Railway vehicle-road-network-tunnel-terrain coupled ice and snow disaster monitoring and early warning system

Abstract

The invention provides a railway vehicle-road-network-tunnel-terrain coupling ice and snow disaster monitoring and early warning system. The railway vehicle-road-network-tunnel-terrain coupling ice and snow disaster monitoring and early warning system comprises a global stereoscopic perception subsystem, a multi-source data fusion communication platform, a multi-physics coupling analysis engine and an intelligent decision and visualization terminal. The global three-dimensional sensing subsystem is used for collecting multi-dimensional ice and snow disaster data of a train bogie, an equipment compartment, a line, a tunnel, an overhead line system, terrain and the like; the multi-source data fusion communication platform is connected with the sensing subsystem to realize fusion and transmission of multi-source heterogeneous data; the multi-physics field coupling analysis engine analyzes the fused data and predicts various ice and snow disaster risks; and the intelligent decision-making and visualization terminal generates early warning information and a control instruction according to the analysis result, and carries out visual display. The system is suitable for a plateau alpine environment, and can realize all-weather, full-coverage and intelligent ice and snow disaster monitoring and early warning.

Description

Railway vehicle-track-network-tunnel-terrain coupled snow and ice disaster monitoring and early warning system Technical Field

[0001] This invention relates to the field of railway disaster prevention and safety monitoring technology, and in particular to a snow and ice disaster monitoring and early warning system that integrates multi-dimensional data of vehicles, tracks (including turnouts), tunnels, overhead contact lines and terrain for complex environments in high-altitude and cold regions. Background Technology

[0002] High-altitude railways are located in frigid regions, where snow and ice disasters are complex and severe, including: blowing snow: strong winds carry snow to the roadbed and tracks, burying the lines and hindering train passage; icing: low temperatures and snow cause ice to form on the overhead contact system, switches, and rail surfaces, reducing the rail adhesion coefficient and affecting train braking and traction performance, while also interfering with the normal operation of electrical equipment; avalanche: snow-covered mountains along the line collapse, impacting and burying railway facilities, directly threatening train safety; tunnel icing: long tunnels in high-altitude areas have complex geological structures and abundant groundwater resources. At extremely low temperatures, water flowing inside the tunnels easily freezes, and the resulting icicles may scrape, fall, or freeze the tracks, endangering train safety; train snow and ice accumulation: trains operate in frigid and snowy environments for extended periods, and severe snow and ice accumulation can easily occur in the bogies and equipment compartments, deteriorating train operation quality and even causing safety accidents.

[0003] Current monitoring methods for ice and snow disasters on plateau railways mainly include: manual inspection and fixed-point meteorological station monitoring: the risk of disaster is judged by regularly inspecting the condition of the line by manpower and combining it with meteorological data collected by fixed-point meteorological stations; video surveillance: the line condition is captured in real time by cameras deployed along the line and disasters are identified by manpower; discrete sensor monitoring: the condition of the line or equipment in a specific area is perceived by sensors deployed in the track circuit or locally.

[0004] Existing solutions suffer from the following drawbacks: Limited and singular monitoring methods: manual patrols are inefficient and risky; fixed weather stations have limited coverage and cannot achieve dynamic tracking; video surveillance is susceptible to wind and snow, failing in low visibility conditions, and relies on manual interpretation, leading to delayed responses; discrete sensors can only perceive local conditions, lacking a global perspective; significant data silos: monitoring data from vehicles, lines, power grids, terrain, and other environments are independent, lacking effective correlation and hindering system-level coupled analysis; one-sided perception and delayed early warning: existing technologies struggle to comprehensively and in real-time reflect the comprehensive risks of the complex coupled system of "vehicle-road-tunnel-network-terrain," with early warnings often being passive alerts after disasters occur, lacking forward-looking predictions based on multi-source data fusion; insufficient adaptability: monitoring schemes for ordinary plains areas do not consider the extreme and variable terrain and climate characteristics of plateaus, easily leading to "acclimatization problems" in plateau environments and failing to meet the needs of precise monitoring; lack of comprehensive risk assessment capabilities: existing technologies struggle to achieve comprehensive risk assessment and collaborative early warning for multiple disaster types, resulting in low disaster prevention and control efficiency.

[0005] Therefore, there is an urgent need to build a high-altitude railway ice and snow disaster monitoring and early warning system that covers multi-dimensional monitoring of the coupled system of "vehicle-road-network-tunnel-terrain", has all-weather, full-coverage, intelligent and forward-looking early warning capabilities, and realizes the upgrade from decentralized monitoring to full-domain coupled early warning, and from passive response to active prevention and control. Summary of the Invention

[0006] This invention provides a vehicle-road-network-tunnel-terrain coupled plateau railway ice and snow disaster monitoring and early warning system, which aims to solve the problems of incomplete monitoring and delayed early warning in the complex ice and snow environment of plateau railways. The overall architecture of this system includes four parts: a full-domain three-dimensional perception subsystem, a multi-source data fusion communication platform, a multi-physics field coupling analysis engine, and an intelligent decision and visualization terminal. Each part works in concert to realize intelligent processing of the entire process from data collection, transmission, analysis to decision and early warning.

[0007] To address the aforementioned technical problems, this invention provides the following technical solution: The railway vehicle-track-network-tunnel-terrain coupled snow and ice disaster monitoring and early warning system proposed in this solution includes: a full-domain three-dimensional perception subsystem, used to comprehensively collect snow and ice disaster-related data from six dimensions: train bogies and equipment compartments, lines, tracks, tunnels, overhead contact lines, and terrain along the line; a multi-source data fusion communication platform, connected to the full-domain three-dimensional perception subsystem, used to fuse and transmit the collected multi-source heterogeneous data; a multi-physics field coupling analysis engine, connected to the multi-source data fusion communication platform, used to perform coupled analysis on the fused data and predict various snow and ice disaster risks; and an intelligent decision-making and visualization terminal, connected to the multi-physics field coupling analysis engine, used to generate early warning information and control commands based on the analysis results and to perform visual display.

