1415results about "Railway vehicle testing" patented technology

The invention discloses a train key component online health monitoring method based on multi-sensor signal cooperation, and the method comprises the steps: selecting an indirect signal as an edge end train online abnormal state early warning index based on the correlation analysis of multi-sensor signals, and diagnosing whether the state of a train key component is abnormal or not; the vibration signals serve as signals for further diagnosis at the cloud, state features of the vibration signals are extracted at the edge end, online optimization of sparse coding is carried out, and the vibration signals are transmitted to the cloud to serve as diagnosis signal features for further diagnosis; real-time training precision and sample complexity are integrated into online estimation of a model training process, optimal features are transferred from a source domain-edge end to a target domain-cloud end, and the cloud end diagnoses the type and grade of a fault by means of the complex operation capability of the cloud end; a dynamic aggregation strategy is provided, the features of a prediction model are determined, optimization of nodes is enhanced, and a feature network is made to adapt to and complete an online prediction task.

The invention provides a state monitoring and fault early warning method for a maglev vehicle system, relates to the technical field of urban rail transit data processing, and solves the problem of how to effectively combine an empirical threshold mode and a machine learning technology and give full play to a combination effect. The method comprises the steps of collecting operation data of a magnetic levitation vehicle system, performing corresponding data preprocessing operation, performing analysis and calculation in a mode of constructing a mathematical model, and evaluating the health state of the magnetic levitation vehicle system; constructing and training a fault prediction model, and inputting the evaluation data into the trained fault prediction model to obtain potential fault data of the magnetic levitation vehicle system; and generating early warning information based on the potential fault data, retrieving an expert experience library by using the early warning information, generating an intelligent decision result, and feeding back the intelligent decision result to the operation and maintenance personnel. According to the method, machine learning and expert experience are effectively combined, so that the real-time performance and the accuracy of fault early warning are improved.

The invention discloses a railway tunnel full-life-cycle model test system and test method considering train vibration, and relates to the technical field of railway tunnel engineering. The model test system adopts modular design and comprises a box frame module, a constant-pressure loading module, an excitation system module, a seepage system module and a monitoring system module. The test method comprises the steps of physical similar model laying, constant-pressure loading control, tunnel excavation and supporting, vibration loading and data monitoring analysis. By accurately simulating an initial stress field and train vibration loads in the construction and operation stages of the tunnel, dynamic monitoring of stress strain, deformation and damage processes of the tunnel surrounding rock and the lining structure in the whole life cycle is achieved. The device has the advantages of compact structure, modular design, high loading precision and stability, real simulation and the like, and a scientific basis can be provided for the fields of performance evaluation, structure optimization, reinforcement design and the like in the whole life cycle of the railway tunnel.

The invention relates to the technical field of artificial intelligence and computer vision, discloses a subway key component fault detection method and system based on an AI visual large model, and aims to solve the problems of low detection precision, weak generalization ability, insufficient multi-mode understanding, poor real-time performance and lack of state evolution modeling in the prior art. The method comprises the following steps: acquiring images of key components through a multi-view industrial camera array, and performing distortion correction, illumination normalization and noise suppression; a pre-trained visual large model is utilized to extract deep space features, and modeling is carried out on a continuous frame feature sequence through bidirectional LSTM to capture a time sequence change trend. By introducing the large-scale visual large model and spatio-temporal joint modeling, the identification capability of tiny defects is improved, the discrimination stability is enhanced, high-precision and low-delay automatic detection is realized, the false alarm rate and the omission ratio are remarkably reduced, and the detection efficiency and the system maintainability are improved.

The invention provides an automatic train inspection method, device and system and a storage medium. The method comprises the steps that a preset trench map is obtained, and the trench map comprises the preset parking position of a train and the position of each detection point under the train; controlling the automatic inspection robot in the trench to move towards the train according to the preset parking position, and acquiring a laser scanning signal of the automatic inspection robot in real time; if the laser scanning signal shows that the automatic inspection robot arrives at one end of the train, updating the position of each detection point according to the current position of the automatic inspection robot; and controlling the automatic inspection robot to move towards the other end of the train, and inspecting the train according to the updated position of each detection point in the moving process. The train inspection method can be simplified, and the inspection efficiency is improved.

