A long time sequence life prediction method for railway signal relay

Abstract

This invention relates to the field of electrical testing technology, specifically a long-term life prediction method for railway signal relays. The method involves acquiring the geographical coordinates, terrain features, and environmental anchor point data of signal relays along the railway line; during relay operation, current waveform acquisition is triggered by coil level jumps to obtain electrical operation parameters; the channel identifier and local geometric scale of the signal relay are assigned to each measurement node in the electrical measurement association topology; based on terrain features, conduction attenuation characteristics are calibrated, and orthogonal conduction attenuation relationships are established to obtain conduction attenuation calibration parameters; using environmental anchor point data as physical reference constraints, multi-point differential compensation detection is performed on the electrical operation parameters to identify multi-physics field distribution detection values ​​and obtain electrical degradation parameters; based on the electrical degradation parameters and the common-mode electrical attenuation parameters of neighboring measurement nodes in the electrical measurement association topology, association detection is performed, and the life prediction result of the signal relay is output.

Description

Technical Field

[0001] This invention relates to the field of electrical testing technology, specifically a long-term life prediction method for railway signal relays. Background Technology

[0002] Railway signal relays are core safety components in signaling systems, performing logic control and circuit switching. The long-term degradation of their contact resistance and coil insulation directly impacts railway traffic safety. Long-term online monitoring of critical electrical quantities in each signal relay along the railway line and prediction of their remaining service life are crucial for ensuring the reliable operation of the signaling system. However, railway lines traverse complex terrain areas such as bridges, tunnels, and cuttings, subjecting signal relays to the combined effects of periodic mechanical vibration loads caused by passing trains and thermal interference from tunnel microclimates. The combined influence of these terrain loads and climatic conditions causes reversible dynamic shifts in electrical quantities such as contact resistance and coil current under normal operating conditions. These shifts are non-permanent electrical responses driven by terrain and climate, and do not correspond to irreversible physical damage caused by mechanical wear of the relay contacts or aging of the coil insulation.

[0003] Existing online detection methods for electrical parameters of signal relays generally treat each signal relay along the line as an independent single-point detection object, only collecting and extracting the operating current waveform of a single relay, without establishing a correlation detection mechanism based on the physical conduction relationship of the terrain structure between adjacent detection locations. Under this single-point detection framework, the reversible common-mode electrical drift caused by external terrain loads and climate interference, and the irreversible electrical degradation caused by relay contact wear and coil insulation aging, are superimposed and mixed in the detection signal, making it impossible to identify and separate them at the electrical quantity detection level.

[0004] Therefore, existing methods for detecting electrical parameters of signal relays cannot accurately detect and separate the net electrical degradation that truly reflects the wear of internal contacts and aging of coil insulation under actual service conditions of complex terrain loads and climatic interference. This results in the long-term life prediction results being systematically biased and inaccurate due to environmental interference.

[0005] To address this, a long-term life prediction method for railway signal relays is proposed. Summary of the Invention

[0006] The purpose of this invention is to provide a long-term lifetime prediction method for railway signal relays. Based on electrical degradation parameters and common-mode electrical attenuation parameters of neighboring measurement nodes in the electrical measurement association topology, the method performs association detection and outputs the lifetime prediction result of the signal relay.

[0007] To achieve the above objectives, the present invention provides the following technical solution: A method for predicting the long-term lifespan of railway signal relays includes: Obtain the geographical coordinates, terrain features, and environmental anchor point data of signal relays along the railway line; When the signal relay operates, the current waveform of the entire process is collected triggered by the jump in coil level, and the transient motion stage and steady-state contact stage are divided based on the current change rate threshold to obtain the electrical action parameters. The channel identifier and local geometric scale of the signal relay are assigned to each measurement node in the electrical measurement association topology; based on the terrain features, the conduction attenuation characteristics of mechanical vibration medium and thermal diffusion fluid are calibrated respectively, and an orthogonal conduction attenuation relationship is established to obtain the conduction attenuation calibration parameters; Using environmental anchor point data as physical reference constraints, the electrical action parameters are detected by multi-point differential compensation using electrical measurement correlation topology and conduction attenuation calibration parameters. The multi-physics distribution detection values ​​of each measurement node are obtained, and the common-mode environmental electrical drift affected by terrain load and climate is removed from the electrical action parameters to obtain electrical degradation parameters. Based on electrical degradation parameters and common-mode electrical attenuation parameters of neighboring measurement nodes in the electrical measurement correlation topology, multi-node electrical quantity correlation detection is performed, and the life prediction results of each signal relay are output.