[0008] Furthermore, the comprehensive three-dimensional perception subsystem includes: a vehicle-based mobile monitoring unit for monitoring the snow and ice accumulation status, track snow cover status, and rail adhesion status of the train bogie and equipment compartment areas; a roadbed fixed monitoring network for monitoring the snow and ice accumulation on turnouts, the dynamic evolution of snow blowing disasters on the track, and the working status of snow melting and de-icing devices; a tunnel environment monitoring unit for monitoring ice accumulation on tunnel linings, dark ice on the rail surface, micro-meteorological gradients, and ground temperature; a catenary icing monitoring unit for monitoring the types of icing on the catenary conductors (rime, rime, mixed rime), icing load, and insulator working status; and a terrain environment monitoring unit for acquiring data on snow thickness, stability, and avalanche precursor signals along the track.

[0009] The vehicle-based mobile monitoring unit includes: a miniature, low-temperature resistant industrial endoscope camera, a supplementary lighting system, and a high-frequency vibration accelerometer installed on key parts of the bogie, used to monitor the thickness of snow and ice accumulation in the bogie area and invert the icing quality of the bogie area through vibration spectrum signals. A temperature, humidity, and pressure differential composite sensor and an infrared thermal imaging probe deployed in the equipment compartment are used to determine filter blockage and monitor the temperature of key components; a train bus data reading module is used to extract real-time train wheel slip / slip signals and real-time changes in traction / braking force data, and invert the change law of the adhesion coefficient of the rail surface caused by snow and ice accumulation. A long-focal-length optical zoom industrial camera deployed on the top of the train's front car is used to clearly capture the icing morphology of the contact wires and catenary cables from a distance, and the relevant collected information is synchronously transmitted to the contact wire icing monitoring unit to assist in the assessment of the contact wire icing status. Furthermore, when the train passes through a tunnel, this long-focal-length optical zoom industrial camera is used to monitor the spatial distribution and size characteristics of ice accumulation on the tunnel lining, and the relevant collected information is synchronously transmitted to the tunnel environment monitoring unit to assist in the assessment of the tunnel lining icing status. Low-temperature resistant wide-angle industrial cameras and lidar deployed at the bottom of the train's locomotive are used to clearly capture the snow and ice cover on the track surface, snow and ice accumulation on switches, and snow cover patterns in the track area. The collected information further assists in monitoring changes in track adhesion coefficient, track snow cover status, and switch snow and ice accumulation conditions. In addition, when the train passes through tunnels, the wide-angle industrial cameras and lidar are also used to simultaneously collect data on the distribution of dark ice on the track surface inside the tunnel, and transmit the relevant information to the tunnel environmental monitoring unit to assist in assessing the tunnel's ice and snow disaster status.

[0010] Furthermore, the roadbed fixed monitoring network includes: fiber optic temperature / stress sensors deployed in the gaps between turnout sleepers to monitor freezing stress; millimeter-wave radar to scan the gap between the switch rail and the stock rail to identify snow accumulation and ice debris; a turnout snow melting device operation status feedback module; and ultrasonic snow depth arrays, small anemometers, and low-temperature resistant wide-angle industrial cameras deployed along the windward side of the line to monitor the snow cover status within the line and the evolution of windblown snow from outside the line to inside the line.

[0011] Furthermore, the tunnel environment monitoring unit includes: a dual-spectrum pan-tilt camera deployed at the tunnel entrance in areas of severe icing and at key seepage points inside the tunnel; temperature and humidity sensors and ground temperature sensors deployed at intervals along the tunnel's depth, used to construct temperature and humidity gradient maps inside and outside the tunnel.

[0012] Furthermore, the contact wire icing monitoring unit includes: a miniature weather station, a dual-spectrum camera (visible light + thermal imaging), a contact wire tension sensor, and an insulator tilt sensor, used to monitor the type of contact wire icing (rime, hoarfrost, mixed rime) and icing load in key sections.

[0013] The terrain environment monitoring unit includes: satellite remote sensing or UAV-borne synthetic aperture radar for periodically scanning the slopes along the route; and snow pressure cells and ground acoustic sensors deployed on the slopes of avalanche-prone areas.

[0014] Furthermore, the multi-source data fusion communication platform includes: a multi-source heterogeneous data fusion engine, used to align, calibrate and deeply fuse monitoring data from different sources, formats and spatiotemporal resolutions under a unified spatiotemporal reference framework to generate a panoramic view of the line health status; and a multi-layer resilient communication network, which comprehensively adopts railway-specific 5G-R, satellite communication and wireless Mesh network to achieve reliable and low-latency transmission of key data in complex plateau environments.

