical resonant frequency, theoretical half-power bandwidth, theoretical damping ratio, and theoretical dynamic stiffness. S2: When there is no train running, a preset excitation signal is applied to the joint to be evaluated, and force sensors and acceleration sensors are used to synchronously collect the time-domain input force signal and time-domain output response signal caused by the excitation signal during the entire excitation period and excitation decay process. S3: Perform frequency domain analysis on the collected time-domain input force signal and time-domain output response signal to solve the measured frequency response function of the joint; extract multi-dimensional feature parameters from the measured frequency response function, including resonance frequency, half-power bandwidth, damping ratio and dynamic stiffness, which together constitute a state feature vector to characterize the current state of the joint to be evaluated. S4: Compare the state feature vector with the benchmark feature vector stored in the theoretical benchmark model to determine the overall performance deviation coefficient; retrieve the nearest joint to be evaluated. The historical state feature vector corresponding to each historical detection record is used to determine the overall performance degradation coefficient based on the historical state feature vector. Specifically: retrieve the most recent seam to be evaluated. The historical state feature vectors and corresponding comprehensive performance deviation coefficients corresponding to each historical detection record are arranged in chronological order; the rate of change of the comprehensive performance deviation coefficients of each pair of adjacent historical detection records is calculated, and their arithmetic mean is calculated to obtain the comprehensive performance degradation coefficient; S5: Construct a joint performance evaluation model based on the comprehensive performance deviation coefficient and the comprehensive performance degradation coefficient. Use the model to determine the comprehensive health index of the joint to be evaluated, and output the graded evaluation results by matching it with the preset grading threshold.

2. The method for evaluating the performance of wide and narrow joints of ballastless track based on multi-sensor data fusion according to claim 1, characterized in that: Based on the joint design standards, a theoretical benchmark model is established using the finite element method. The specific logic behind this is as follows: Based on the design drawings of the joint to be evaluated, its geometric parameters are obtained, including the size of the ballastless track slab, the shape of the ballastless track slab, the concrete type, and the ballastless track structure; based on the geometric parameters, a three-dimensional parametric geometric model of the joint area is established. Material properties are assigned to each component in the geometric model, including the elastic modulus, density, and Poisson's ratio of the ballastless track slab. Subsequently, the geometric model is meshed using finite element methods, employing hexahedral dominant elements. Local seeds are set in the weld area and around bolt holes, and the element size of the weld area and surrounding bolt holes is controlled to match the global element size. to To achieve mesh refinement; Modal analysis was performed on the established finite element model to calculate its front-end performance under healthy conditions. Natural frequencies and mode shapes; Based on the modal analysis results, the nodes in the model corresponding to the actual operating positions of the field excitation device are defined as input points. A unit harmonic force excitation is applied to the input points, and the nodes in the model corresponding to the actual installation positions of the field acceleration sensors are defined as output points. The acceleration response of the output points is read, and the theoretical frequency response function between the input points and the output points is calculated. Extract the preceding frequency response from the theoretical frequency response function curve. The theoretical resonant frequency, theoretical half-power bandwidth, and theoretical damping ratio corresponding to the first mode, and the theoretical dynamic stiffness obtained by calculating the reciprocal of the origin dynamic compliance of the theoretical frequency response function, together constitute the reference eigenvector.

3. The method for evaluating the performance of wide and narrow joints of ballastless track based on multi-sensor data fusion according to claim 1, characterized in that: The specific execution process of S2 is as follows: During the track maintenance window when no trains are running, a vibrator is used as the excitation device to apply a preset transient excitation signal to the ballastless track slab; the time-domain input force signal generated by the force sensor installed on the excitation device, as well as the time-domain output response signal generated by the acceleration sensor during the entire excitation process and the excitation decay process are collected simultaneously; the starting point of the excitation signal is used as the synchronous trigger reference for the acquisition process. For longitudinal joints between track slabs, acceleration sensors are fixed on the sides of the track slab support platform on both sides of the joint, 0.3-0.5m from the edge of the joint; for transverse joints, acceleration sensors are installed on the sides of the track slabs on both sides of the joint; the acceleration sensors are connected to the foundation structure through stainless steel brackets.

4. The method for evaluating the performance of wide and narrow joints of ballastless track based on multi-sensor data fusion according to claim 3, characterized in that: Perform Fourier transforms on the acquired time-domain input force signal and time-domain output response signal to obtain the corresponding frequency-domain signals. and ,in accordance with and Calculate the autopower spectrum of the force signal and the cross-power spectrum of the force signal and the response signal. : In the formula, yes The complex conjugate; The measured frequency response function is calculated based on the following formula. : In the formula, (w) is the measured frequency response function. The cross-power spectrum of the force signal and the response signal. The power spectrum of the force signal; From the measured frequency response function Amplitude spectrum | | Identify the peaks in the data, and the frequency corresponding to each peak is the resonant frequency of the corresponding order, denoted as: ,in, Index of order; For the Calculate the two frequencies corresponding to the 3dB drop in amplitude at the first resonance peak. and The half-power bandwidth is calculated using the following formula: In the formula, The bandwidth is the half-power bandwidth of the i-th order; According to the The first resonant frequency and its half-power bandwidth are calculated using the following formula: Step damping ratio : Take amplitude spectrum The average value in the 0-50Hz low-frequency range is used to calculate the dynamic stiffness, and its reciprocal is calculated. In the formula, For dynamic stiffness, Amplitude spectrum The average value in the low-frequency range of 0-50Hz; The extracted front The resonance frequency, half-power bandwidth, damping ratio, and dynamic stiffness are combined to form a state feature vector that characterizes the current state of the joint to be evaluated.