The invention provides a wheel tread derailment detection method and system for a railway vehicle. According to the method, jacking driving devices with jacking rotating wheels are symmetrically arranged on the two longitudinal sides of a track, a pressure sensing unit is arranged between the two jacking rotating wheels, optical detection units with orthogonal optical paths are installed on the outer sides, and displacement measurement units are integrated on the jacking rotating wheels, so that a three-dimensional detection network is formed. During detection, a top rotating wheel lifts a wheel step by step, a blocking signal, tread pressure and a vertical displacement parameter are synchronously collected, and a space mapping relation between a pressure decreasing path and a displacement abnormal point is established; converting the blocking signal into a blocking release rate, and calculating a displacement accumulated value in combination with a space mapping relation; and when the displacement accumulated value is continuously increased and the blocking release rate exceeds a threshold value, wheel-rail separation is judged, and jacking is controlled to be stopped. Through cooperative detection and analysis of multiple sensors, accurate judgment and self-adaptive jacking control of the wheel-rail separation state are achieved, and the problem of wheel diameter adaptability of a traditional fixed jacking mode is effectively solved.

The invention discloses a high-speed train body health monitoring system and method based on a distributed IMU, and belongs to the technical field of mechanical vibration testing, and the method comprises the steps: obtaining a preset structure topology model of a high-speed train body, recognizing a key structure region, and constructing a physical topology grouping mapping table; deploying a plurality of inertial measurement unit nodes to the key structure area and carrying out relevance division to obtain a physical topology group; the method comprises the following steps: receiving a real-time operation condition signal from a train control system, performing feature extraction on vibration data acquired by inertial measurement unit nodes in a physical topology group, generating a region collaborative feature vector, performing space-time anomaly analysis, positioning a damage generation region, and generating a residual life evaluation parameter according to the vibration feature variation of the damage generation region. According to the method, physical topology grouping mapping and working condition self-adaptive feature extraction technologies are adopted, space-time transmission consistency verification and an aging dynamic correction mechanism are combined, and accurate positioning of a damaged area and reliable evaluation of the residual life can be achieved.

The invention discloses a high-speed rail vehicle-mounted equipment fault prediction and health management method, and relates to the technical field of high-speed rail vehicle-mounted equipment fault prediction, and the method comprises the steps: monitoring high-speed rail vehicle-mounted equipment through employing a high-precision sensor network, and obtaining an original monitoring data set; cleaning the original monitoring data set by adopting a multi-source data integration and preprocessing method to obtain a high-quality data set, and extracting key features based on the high-quality data set to obtain a feature data set; constructing a fault prediction model based on a deep learning method and the feature data set, inputting the high-quality data set to the fault prediction model, and outputting a fault early warning result; based on the fault early warning result, a decision support method is adopted to evaluate the equipment state, and a maintenance strategy suggestion is obtained; obtaining a comprehensive analysis report based on the maintenance strategy suggestion and the equipment evaluation state; environmental data and historical maintenance record data from a ground monitoring system are integrated, and the data are fused by using a weighted average method.

The invention discloses a motor bearing fault diagnosis method, device and equipment and a medium, and relates to the technical field of motor bearing fault diagnosis, and the method comprises the steps: obtaining a vibration signal and a current signal of a motor bearing, and carrying out the feature extraction of the vibration signal and the current signal, obtaining a first time domain feature, a first frequency domain feature and a first energy feature of the vibration signal and a second time domain feature, a second frequency domain feature and a second energy feature of the current signal; carrying out feature fusion and feature dimension reduction on the first time domain feature, the first frequency domain feature, the first energy feature, the second time domain feature, the second frequency domain feature and the second energy feature to obtain a feature after fusion dimension reduction, and taking the feature after fusion dimension reduction as input; and determining whether the motor bearing has a fault and the fault type when the fault exists by using the fault diagnosis model. According to the invention, the fault diagnosis precision can be improved.

A wheel test device includes a rail wheel support rotatably supporting a rail wheel, a wheel support rotatably supporting a test wheel in contact with the rail wheel, a first electric motor, a power distributor distributing power generated by the first electric motor to the rail wheel and the test wheel, and a torque generator generating torque to be applied to the test wheel. The torque generator includes a rotating frame rotationally driven by the first electric motor, and a second electric motor mounted on the rotating frame. One of or both the rail wheel and the test wheel is connected to the first electric motor via the torque generator. A rated output of the second electric motor is equal to or more than 3 kW, and moment of inertia of a rotating part of the second electric motor is equal to or less than 0.01 kg·m2.