[0015] Furthermore, the multiphysics coupling analysis algorithm engine includes: a coupled numerical model of wind and snow migration-erosion-deposition multi-transport characteristics in railway subgrade areas, used to predict the accumulation location and rate of blown snow in the subgrade and turnout areas based on digital elevation model data, real-time wind speed and direction data, and snow cover morphology in key areas; a coupled numerical model of wind, snow, water, and ice multi-phase transport in train bogies / equipment compartments, used to predict the snow and ice accumulation quality and spatial distribution characteristics in the train bogie and equipment compartment areas based on data such as train speed, refined geometric models of train bogies and equipment compartments, ambient temperature and humidity, snow cover thickness, and snow cover physical properties; and a quasi-steady-state numerical model of the evolution characteristics of icing morphology in railway catenary / rail surface. The system includes: a train speed and departure density model for predicting the dynamic growth characteristics of icing on railway catenary and track surfaces based on data such as train speed, departure density, ambient temperature and humidity, catenary current, and freezing rain / snowfall type; a cold-region tunnel-atmosphere thermodynamic coupling model for predicting the dynamic movement of icing boundaries within tunnels based on tunnel internal temperature data, spatial temperature linear distribution data, external temperature data, and seepage location and flow parameters within tunnels; and a numerical model for avalanche dynamic risk assessment along high-altitude and cold-region railways for predicting the probability of avalanches and their impact range based on data such as snow thickness, fault depth, collapse area, slope orientation, terrain slope, slope curvature, slope roughness, temperature rise rate, and ground acoustic signals.

[0016] Furthermore, the intelligent decision-making and visualization terminal includes: a graded early warning module, used to automatically trigger four levels of early warning (blue, yellow, orange, and red) based on a comprehensive risk index; an intelligent linkage control module, used to automatically activate vehicle-mounted snow melting and de-icing devices, turnout rapid de-icing devices, tunnel lining de-icing, and adjust tunnel entrance insulation air curtains or heating strips, railway line rapid snow removal, catenary rapid de-icing, and send suggested speed curves to locomotives after an early warning is triggered; and a digital twin visualization module, used to provide an integrated visual display of line terrain, equipment status, monitoring data, and early warning information.

[0017] The railway vehicle-track-network-tunnel-terrain coupled ice and snow disaster monitoring and early warning system proposed in this invention has the following beneficial effects achieved by adopting the above technical solution: 1. Comprehensive monitoring dimensions and wide coverage: a five-dimensional three-dimensional perception network is constructed to achieve all-weather, full-coverage monitoring without blind spots, and the monitoring coverage is improved by more than 80% compared with the existing technology.

[0018] 2. Strong data fusion capabilities and accurate analysis: Breaking down data silos, achieving deep integration and collaborative analysis of multi-dimensional data, improving the accuracy of comprehensive risk assessment by more than 60%.

[0019] 3. Forward-looking early warning mode and timely response: Based on the multi-physics coupling model, it realizes pre-prediction, and the early warning can be several hours to tens of hours in advance.

[0020] 4. Strong environmental adaptability: Specially designed for extreme plateau environments, it can operate stably in environments with temperatures ranging from -40℃ to 5℃ and wind speeds ≤30m / s.

[0021] 5. High level of intelligence in prevention and control: It realizes four-level hierarchical early warning and automatic linkage control, and the prevention and control response time is shortened to the minute level. Attached Figure Description

[0022] Figure 1 is a schematic diagram of the overall system of the railway vehicle-road-network-tunnel-terrain coupled ice and snow disaster monitoring and early warning system of the present invention; Figure 2 is a schematic diagram of the vehicle-based mobile monitoring unit of the present invention; Figure 3 is a schematic diagram of the roadbed fixed monitoring unit of the present invention; Figure 4 is a schematic diagram of the tunnel environment monitoring unit of the present invention; Figure 5 is a schematic diagram of the contact wire icing monitoring unit of the present invention; Figure 6 is a schematic diagram of the terrain environment monitoring unit of the present invention; Figure 7 is a schematic diagram of the overall system architecture and collaborative relationship of the present invention; Figure 8 is a schematic diagram of the decision logic of the present invention. Detailed Implementation

[0023] The technical inventions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0024] It should be noted that the terms “front,” “back,” “left,” “right,” “up,” and “down” used in the following description refer to the directions shown in the attached diagram, while the terms “inside” and “outside” refer to the directions toward or away from the geometric center of a specific component, respectively.

[0025] As shown in Figures 1-8, the technical solution adopted by this invention is as follows: The proposed railway vehicle-road-network-tunnel-terrain coupled snow and ice disaster monitoring and early warning system includes a full-domain three-dimensional perception subsystem, a multi-source data fusion communication platform, a multi-physics field coupling analysis engine, and an intelligent decision-making and visualization terminal connected in sequence. Each subsystem is connected through a multi-layer resilient communication network to realize real-time data transmission and processing.

[0026] In a specific implementation, the vehicle-based mobile monitoring unit is installed on the bogies and equipment compartments of the high-altitude EMU; the roadbed fixed monitoring network is deployed along the line, covering key areas such as tunnel entrances and sections severely affected by blowing snow; the catenary icing monitoring unit is installed on the catenary supports; and the terrain environment monitoring unit is deployed using a combination of satellite remote sensing or UAV-borne synthetic aperture radar. Multi-source data is fused and input into a coupled analysis engine, which outputs a risk index to the decision-making terminal, triggering corresponding early warning and control commands. This achieves comprehensive vehicle-based mobile monitoring of bogie monitoring equipment, equipment compartment monitoring equipment, lead car roof monitoring equipment, lead car bottom monitoring equipment, vibration sensors, acceleration sensors, and impact monitoring video equipment.

[0027] To enable those skilled in the art to better understand and implement the technical solutions of this invention, the specific implementation methods, model algorithms, and processes of the key technologies are described in complete and detailed manner below.