[0008] Preferably, the geographical coordinates, terrain features, and environmental anchor point data of relays along the railway line are obtained, specifically including: The longitude coordinates, latitude coordinates, and rail surface elevation of each signal relay are obtained to obtain the spatial position parameters of the signal relay, which are used as geographical coordinates. The load-bearing structures along the railway line are measured to distinguish between bridge sections, tunnel sections, cutting sections and roadbed sections. For bridge sections, the bridge span and pier spacing are measured and calculated. For tunnel sections, the clearance height and cross-sectional area are measured to obtain the geometric scale parameters of each section as topographic features. Multi-parameter intelligent sensing nodes are deployed at signal relays along the line. These nodes integrate vibration acceleration detection units, temperature detection units, and humidity detection units. They continuously collect vibration acceleration, temperature, and humidity data at the measurement points, and after local preprocessing, upload the data through a communication interface to obtain environmental benchmark detection values, which serve as environmental anchor point data.

[0009] Preferably, the process of obtaining electrical operating parameters specifically includes: Intelligent electrical parameter sensing units are configured at each signal relay. The intelligent electrical parameter sensing unit uses the level change of the signal relay drive control signal as a synchronous trigger signal. At the level change moment, it triggers the acquisition of the current waveform of the coil current and records the current waveform of the entire process from the trigger moment until the coil current returns to stability, thereby obtaining the action current waveform detection signal. The current change rate is detected point by point for the action current waveform detection signal. The action current waveform detection signal is divided into transient motion detection segment and steady-state contact detection segment, with the time when the current change rate is lower than the current change rate threshold and the stability condition is met as the dividing point. In the transient motion detection section, the armature engagement time, contact bounce count, bounce duration, and current dip depth are measured to obtain transient motion electrical parameters. In the steady-state contact detection section, the contact resistance value, average coil operating current, and operating current ripple coefficient are measured to obtain steady-state contact electrical parameters. The transient motion electrical parameters and steady-state contact electrical parameters are combined to obtain the electrical operating parameters of the signal relay.

[0010] Preferably, the process of constructing the electrical measurement association topology includes: Using signal relays along the line as measurement nodes, a channel identifier is assigned to each measurement node. The channel identifier includes the station code, cabinet number, and slot identifier, and the signal relay corresponding to the measurement node is uniquely calibrated. The topographic features of the terrain area where each measurement node is located are used as the local geometric scale and assigned to the corresponding measurement node. Based on the geographical coordinates of each measurement node, measurement association channels are established for measurement node pairs whose spatial distance is less than the preset neighborhood radius. The conduction strength calibration value of the measurement association channel is jointly calibrated by the spatial distance of the measurement node pairs and the type of terrain bearing structure traversed. The channel identifier, local geometric scale of each measurement node and the conduction strength calibration value of each measurement association channel are combined to form an electrical measurement association topology covering the entire line.

[0011] Preferably, the process of obtaining the conduction attenuation calibration parameters specifically includes: For mechanical vibration propagation with solid structure as medium, the vibration energy attenuation characteristics of the solid propagation path between adjacent measurement node pairs are physically calibrated based on the local geometric scale of each measurement node in the electrical measurement association topology, so as to obtain the vibration transmission attenuation calibration value of each measurement node pair. For thermal diffusion propagation using fluid air as the medium, the attenuation characteristics of heat transfer between adjacent measurement node pairs are physically calibrated based on the cross-sectional area and connection length of the enclosed space where each measurement node is located, and the thermal transfer attenuation calibration value of each measurement node pair is obtained. The vibration transmission attenuation calibration value and the thermal transmission attenuation calibration value correspond to independent physical propagation processes, forming an orthogonal conduction attenuation relationship; the signal relay measurement node pair is used as the calibration node pair, and the vibration transmission attenuation calibration value and the thermal transmission attenuation calibration value are used as corresponding components to form conduction attenuation calibration parameters.

[0012] Preferably, the process of obtaining electrical degradation parameters includes: By utilizing the conduction strength calibration value and attenuation calibration value of the conduction attenuation calibration parameter of each measurement association channel in the electrical measurement association topology, multi-point weighted differential compensation detection is performed on the influence of the superposition of physical quantities of the neighboring environment on each measurement node, so as to obtain the multi-physics field distribution detection value of each measurement node at the current time. Based on the multi-physics field distribution detection values ​​and the pre-calibrated environmental physical quantity and electrical quantity response calibration relationship, the common-mode environmental electrical drift caused by terrain load and climate at each measurement node is detected; the electrical action parameters of each measurement node at the current moment are used as the detection values ​​to be compensated, and the common-mode environmental electrical drift is used as the interference compensation amount. Differential interference suppression is performed on the detection values ​​to be compensated to remove the common-mode environmental electrical interference from the electrical action parameters, and further electrical degradation parameters reflecting the accumulation of irreversible damage inside the relay are obtained.