The invention relates to the field of gear fault diagnosis, in particular to a rack rail train driving gear fault diagnosis method, system and device, and solves the problem that fault features cannot be effectively extracted from rack rail section vibration signals in the prior art. The method comprises the following steps: identifying a transition section vibration signal of a transition device; identifying a toothed rail section vibration signal; the rack rail section vibration signals are processed; and fault diagnosis is realized according to the meshing frequency and the sideband information. According to the time domain characteristics of the vibration signals, the vibration signals generated when the driving gear passes through the rack rail section can be accurately recognized from the vibration signals, and self-adaptive recognition of the vibration signals of the rack rail section is achieved. According to the method, noise interference in time-frequency distribution of toothed rail section vibration signals is removed, time-frequency aggregation of the time-frequency distribution is improved, fault features of the driving gear are extracted from the time-frequency distribution, fault diagnosis of the driving gear is achieved, health state monitoring of the toothed rail train driving gear is achieved, and train operation safety is guaranteed.

The invention discloses a bogie detection system and a control method thereof, and the system comprises a guide rail which is provided with a detection station; the height detection device is arranged above the detection station, and the height detection device is used for detecting the height of the bogie; the first image detection device is arranged below the detection station, and the first image detection device is used for carrying out image detection on the bottom of the bogie while carrying out height detection; and the second image detection device is arranged on the side surface of the detection station and is used for carrying out image detection on the side surface of the bogie while carrying out height detection. Bogie height parameter and multi-view image acquisition and multi-modal data fusion analysis are synchronously completed at a single station, the defect recognition accuracy is improved, the labor intensity of detection personnel is remarkably reduced, the repeated configuration of equipment is reduced through station integrated design, the detection period is effectively shortened, and the workshop space is saved. And an efficient intelligent detection solution is provided for manufacturing of rail transit equipment.

The invention discloses an electric inertia simulation braking power test bench and an electric inertia control method thereof, and relates to the technical field of electric transmission simulation of flywheels, the electric inertia simulation braking power test bench comprises a motor, an incremental encoder, a main shaft, a flywheel, a force arm frame, a mechanical friction braking torque sensor, an electric transmission control system and a brake disc; an incremental encoder is installed at the rear end of the motor, the transmission end of the motor is connected with the main shaft through a coupler, the flywheel is installed on the main shaft, a brake disc is installed on the force arm frame, and meanwhile the force arm frame is connected with a mechanical friction braking torque sensor. According to the invention, the angular speed, the angular displacement and the mechanical friction braking torque are synchronously collected, and the motor torque is calculated and given, so that the test bench operates in a state of approaching an equivalent flywheel and equivalently having no system resistance; the algorithm can control the total error within a very small range; accurate matching of rotational inertia is achieved, and the control precision can meet the high-precision test requirement. And investment and operation cost are saved.

A door leaf abnormity identification method based on a door plate state and a door leaf monitoring system relate to the field of door leaf detection, and the method comprises the following steps: determining whether a vehicle door is locked in place based on a motor encoder, a door in-place switch and a lock in-place switch; when it is judged that locking is in place, the state of the door leaf monitoring switch is monitored, and the door leaf monitoring switch is used for monitoring the state of the door leaf; if the vehicle door is locked in place and the door leaf monitoring switch is in a non-closed state, judging that the door leaf state of the door leaf is an abnormal state; and when the door leaf state is an abnormal state, generating a door leaf abnormal fault record based on the abnormal state, and sending the door leaf abnormal fault record to the first client. By implementing the method, the reliability of subway door detection can be improved.

The invention provides the unsupervised learning-based fault diagnosis method and system for detecting wheel out of round, which belongs to the technical field of machine learning fault diagnosis. Characteristic signals of train wheels will be acquired. The pre-trained detection model is adopted to process the characteristic signals of the train wheels to be detected, so as to obtain the wheel roundness state results. As for the collected characteristic signals, the invention constructs a subway wheel out of round detection method based on unsupervised learning for collected characteristic signals, which are deployed to computer equipment capable of executing computer programs, inputting characteristic signal data collected by a subway wheel out of round monitoring device based on rail wayside response into the computer equipment to obtain the wheel roundness state.