[0028] I. Specific Technical Scheme for Inversion of Icing Mass in Vehicle-Based Moving Monitoring Unit One of the core functions of the vehicle-based moving monitoring unit is to invert the icing mass of the bogie by analyzing its vibration signals. The specific scheme is as follows: (I) Core Principle Icing of the bogie causes changes in the mass distribution of the wheelset and frame, forming an unbalanced mass, which in turn causes changes in vibration characteristics (such as vibration frequency, amplitude, and phase). This scheme is based on the mapping relationship of "vibration response - mass imbalance". Vibration signals are collected by a high-frequency vibration accelerometer, and combined with a preset benchmark model and algorithm, the additional mass generated by icing is inverted. The core logic is: Icing mass → Unbalanced torque → Change in vibration signal characteristics → Algorithm inversion → Quantification of icing mass.

[0029] (II) Hardware Deployment and Parameter Configuration: Sensor Installation Location and Quantity: Three key measuring points are selected: the bogie wheelset axle box, the midpoint of the frame crossbeam, and the end of the frame side beam. A double-fixed system using magnetic adsorption and bolt reinforcement is employed, ensuring the sensor's contact gap with the bogie surface is ≤0.5mm, and low-temperature damping adhesive is applied.

[0030] Sensor parameters: It adopts a miniature low-temperature resistant piezoelectric accelerometer with an operating temperature range of -40℃ to 60℃, a measurement range of ±50g, a frequency response of 10Hz to 10kHz, a sampling rate of ≥2kHz, and an accuracy of ±0.5% FS.

[0031] Auxiliary hardware collaboration: Works in conjunction with an endoscopic camera and supplemental lighting system (light intensity ≥ 500 lux) to provide spatial reference for regional correlation analysis of vibration signals. The data acquisition module supports multi-channel synchronous acquisition with a transmission latency ≤ 10ms and a storage capacity ≥ 16GB.

[0032] (III) Baseline Model Construction (Calibration under Icing-Free Conditions) Establishment of Baseline Vibration Database: Under ice-free weather conditions (ambient temperature ≥ 5℃), vibration data of the train during stable operation at speeds ranging from 20km / h to 120km / h (divided according to a 20km / h gradient) were collected. The signals were filtered (5th-order Butterworth low-pass filter, cutoff frequency 5kHz) and de-trending processed to extract the baseline characteristic parameters at each speed, including: fundamental frequency vibration amplitude A0 (corresponding to wheelset rotation frequency f0 = v / (2πr)), harmonic amplitude ratio K0 (ratio of 2nd and 3rd harmonic amplitudes to fundamental frequency amplitude), and phase difference φ0 (phase relationship of fundamental frequency signals between 3 measuring points).

[0033] Unbalanced mass-vibration characteristic mapping model: Based on multibody dynamics simulation, a bogie dynamic model is constructed. Simulated ice blocks of different locations (wheel flange, frame crossbeam, brake disc) and different masses (0.1kg~5kg, in 0.1kg increments) are input to simulate vibration characteristic parameters under various working conditions. Establish a mapping relationship model: Where M is the mass of ice (kg), v is the real-time speed of the train (km / h), and A, K, and φ are real-time vibration characteristic parameters. These are the baseline characteristic parameters for the corresponding speed.

[0034] (iv) Real-time inversion algorithm process data preprocessing: adaptive filtering is performed on the real-time vibration signal to remove interference such as wind and snow impact and track irregularities.

[0035] The vibration signal and the endoscopic camera image data are synchronized by timestamp (time error ≤ 1ms).

[0036] The fundamental frequency amplitude A, harmonic amplitude ratio K, and phase difference φ of the measuring point are calculated in real time at the current velocity.

[0037] Ice quality calculation: Calculate the deviation between real-time characteristic parameters and baseline parameters: ΔA=A-A0, ΔK=K-K0, Δφ=φ-φ0.

[0038] Input ΔA, ΔK, Δφ and real-time velocity v into the preset mapping model, and output the initial icing mass M1.

[0039] Correction and compensation: Temperature compensation: Introduce a temperature correction coefficient α (α=1+0.005×(T0-T), where T0 is the reference temperature and T is the real-time temperature).

[0040] Position compensation: The actual position of the ice is identified by image recognition and corrected according to the preset position weight coefficient β (wheel flange β=1.0, frame beam β=0.8, brake disc β=0.9).

[0041] The final mass of the ice is obtained as: M = α × β × M1.

[0042] Result verification and anomaly judgment: If the icing quality fluctuation calculated over 5 consecutive sampling periods (10ms / period) is ≤5%, it is judged as a valid result.

[0043] When M≥0.5kg, it is judged as "risk of freezing" and relevant data is uploaded.

[0044] II. Specific process of data alignment, calibration and fusion in multi-source data fusion communication platform The platform's processing of multi-source heterogeneous data includes the following complete process: (I) Preparatory work: Construction of unified spatiotemporal reference framework Time reference unification: The railway-specific UTC time synchronization system is adopted, with satellite time service (GPS / BeiDou dual mode) as the core reference, and the time accuracy reaches the nanosecond level (±10ns).

[0045] The fixed monitoring unit synchronizes every 10 seconds via the 5G-R private network NTP / PTPv2 protocol (error ≤ 1ms); the mobile monitoring unit is calibrated every 5 seconds via dual-mode synchronization with satellite time synchronization via the train bus (error ≤ 2ms).

[0046] All data uses the same timestamp format: “YYYY-MM-DD HH:MM:SS.ssssss”.

[0047] Unified spatial reference: The coordinate system adopts the National Geodetic Coordinate System 2000 (CGCS2000), the plane coordinates adopt the Gauss-Kruger projection (3° zone), and the elevation adopts the 1985 National Elevation Datum.

[0048] The fixed monitoring points obtain precise coordinates (error ≤ ±5cm) through static GPS measurement, forming a coordinate database.

[0049] The mobile monitoring point uses vehicle-mounted GPS / BeiDou positioning module to fuse and correct data with odometer data (accuracy ≤ ±10cm).