5. The method for evaluating the performance of wide and narrow joints of ballastless track based on multi-sensor data fusion according to claim 4, characterized in that: The state feature vector is compared with the benchmark feature vector stored in the theoretical benchmark model to determine the overall performance deviation coefficient, based on the following formula: In the formula, The comprehensive performance deviation coefficient is used to characterize the degree of deviation between the current overall performance status of the joint being evaluated and the health benchmark. For the feature parameter index used in the comparison, The number of feature parameters used in the comparison, and , It is the order; For the first Each feature parameter value involved in the comparison For the first The reference values ​​of the characteristic parameters are obtained from the reference feature vector acquired by S1; the characteristic parameters include resonant frequency, half-power bandwidth, damping ratio, and dynamic stiffness. The formula used to calculate the overall performance degradation factor is as follows: In the formula, It is a comprehensive performance degradation index, used to characterize the average rate at which the performance status of a joint deteriorates over time. The total number of historical detection records invoked. An index for historical detection records. According to the first The comprehensive performance deviation coefficient is calculated from the historical test records. According to the first The comprehensive performance deviation coefficient calculated from the historical test records; Indicates the first The first historical detection record and the first The time interval between each historical detection record.

6. The method for evaluating the performance of wide and narrow joints of ballastless track based on multi-sensor data fusion according to claim 5, characterized in that: Using the comprehensive performance deviation coefficient and comprehensive performance degradation coefficient as input variables, a joint performance evaluation model is constructed, and the comprehensive health index of the joint to be evaluated is determined using the joint performance evaluation model; the expression of the joint performance evaluation model is as follows: In the formula, The comprehensive health index of the seam to be evaluated is used to characterize the overall health status of the seam to be evaluated. This is the preset upper limit for the health index; It is a natural constant; and As preset weights, and , .

7. The method for evaluating the performance of wide and narrow joints of ballastless track based on multi-sensor data fusion according to claim 6, characterized in that: Based on the matching results between the comprehensive health index and the preset joint performance grading threshold, the grading evaluation results of the joint to be evaluated are output. The specific logic is as follows: Preset first threshold Second threshold And satisfy: ; like If the joint performance level is determined to be "healthy", then the joint to be evaluated is determined to be "healthy". like If so, the performance level of the joint to be evaluated is determined to be "Caution"; like If so, the performance level of the joint to be evaluated is determined to be "abnormal".

8. A performance evaluation system for wide and narrow joints of ballastless track based on multi-sensor data fusion, characterized in that: The aforementioned multi-sensor data fusion-based performance evaluation system for ballastless track width-narrow joints is used to execute the multi-sensor data fusion-based performance evaluation method for ballastless track width-narrow joints as described in any one of claims 1-7, comprising: The theoretical benchmark modeling module, based on the joint design standard, establishes a theoretical benchmark model of the joint to be evaluated under a healthy state using the finite element method. The benchmark feature vector is obtained based on the theoretical benchmark model. The benchmark feature vector includes the theoretical resonant frequency, theoretical half-power bandwidth, theoretical damping ratio, and theoretical dynamic stiffness. The signal acquisition module is used to apply a preset excitation signal to the joint to be evaluated when there is no train running, and to use force sensors and acceleration sensors to synchronously acquire the time-domain input force signal and time-domain output response signal caused by the excitation signal during the entire excitation period and excitation decay process. The feature vector construction module is used to perform frequency domain analysis on the acquired time-domain input force signal and time-domain output response signal to solve the measured frequency response function of the joint; multi-dimensional feature parameters, including resonance frequency, half-power bandwidth, damping ratio and dynamic stiffness, are extracted from the measured frequency response function to form a state feature vector that characterizes the current state of the joint to be evaluated. The coefficient calculation module compares the state feature vector with the benchmark feature vector stored in the theoretical benchmark model to determine the overall performance deviation coefficient; it also retrieves the nearest joint to be evaluated. The historical state feature vector corresponding to each historical detection record is used to determine the overall performance degradation coefficient based on the historical state feature vector. The health grading evaluation module is used to construct a joint performance evaluation model based on the comprehensive performance deviation coefficient and the comprehensive performance degradation coefficient. The model is used to determine the comprehensive health index of the joint to be evaluated, and the grading evaluation result is output by matching it with the preset grading threshold.

Images