The invention relates to the technical field of rail transit monitoring, and discloses a multi-sensor fusion rail transit electric passenger car bogie dynamic performance monitoring method. The method comprises the steps that multi-source heterogeneous sensing data such as bogie vibration acceleration time sequence signals, wheel-rail contact force distribution data, bearing temperature gradient data and structural strain field data are collected; analyzing the vibration signal to extract a frequency band energy feature vector, performing spatial gridding mapping on the contact force data to generate a distribution matrix, and performing time-space alignment fusion on the two to form a primary fusion feature set; positioning a bearing temperature anomaly region through a deep convolutional network, and generating a secondary fusion feature set in combination with a structural strain concentration coefficient; and inputting the two-stage feature set into a dynamic performance evaluation model, outputting a performance degradation index set of scores of axle box bearing wear degree, framework fatigue cumulant, wheel set tread damage and the like, generating a maintenance priority sequence and triggering a preventive maintenance instruction according to the performance degradation index set, and realizing accurate monitoring and efficient operation and maintenance management and control of the dynamic performance of the bogie.

The invention provides a train sliding plug door fault diagnosis method and system based on an acoustic signal, and belongs to the technical field of fault diagnosis, and the method specifically comprises the steps: carrying out the filtering processing of a preset type of interference signal in the acoustic signal, obtaining a processing signal, decomposing the processing signal, obtaining a plurality of dimensions of acoustic feature signals, and carrying out the processing of the acoustic feature signals; the method comprises the following steps: acquiring an acoustic characteristic signal of a sliding plug door, sending the acoustic characteristic signal to a prediction model of a Transform-based double-branch model framework, determining the abnormal probability of the sliding plug door under different fault types, and when the fault type of which the abnormal probability does not meet the requirement exists, acquiring a measurement result of a historical acoustic signal in an adjacent time period, according to the method, the consistency of the multiple groups of historical acoustic signals in the adjacent time period and the acoustic signals is determined, and whether an early warning signal needs to be output or not is determined by combining the historical acoustic signals, the abnormal probability of the fault type of which the abnormal probability does not meet the requirement under the acoustic signals and the measurement reliability, so that the accuracy of fault diagnosis processing is improved.

The invention relates to the technical field of subway tunnel fire experiment, and discloses a variable-pressure reduced-size subway tunnel / train compartment double-long-narrow-space fire experiment platform which comprises a tunnel body, a ventilation unit, a monitoring unit, a rail unit and a train unit, and a fire source structure is arranged in the train unit. According to the variable-pressure reduced-size subway tunnel / train compartment double-long-narrow-space fire experiment platform, a fire source structure is arranged, when a fire experiment needs to be carried out, a top fan, two side fans and a main fan are started through a wireless receiver, the wind speed during high-speed running of a train is simulated, and then a remote control igniter is remotely and wirelessly connected; the top fan, the two side fans and the main fan are remotely turned off after a certain time, a real fire disaster is simulated, the fire source structure can simulate the effect that the fire source continuously burns the multiple carriages after a train runs at a high speed, and the authenticity is high.

The invention discloses a heavy-load train coupler swing angle measuring method based on machine vision, and relates to the technical field of heavy-load train coupler swing angle measurement. The method comprises the following steps: S1, acquiring an image of a coupler area of a heavy-load train, and preprocessing the image to obtain a preprocessed coupler area image; s2, performing coupler feature target identification on the preprocessed coupler region image to obtain pixel coordinates of a coupler; and S3, calculating a coupler swing angle according to the pixel coordinates of the coupler. The coupler swing angle of the heavy-load train is quantitatively recognized based on machine vision, high-precision non-contact measurement of the coupler swing angle of the heavy-load train is achieved, and the method has the advantages of being low in installation and maintenance cost, small in potential safety hazard and the like and is expected to be developed into a long-term monitoring method for the coupler swing angle of the heavy-load train.

The invention relates to the technical field of wheel-rail force, in particular to a train stability analysis method and system based on wheel-rail force, and the method comprises the following steps: based on vertical force and transverse force data collected in real time, carrying out segmented statistics on stress value distribution of contact points of train wheels and a track, and analyzing abnormal amplitude in force fluctuation values. According to the method, the stress values of the contact points of the train wheels and the rail are counted in a segmented mode, abnormal amplitude analysis and periodic abnormal parameter screening are combined, the irregular vibration frequency and position are effectively recognized, the vibration recognition accuracy and processing speed are improved, the triggering sensitivity and the working threshold value of the force sensor are adjusted, and the accuracy of vibration recognition is improved. In combination with optimization of frequency distribution of a track irregularity area, the response capability and adaptability of a monitoring system are enhanced, through a waveform characteristic parameter set extracted from a wheel power signal, the influence of track gauge change on waveform characteristics can be evaluated more comprehensively, dynamic stability analysis of a train is optimized, and data support is provided for train stability.