[0050] The entire line is divided into "line mileage segments" in 10-meter units, and a two-way mapping of "physical coordinates - mileage index" is established.

[0051] (II) Accurate alignment of multi-source data (spatiotemporal dimension matching) Time alignment process: downsample high-frequency data (such as 2kHz vibration signal) to 10Hz, and complete low-frequency data (such as remote sensing data once a day) to 10Hz through linear interpolation.

[0052] A time-data index table is constructed using a unified time base as the axis for matching. Data with timestamp deviations exceeding ±5ms is corrected using a linear offset.

[0053] Divide the time window into 1-second intervals, and consider the data in each window containing 10 sampling points as "data from the same period".

[0054] Spatial alignment process: Mapping fixed, mobile, and remote sensing data to the corresponding "route mileage segment" according to their coordinates.

[0055] Ultimately, a two-dimensional spatiotemporal grid is formed, consisting of a "time window (1 second) + mileage segment (10 meters)," with each grid containing all multi-source data under that spatiotemporal node.

[0056] (III) Multi-source data multi-dimensional calibration (error correction and standardization) system error calibration: sensor error calibration: based on laboratory calibration data, establish an error correction model (such as the linear model y=ax+b) for each sensor to correct the original measurement value.

[0057] Environmental interference calibration: Introduce a temperature correction coefficient, use differential signal processing to eliminate electromagnetic interference, and use terrain shadow correction algorithms (such as C-correction method) to correct terrain deviations in remote sensing data.

[0058] Data transmission error calibration: Lost data is supplemented by "interpolation of data from previous and next frames + trend prediction"; for data with a delay of more than 100ms, the correct spatiotemporal node is calculated in reverse based on the delay time.

[0059] Random error removal: Outliers are detected and removed using the “3σ criterion” and the “local outlier algorithm”, followed by smoothing by moving average filtering (window size of 5 sampling points).

[0060] Data format standardization: Convert all data to floating-point (retaining 6 decimal places), unify the physical quantity units to the International System of Units (SI), and encapsulate them using the standard JSON format.

[0061] (iv) Deep fusion of multi-source data (information complementarity and value mining) Fusion layer design: Data layer fusion: For the original data of the same physical quantity from multiple sources, the "weighted average fusion method" is adopted, and the weights are dynamically determined according to the sensor accuracy and confidence level.

[0062] Feature layer fusion: Extracting features from different types of data. Numerical data is used to extract statistical features such as mean and variance; signal data is used to extract frequency / time domain features through Fourier transform and wavelet transform; image data is used to extract features such as ice area through target detection algorithms.

[0063] Decision-making level integration: Using "Bayesian network + weighted voting method", the individual disaster risk indicators are integrated into a comprehensive risk index.

[0064] Core fusion algorithm applications: For different data types, we employ adaptive weighted fusion algorithms, deep learning fusion algorithms (such as CNN-LSTM), support vector machine (SVM) algorithms that combine feature-level and decision-level approaches, and Kalman filtering that combines static and dynamic data.

[0065] Fusion result output: Output "Panoramic Dataset of Line Health Status", which includes the fusion values ​​of basic physical quantities, feature parameters, risk assessment results and data confidence of each spatiotemporal grid, and is pushed to the analysis engine with a delay of ≤100ms.

[0066] III. Core Model Algorithm Details of Multiphysics Field Coupling Analysis Engine (I) Topography-Wind Field-Snow Accumulation Coupled Model: Using digital elevation model (DEM) topographic data and real-time wind speed and direction data, it simulates 3D wind field streamlines under complex terrain, and combines snow particle motion equations to accurately calculate and predict the accumulation location and accumulation rate of windblown snow in the roadbed and turnout area.

[0067] The Euler-Euler simulation method was adopted, and the snow phase control equation was added to the air phase control equation to simulate the snow particle motion through a hybrid model.

[0068] Snow-type governing equations:

[0069] in For snow density, snow volume fraction This represents the relative slip velocity of the snow phase, simplified to 0.

[0070] Continuity equation for the hybrid model: The momentum equation for the hybrid model can be obtained by summing the momentum equations of all phases, and is expressed as:

[0071] in The drift velocity of the next phase k:

[0072] Determination of deposition location and deposition rate: The erosion or deposition of snow surface depends on the frictional velocity near the wall ( ) and critical friction speed ( The size relationship of ). When At that time, snow particles on the wall surface move into the air with the airflow, and the snow surface is eroded. At this time, snow particles on the wall surface accumulate. The formula for calculating the wall friction speed is:

[0073] in This refers to the shear stress between the airflow and the snow surface. This refers to air density.

[0074] Snow erosion and deposition flux models: Erosion flux:

[0075] Sediment flux:

[0076] In the formula: It is a constant coefficient, and its value is 7e. -4 ; This is the snow settling velocity, which can be taken as 0.5 m / s; It is the threshold friction velocity, obtained from field measurements; The mass concentration of snow is given by a value of [value]. .

[0077] The change in snow surface height per unit time is as follows: (II) The integrated icing growth model of vehicle-track-net combines data such as train running speed (wind cooling effect), ambient temperature and humidity, contact network current (Joule heating) and precipitation type to establish a dynamic heat balance equation and accurately predict the icing growth rate of contact network and undercarriage bogie.

[0078] Governing equations for the continuous phase (air and water vapor): Mass conservation equation:

[0079] in, It is a velocity vector; For quality source items ( ), which represents the rate of gas mass generation or consumption within a local volume.

[0080] Momentum equation:

[0081] in, air density; For pressure; Aerodynamic viscosity; Turbulent viscosity; It is the vector of gravitational acceleration; The momentum source term is applied to the discrete relative continuous phase.

[0082] Energy equation:

[0083] in, Specific heat capacity at constant pressure; For temperature; Thermal conductivity; The term represents the energy source, indicating the heat exchange effect of droplets releasing / absorbing latent heat.

[0084] Water vapor transport equation:

[0085] in, It represents the mass fraction of water vapor. The effective diffusion coefficient; This refers to the mass source term generated by droplet evaporation / condensation.

[0086] Lagrange tracking of discrete phases (freezing rain droplets): Equations of particle motion:

[0087]

[0088] in, The particle position; The particle velocity; For particle mass; The density of liquid water; The air resistance to the droplet; For added mass force; For other forces.

[0089] Particle thermal equilibrium equation:

[0090] in, The particle temperature; Specific heat capacity of the liquid; The convective heat transfer coefficient; The particle surface area; The mass of phase transition per unit time; This is the latent heat of phase transition.

[0091] Bidirectional coupling with the continuous phase is achieved through the source term:

[0092]

[0093]

[0094]

[0095] in, This represents the volume of a single computational cell within a CFD mesh.

[0096] Wall collision and film icing: Droplet behavior after wall collision (adhesion, splashing, breakup) Weber number ( The diameter of the droplet; The impact normal velocity; (Surface tension coefficient of droplet), Ohnesorge number (describes the coupling between viscosity and surface tension), and angle of incidence. Thus, parameters such as the breakup conditions and the distribution of droplet diameter, velocity distribution, and quantity are determined by the SSD standard and user-defined criteria.

[0097] User-defined running criteria: Allowed to run according to Trigger secondary splitting or splashing, and overlay droplet distribution functions and empirical formulas for bounce / slip angles to absorb experimental or field data.

[0098] Thin film thickness evolution equation:

[0099] in, For film thickness; The tangential divergence along the surface; The wall velocity; For the mass flux of foreign attached droplets; The mass flux consumed for freezing; This refers to the mass flux lost due to slippage.

[0100] Thin film energy equation:

[0101] in, This refers to the film temperature; Thermal conductivity of the liquid; This refers to the free-flow temperature.

[0102] Stefan condition for ice-liquid interface (one-dimensional normal approximation):

[0103] in, The density of ice; This refers to the thickness of the ice layer. The thermal conductivity of ice.

[0104] Ice layer shape update: Calculation of mesh node displacements using radial basis function interpolation.

[0105] in, The coordinates of the target node; For point The displacement vector; The number of all boundary nodes involved in the interpolation; These are the RBF interpolation coefficients; The selected RBF kernel function; These are the coordinates of the boundary nodes; Let be the Euclidean distance from the point to the boundary point; These are linear polynomial terms.

[0106] (III) The tunnel-atmosphere thermo-coupling model simulates the cold air inflow process under the "piston wind" effect based on meteorological data outside the tunnel and measured gradient data inside the tunnel, and accurately predicts the dynamic movement trajectory of the icing limit (0℃ isotherm) inside the tunnel.

[0107] Governing equations: Continuity equations:

[0108] in, For time; For volume; It is an area vector; Density; It is a velocity vector; Energy sources that contribute energy, such as radiation sources, phase-to-phase energy sources, or energy sources resulting from chemical reactions.

[0109] Momentum equation:

[0110] in, Density; It is the viscous stress tensor; It is a volume force vector.

[0111] Energy equation:

[0112] in, Total energy; Total enthalpy; This is the heat flux vector.

[0113] in For total energy, ν is the total enthalpy, and h is the static enthalpy.

[0114] Simulation and prediction process: The train motion is simplified into transient boundary conditions or momentum source terms, and the aforementioned governing equations are solved to obtain the unsteady flow field and temperature field inside the tunnel. The spatial position of the 0℃ isothermal surface is dynamically extracted through the real-time three-dimensional temperature field, and its movement trajectory is tracked.

[0115] (iv) The avalanche dynamic risk model integrates multi-source data such as snow thickness, terrain slope, temperature rise rate and ground sound signal to calculate the probability of avalanche occurrence and impact range.

[0116] Avalanche probability model: Snow layer thickness index:

[0117] —The critical snow depth at which an avalanche occurs; —Snow density; —Acceleration due to gravity; —The coefficient of friction between the snow and the mountain; —Mountain slope.

[0118]

[0119] —Snow layer thickness index; —Current snow depth; —Stable snow layer thickness threshold.

[0120] The critical snow layer thickness at which an avalanche occurs: when When the snow layer is less than 0, the snow layer condition is relatively stable; When ≈1, the snow layer is on the edge of an unstable state; When the value is greater than 1, cracks appear in the snow layer and slippage occurs.

[0121] Slope instability index: Shear stress of snow layer along the slope direction:

[0122] —Shear force of snow layer along slope; —Snow density; —Acceleration due to gravity; —Snow layer thickness; —The angle between the slope and the horizontal plane.

[0123] (2) Shear strength of the snow layer:

[0124] —The shear strength of the snow layer; —The cohesive force of the snow layer; —Normal pressure acting on the slip surface; —The pore water pressure in the snow layer; —Internal friction angle.

[0125] Dimensionless treatment of slope instability indices:

[0126] when When <0, the snow layer is in a stable state; when When ≈0, the snow layer is in a critical state of instability; when When the value is greater than 0, the snow layer is in an unstable state and the instability trend gradually intensifies.

[0127] Temperature rise rate index:

[0128] —The cumulative disturbance of snow accumulation caused by rising temperatures; —Cumulative time; — Rate of change of temperature; —The minimum rate of temperature rise that causes significant snowmelt.