The invention discloses a recreation facility track error measuring vehicle, and relates to the field of recreation facility track detection, a driving system comprises two driving devices symmetrically arranged on the two sides of the front portion of a vehicle frame, and the driving devices are used for being clamped on a track and can drive the vehicle frame to run along the track; the front measuring system comprises a horizontal sensor and two front walking devices symmetrically arranged on the two sides of the rear portion of the vehicle frame, and the horizontal sensor is arranged on the vehicle frame. The rear measuring system comprises a middle piece, an acceleration sensor and two rear walking devices symmetrically arranged on the two sides of the middle piece, the front end of the connecting piece is connected with the frame through a first universal connecting part, the rear end of the connecting piece is connected with the middle piece through a second universal connecting part, and the acceleration sensor is arranged on the middle piece; a marking mechanism is arranged on the rear walking device and used for marking on the track. The recreation facility track error measuring vehicle can measure manufacturing errors and installation errors at the same time, and the measuring efficiency is improved.

The invention relates to the technical field of rail traffic monitoring, and discloses an urban rail vehicle driving analysis system based on digital twinning, which is characterized in that vehicle vibration and traction motor current signals are collected in real time through a vehicle-mounted sensing unit, and an edge feature extraction unit calculates energy ratio features of a predefined fault frequency band; the data transmission unit transmits the characteristic data and the working condition information to a digital twin platform, a digital fault tree model is arranged in the platform, and a cross-physical domain fault conduction path is analyzed and recognized through multi-source data coupling. Through combination of frequency band energy ratio characteristics and fault tree logical reasoning, electromechanical coupling fault risks which cannot be found by traditional isolated monitoring are effectively captured, meanwhile, a diagnosis threshold value is dynamically adjusted according to the wheel-rail adhesion state, and the diagnosis reliability under the complex working condition is improved.

The invention discloses an intelligent train bogie defect detection and fault identification system and method. The intelligent train bogie defect detection and fault identification system comprises a control module, a data acquisition module, a single-chip microcomputer module and a data sending module. The control module controls other modules to work and transmit data; the data acquisition module, the single chip microcomputer module and the data sending module are respectively connected with the control module; the data acquisition module comprises a high-definition imaging unit, a radar ranging unit and an infrared temperature measurement unit; the single-chip microcomputer module comprises an image processing unit, a defect detection unit and a fault identification unit. The data sending module comprises a coding unit and a wireless communication unit; and the fault judgment unit clusters the processed fault feature map and the fault information trained by the database to obtain the fault information, so that intelligent train bogie defect detection and fault identification are realized, the labor amount of workers is reduced, the bogie maintenance efficiency is improved, and the train operation safety is guaranteed.

The invention provides a water-rich stratum tunnel train simulation excitation signal generation method and device, and relates to the technical field of engineering test, and the method comprises the steps: obtaining train operation parameters, wheel track vibration waves and track characteristic parameters; performing frequency band division on the wheel track vibration wave, calculating time delay, performing phase compensation on each frequency band according to the time delay, and generating a composite waveform to obtain a first signal; performing frequency domain correction on the first signal according to the train operation parameters, and adjusting the amplitude of each frequency band according to the corrected normalized energy of each frequency band to obtain a second signal; frequency bands of the second signal are divided according to the track characteristic parameters, the center frequency of each frequency band is determined, phase compensation is carried out according to impedance corresponding to the center frequency, and a third signal is obtained; and generating a simulation excitation signal based on the third signal and the train operation parameters. The excitation signal generated by the invention has the composite characteristics of high-frequency transient impact and low-frequency rigid motion, and can provide high-fidelity excitation input for related tests of train vibration load.

The utility model discloses an integrated rigidity measuring device for a tumbler node. The integrated rigidity measuring device comprises integrated rigidity measuring equipment and auxiliary equipment, the integrated rigidity measuring device comprises a bolt, an inner positioning pin, a sleeve, a clamp spring and a force transmission pin. The two to-be-tested rotating arm nodes are arranged face to face and are connected through a force transmission pin; an inner positioning pin is arranged in the matching surface of the two to-be-tested rotating arm nodes, and the inner positioning pin penetrates through the two to-be-tested rotating arm nodes to form a fixed whole; and the fixed whole body is placed in the sleeve. According to the utility model, two nodes can be effectively combined, relative slippage caused by assembly factors is avoided, and the measurement precision can be greatly improved. The two nodes are combined together through the sleeve, the sleeve and the nodes are in interference fit and are installed into a whole, rigidity tests in the axial direction, the radial direction, the torsion direction and the deflection direction can be carried out, repeated disassembly and assembly are not needed, and measurement convenience is improved.