[0129]

[0130] Ground acoustic signal indicators: Increased ground acoustics are often a precursor to avalanches; ground acoustics can typically reflect internal fissures and slippage within snow bodies. Ground acoustic anomaly indicators:

[0131] —Sound anomaly indicators; —The recorded ground acoustic signal energy; —The ground acoustic energy threshold at which the snow layer transitions from a stable to an unstable state.

[0132] when When the value is less than 1, the snow layer condition is relatively stable; When ≈1, the snow layer is on the edge of an unstable state; When the value is greater than 1, cracks appear in the snow layer and slippage occurs.

[0133] Snow thickness index normalized to 1 for each indicator:

[0134] Slope instability indicators:

[0135] Temperature rise rate index:

[0136] Ground acoustic signal indicators:

[0137] Probability calculation: weighted summation

[0138] , , , The weights for four indicators—snow thickness, slope instability, rate of temperature rise, and ground acoustic signal—are determined, and historical avalanche data are used to assign weights to these four indicators. , , , .

[0139] Based on the calculated probability of avalanche occurrence, avalanches are divided into three levels: low-risk avalanches when P(A) < 0.25, medium-risk avalanches when 0.25 ≤ P(A) < 0.5, and high-risk avalanches when P(A) ≥ 0.5.

[0140] Avalanche impact range calculation: Maximum slip distance:

[0141]

[0142] —Relative elevation difference in the avalanche source area; —Coefficient of friction; Impact width and thickness:

[0143]

[0144] , —Empirical coefficient; — Terrain diffusion angle.

[0145] Arrival probability model:

[0146] —The distance from the avalanche front to the target point; — Dispersion coefficient of running distance.

[0147] Railway clearance intrusion determination: Avalanche intrusion into the railway clearance is determined when both of the following conditions are met simultaneously: Maximum running distance condition: (Distance from the avalanche source area to the railway) Deposition thickness conditions: (Railway Safety Clearance Control Values) IV. System Implementation Results Based on the aforementioned specific technical solutions and algorithm models, this system has been deployed and validated on a plateau railway test section. Results show that the system operates stably and reliably in extreme environments with temperatures ranging from -40℃ to 5℃ and wind speeds ≤30m / s. The comprehensive three-dimensional sensing network achieves complete coverage of monitoring blind spots, significantly improving the monitoring range compared to traditional methods. Through deep fusion of multi-source data and precise analysis using multi-physics coupling models, the accuracy of comprehensive risk assessment is significantly improved, with early warning lead times reaching several hours to tens of hours. The response time of intelligent linkage control is shortened to the minute level, significantly enhancing the proactive, precise, and collaborative prevention and control capabilities against ice and snow disasters on plateau railways.

[0148] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

[0149] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, material, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, material, or apparatus.

Claims

1. A railway vehicle-track-network-tunnel-terrain coupled snow and ice disaster monitoring and early warning system, characterized in that: include: The all-domain three-dimensional perception subsystem is used to comprehensively collect ice and snow disaster-related data from six dimensions: train bogies and equipment compartments, lines, tracks, tunnels, overhead contact lines, and terrain along the line. A multi-source data fusion communication platform is connected to the full-domain three-dimensional perception subsystem for fusing and transmitting collected multi-source heterogeneous data; A multiphysics coupling analysis engine, connected to the multi-source data fusion communication platform, is used to perform coupling analysis on the fused data and predict the risks of various ice and snow disasters. The intelligent decision-making and visualization terminal is connected to the multiphysics coupling analysis engine and is used to generate early warning information and control commands based on the analysis results, and to display them visually.

2. The railway vehicle-track-network-tunnel-terrain coupled ice and snow disaster monitoring and early warning system according to claim 1, characterized in that: The comprehensive three-dimensional perception subsystem includes: a vehicle-based mobile monitoring unit for monitoring the snow and icing status of the train bogie and equipment compartment areas, the snow cover status of the track, and the rail adhesion status; a roadbed fixed monitoring network for monitoring the snow and icing status of turnouts, the dynamic evolution of snow-blown disasters on the track, and the working status of snow and ice melting devices; a tunnel environment monitoring unit for monitoring icing on tunnel linings, dark ice on the rail surface, micro-meteorological gradients, and ground temperature; a catenary icing monitoring unit for monitoring the icing load on the catenary conductors and the operating status of insulators; and a terrain environment monitoring unit for acquiring data on snow thickness and stability along the track and avalanche precursor signals.

3. The railway vehicle-track-network-tunnel-terrain coupled ice and snow disaster monitoring and early warning system according to claim 2, characterized in that: The vehicle-based mobile monitoring unit includes: a miniature low-temperature resistant industrial camera, a supplementary lighting system, and a high-frequency vibration accelerometer installed on key parts of the bogie, used to monitor the thickness of snow and ice accumulation in the bogie area and invert the icing quality of the bogie area through vibration spectrum signals; a temperature, humidity, and pressure differential composite sensor and an infrared thermal imaging probe deployed in the equipment compartment, used to determine filter blockage and monitor the temperature of key components; a train bus data reading module, used to extract real-time train wheel slip / slip signals and real-time changes in traction / braking force data, and invert the law of changes in the adhesion coefficient of the rail surface caused by snow and ice accumulation; and a long-focal-length optical zoom industrial camera deployed on the top of the train's front car, used to clearly capture the snow and ice adhesion morphology of the contact wires and catenary cables from a distance, and synchronously transmit the relevant collected information to the contact wire icing monitoring unit, supplementing... To aid in assessing the icing status of the overhead contact line, this long-focal-length optical zoom industrial camera is used to monitor the spatial distribution and size characteristics of ice buildup on the tunnel lining when the train passes through the tunnel. The collected information is simultaneously transmitted to the tunnel environmental monitoring unit to assist in assessing the icing status of the tunnel lining. Low-temperature resistant wide-angle industrial cameras and lidar deployed at the bottom of the train's locomotive are used to clearly capture ice and snow cover on the track surface, snow accumulation and icing on the switches, and the snow cover pattern in the track area. The collected information further assists in monitoring changes in the track adhesion coefficient, the snow cover status of the track, and the snow accumulation and icing status of the switches. In addition, when the train passes through the tunnel, the wide-angle industrial camera and lidar are also used to simultaneously collect information on the distribution of dark ice on the track surface within the tunnel, and the collected information is simultaneously transmitted to the tunnel environmental monitoring unit to assist in assessing the tunnel's ice and snow disaster status.

4. The railway vehicle-track-network-tunnel-terrain coupled ice and snow disaster monitoring and early warning system according to claim 2, characterized in that: The roadbed fixed monitoring network includes: fiber optic temperature / stress sensors deployed in the gaps between turnout sleepers to monitor freezing stress; millimeter-wave radar to scan the gap between switch rails and stock rails to identify snow accumulation and ice debris; a turnout snow melting device operation status feedback module; and ultrasonic snow depth arrays, small anemometers, and low-temperature resistant wide-angle industrial cameras deployed along the windward side of the line to monitor the snow cover status within the line and the evolution of windblown snow from outside the line to inside the line.

5. The railway vehicle-track-network-tunnel-terrain coupled ice and snow disaster monitoring and early warning system according to claim 1, characterized in that: The tunnel environmental monitoring unit includes: dual-spectrum pan-tilt cameras deployed at the tunnel entrance and key seepage points inside the tunnel; temperature and humidity sensors and ground temperature sensors in key areas are deployed at intervals along the tunnel's depth to construct temperature and humidity gradient maps inside and outside the tunnel.

6. The railway vehicle-track-network-tunnel-terrain coupled ice and snow disaster monitoring and early warning system according to claim 1, characterized in that: The overhead contact line icing monitoring unit includes a miniature weather station, a dual-spectrum camera (visible light + thermal imaging), an overhead contact line tension sensor, and an insulator tilt sensor, used to monitor the type of icing (rime, hoarfrost, mixed rime) and icing load of the overhead contact line in key sections.

7. The railway vehicle-track-network-tunnel-terrain coupled ice and snow disaster monitoring and early warning system according to claim 1, characterized in that: The terrain environment monitoring unit includes: satellite remote sensing or UAV-borne synthetic aperture radar for periodically scanning the slopes along the route; and snow pressure cells and ground acoustic sensors deployed on the slopes of avalanche-prone areas.

8. The railway vehicle-track-network-tunnel-terrain coupled ice and snow disaster monitoring and early warning system according to claim 1, characterized in that: The multi-source data fusion communication platform includes: a multi-source heterogeneous data fusion engine, used to align, calibrate and deeply fuse monitoring data from different sources, formats and spatiotemporal resolutions under a unified spatiotemporal reference framework to generate a panoramic view of the line health status; and a multi-layer resilient communication network, which comprehensively adopts railway-specific 5G-R, satellite communication and wireless Mesh network to achieve reliable and low-latency transmission of key data in complex plateau environments.

9. The railway vehicle-track-network-tunnel-terrain coupled ice and snow disaster monitoring and early warning system according to claim 1, characterized in that: The multiphysics coupling analysis engine includes a coupled numerical model of multiple transport characteristics of wind and snow migration-erosion-deposition in railway subgrade areas, which is used to predict the accumulation location and rate of windblown snow in the subgrade and turnout areas based on digital elevation model data, real-time wind speed and direction data and snow cover morphology in key areas. Numerical models for multiphase coupled transport of wind, snow, water, and ice in train bogies / equipment compartments are developed, including: a numerical model for predicting the quality and spatial distribution of snow and ice accumulation in the bogie and equipment compartment areas based on data such as train speed, refined geometric models of train bogies and equipment compartments, ambient temperature and humidity, snow thickness, and physical properties of snow cover; a quasi-steady-state numerical model for the evolution of icing morphology in railway catenary / rail surface based on data such as train speed, departure density, ambient temperature and humidity, catenary current, and freezing rain / snowfall type; a cold-region tunnel-atmosphere thermodynamic coupling model for predicting the dynamic movement of icing boundaries within tunnels based on data such as tunnel internal temperature, spatial temperature linear distribution, external temperature, and seepage location and flow parameters; and a numerical model for avalanche dynamic risk assessment along high-altitude and cold-region railways for predicting the probability and impact range of avalanches based on data such as snow thickness, fault depth, collapse area, slope orientation, terrain slope, slope curvature, slope roughness, temperature rise rate, and ground acoustic signals.

10. The railway vehicle-track-network-tunnel-terrain coupled ice and snow disaster monitoring and early warning system according to claim 1, characterized in that: The intelligent decision-making and visualization terminal includes: a graded early warning module, used to automatically trigger four levels of early warning (blue, yellow, orange, and red) based on a comprehensive risk index; an intelligent linkage control module, used to automatically activate vehicle-mounted snow melting and de-icing devices, turnout rapid de-icing devices, tunnel lining de-icing, and adjust tunnel entrance insulation air curtains or heating strips, railway line rapid snow removal, catenary rapid de-icing, and send suggested speed curves to locomotives after an early warning is triggered; and a digital twin visualization module, used to provide an integrated visual display of line terrain, equipment status, monitoring data, and early warning information.

Images