The invention discloses a high-speed train bogie gearbox bearing fault diagnosis method, which comprises the following steps of: acquiring vibration acceleration signals of a bearing in a train bogie gearbox in different health states, preprocessing the vibration acceleration signals into a training sample set and a test sample set, and constructing a fault diagnosis classification model of noise label robustness; and training the classifier model by adopting a self-paced learning strategy, and inputting the test sample set into the trained classification model to obtain a train bogie gearbox bearing fault diagnosis result. According to the method, through the provided adaptive course sample selection strategy, the threshold value of confidence sample selection is associated with the development of the network, and the introduction of a resistance loss function can effectively resist the over-fitting problem easily generated by the model on a noise label. According to the method, the diagnosis accuracy of the data set with the noise label can be improved, and the high-speed train bogie gearbox bearing fault diagnosis precision is improved under the condition of high noise label rate.

The invention relates to the technical field of train fault diagnosis, in particular to a fault diagnosis method for a train transmission system, and overcomes the defects that weak features are difficult to extract and fault features are difficult to effectively distinguish in the prior art. In combination with the strong feature extraction capability of a deep network model constructed by an InceptionResNetV2 network, local and global features are captured by using a parallel multi-branch structure and a residual adaptive fusion mechanism, and through multi-scale perception and residual cross-layer connection, the capture capability of the super-huge bearing weak fault time sequence-frequency domain features can be significantly enhanced, and the fault detection accuracy is improved. And finally, pre-training a model by using source domain data, extracting parameterized knowledge of the model, and migrating to a target domain to realize parameter fine tuning and feature adaptation, so that the generalization performance and convergence speed of fault diagnosis can be improved, different fault categories can be effectively separated, and feature aggregation of the same category can be realized.

Railcar inspection systems, methods, and apparatuses are disclosed, including a railcar inspection portal. The railcar inspection portal includes a physical structure positioned around a railroad track, and through which a railcar can travel. The railcar inspection portal can include wheel detection sensors along the railroad track for detecting the presence of a railcar passing over the sensors. The sensors can transmit signals, corresponding to railcars passing over the sensors, to computing devices for determining railcar speeds. The railcar inspection portal can include imaging devices configured to capture images and readings of railcars passing through the inspection portal. Based on a determined speed corresponding to a passing railcar, the computing devices can control the imaging devices to capture specific areas or components of the passing railcar, or individual cars thereon. The computing devices can process the captured images to detect defects corresponding to the passing railcar, or individual cars thereon.

The invention discloses a railway vehicle body abnormal vibration diagnosis and recognition method. The method comprises the following steps that 1, an accelerometer sensor is used for collecting vehicle body vibration signals; step 2, using EMD to decompose the vibration signal x (t), wherein each mode IMFi (t) after decomposition corresponds to a certain frequency component in the vibration signal; step 3, after preprocessing the data, selecting modal data of the first 3-5 orders, and then performing fast Fourier transform on each modal IMFi (t) to obtain frequency spectrum information of each modal IMFi (t); step 4, smoothing processing is carried out on the frequency spectrum image; 5, calculating the energy ratio P of the vibration dominant frequency; step 6, calculating a performance index Q; and 7, judging whether the vehicle is in an abnormal state or not according to the P and the Q. According to the method, the vibration energy ratio index P and the performance index Q are used in cooperation, whether the stability, comfort and other performance of vehicle operation are good or not can be clearly known, and meanwhile whether the vehicle body or the bogie has obvious harmonic vibration or not can be judged.

The invention relates to the technical field of train data analysis, in particular to a data analysis method for train operation faults, which comprises the following steps: carrying out data fusion processing on train operation brake signals acquired by a vibration sensor, an acoustic emission sensor and a state detection device to generate multi-dimensional brake time sequence data; brake fusion features are generated from the multi-dimensional brake time sequence data through a cross-modal fusion model, and the cross-modal fusion model is constructed based on a gating feature interaction layer and a convolution feature extraction layer; and the brake fusion features are used to generate brake fault severity through a space-time joint model, the space-time joint model is constructed based on an improved LSTM architecture, and optimization training is carried out through a physics constraint loss function. According to the method, the accuracy and reliability of predictive fault diagnosis of the subway braking system are remarkably improved through the LSTM architecture of multi-modal data fusion, cross-modal feature modeling and physical constraint embedding on the train operation braking data detected beside the track and the detection data of the braking system.