[0013] Preferably, the process of obtaining the remaining life prediction results of the signal relay includes: The electrical degradation parameters of the signal relay are detected to obtain a sequence of individual electrical degradation values ​​for the signal relay. The historical common-mode environmental electrical drift of the neighboring measurement nodes connected to the target signal relay via the measurement association channel is extracted from the electrical measurement association topology. The conduction strength calibration value of the measurement association channel of each neighboring measurement node is used as a coefficient to weight and merge the historical common-mode environmental electrical drift to obtain the common-mode electrical attenuation parameter of the neighboring measurement nodes. The change rate of the common-mode electrical attenuation parameter is detected. When the change rate of the common-mode electrical attenuation parameter continuously exceeds a preset rate detection threshold, it is determined that there is environmental degradation in the area where the target signal relay is located, and multi-node electrical quantity correlation detection is initiated. In the multi-node electrical quantity correlation detection, the change rate of the neighboring common-mode electrical attenuation parameter is used as an environmental acceleration detection factor, which is combined with the initial lifetime prediction detection value obtained from the sequence of individual electrical degradation detection values. The initial lifetime prediction detection value is then reduced and corrected to obtain a lifetime prediction result corrected by multi-node joint electrical quantity detection. Based on the lifetime prediction result, graded early warning information for the signal relay is output.

[0014] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention precisely divides the transient motion detection segment and the steady-state contact detection segment using dual criteria of current change rate and stable duration. This allows the transient motion electrical parameters, such as the number of contact bounces and armature engagement time, obtained from the two segments respectively, and the steady-state contact electrical parameters, such as the contact resistance value and the average coil operating current value, to each correspond to a clear and independent physical failure mechanism at the detection level. The transient motion electrical parameters reflect the mechanical wear state of the contact, while the steady-state contact electrical parameters reflect the contact plating erosion and coil insulation aging state. The underlying electrical quantity acquisition ensures the physical fidelity of the electrical action parameters, providing high-quality detection inputs with clear physical meaning for subsequent common-mode interference elimination and life prediction.

[0015] 2. This invention constructs an electrical measurement correlation topology covering the entire line, unifying the terrain geometry attributes and spatial conduction correlation of each signal relay into a measurement reference system. Simultaneously, it performs physical calibration of conduction attenuation characteristics for two independent physical processes: mechanical vibration propagation in solid media and thermal diffusion propagation in fluid air. This forms orthogonal conduction attenuation calibration parameters, accurately quantifying the physical conduction intensity of environmental stress between each pair of measurement nodes along the entire line. Using the measured physical quantities of each anchor point sensor as physical boundary constraints, multi-point weighted differential compensation detection is implemented for each measurement node using the electrical measurement correlation topology and conduction attenuation calibration parameters. This accurately detects the common-mode environmental electrical drift of each measurement node at the current moment due to terrain loads and climate influences, improving the anti-interference capability and long-term stability of railway signal relay electrical degradation parameter detection.

[0016] 3. While continuously detecting the electrical degradation of each individual relay, this invention extracts the historical common-mode environmental electrical drift of neighboring measurement nodes and weights and merges them using the conduction strength calibration value as a coefficient to obtain the neighboring common-mode electrical attenuation parameter. When the rate of change of this parameter continuously exceeds the preset rate detection threshold, it is determined that there is environmentally driven degradation in the target area. The initial life prediction detection value of the single unit is reduced and corrected by a normalized dimensionless environmental acceleration detection factor, and the life prediction result corrected by spatial group perception is output, which improves the timeliness of perception of hidden group environmentally driven degradation. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating a long-term lifetime prediction method for railway signal relays according to the present invention. Figure 2 This is a logic diagram of a long-sequence lifetime prediction method for railway signal relays according to the present invention; Figure 3 This is a flowchart illustrating the electrical operating parameters of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1:

[0020] This invention proposes a long-term timing lifetime prediction method for railway signal relays. The process of the method is as follows: Figure 1 As shown, the logic of the method is as follows: Figure 2 As shown, it includes: Obtain the geographical coordinates, terrain features, and environmental anchor point data of signal relays along the railway line; When the signal relay operates, the current waveform of the entire process is collected triggered by the jump in coil level, and the transient motion stage and steady-state contact stage are divided based on the current change rate threshold to obtain the electrical action parameters. The channel identifier and local geometric scale of the signal relay are assigned to each measurement node in the electrical measurement association topology; based on the terrain features, the conduction attenuation characteristics of mechanical vibration medium and thermal diffusion fluid are calibrated respectively, and an orthogonal conduction attenuation relationship is established to obtain the conduction attenuation calibration parameters; Using environmental anchor point data as physical reference constraints, the electrical action parameters are detected by multi-point differential compensation using electrical measurement correlation topology and conduction attenuation calibration parameters. The multi-physics distribution detection values ​​of each measurement node are obtained, and the common-mode environmental electrical drift affected by terrain load and climate is removed from the electrical action parameters to obtain electrical degradation parameters. Based on electrical degradation parameters and common-mode electrical attenuation parameters of neighboring measurement nodes in the electrical measurement correlation topology, multi-node electrical quantity correlation detection is performed, and the life prediction results of each signal relay are output.

[0021] Preferably, the geographical coordinates, terrain features, and environmental anchor point data of relays along the railway line are obtained, specifically including: The longitude coordinates, latitude coordinates, and rail surface elevation of each signal relay are obtained to obtain the spatial position parameters of the signal relay, which are used as geographical coordinates. The load-bearing structures along the railway line are measured to distinguish between bridge sections, tunnel sections, cutting sections and roadbed sections. For bridge sections, the bridge span and pier spacing are measured and calculated. For tunnel sections, the clearance height and cross-sectional area are measured to obtain the geometric scale parameters of each section as topographic features. Multi-parameter intelligent sensing nodes are deployed at signal relays along the line. These nodes integrate vibration acceleration detection units, temperature detection units, and humidity detection units. They continuously collect vibration acceleration, temperature, and humidity data at the measurement points, and after local preprocessing, upload the data through a communication interface to obtain environmental benchmark detection values, which serve as environmental anchor point data.

[0022] The geographical coordinates of signal relays are obtained through on-site surveying and measurement. During the railway line survey phase, three-dimensional coordinate measurements are performed on the location of each signal relay along the line, recording its longitude, latitude, and the elevation value of the rail surface, thus forming the spatial position parameters of the relay. The inclusion of elevation data helps to reflect the impact of vertical elevation differences on the environmental stress transmission path in subsequent spatial measurement and topology association, improving the accuracy of transmission direction modeling.

[0023] Topographic features were determined by segmenting and classifying the load-bearing structures along the railway line. The types of surface load-bearing structures were identified segment by segment along the line, sequentially labeling bridge sections, tunnel sections, cutting sections, and roadbed sections. For bridge sections, the calculated span of each span and the distance between adjacent piers needed to be measured on-site. For tunnel sections, the clearance height and the cross-sectional area of ​​the inner contour of the lining needed to be measured; these two measurements together determine the effective channel size for heat diffusion within the tunnel and are directly input as spatial references during subsequent conduction attenuation calibration.

[0024] Environmental anchor point data acquisition employs a sensor deployment strategy based on representative nodes in different zones. The entire route is divided into several micro-topographic units according to terrain micro-environment types. Within each micro-topographic unit, at least two representative signal relay installation locations are selected, and multi-parameter intelligent sensor nodes are deployed. After completion, the range of each sensor is calibrated to confirm that its accuracy within the effective range meets the requirements. After calibration, each multi-parameter intelligent sensor node continuously collects vibration acceleration, ambient temperature, and relative humidity at the measurement points at a preset sampling frequency, forming continuously updated environmental benchmark detection values ​​over time, which serve as the physical benchmark constraint for subsequent multi-point differential compensation detection. The environmental state of signal relay measurement nodes without installed sensors is inferred through the established neighborhood conduction relationships in the electrical measurement association topology.

[0025] This invention provides a precise geographic and physical reference foundation for subsequent construction of spatial measurement correlation topology and implementation of multi-physics field distribution detection by collecting spatiotemporal reference information of signal relays along railway lines. The sensor deployment strategy of representing nodes by region effectively controls the deployment scale while ensuring the integrity of spatial coverage of environmental anchor points, balancing engineering feasibility and monitoring accuracy, and providing a stable reference input for long-term continuous monitoring.

[0026] Preferably, the process for obtaining electrical action parameters is as follows: Figure 3 As shown, it specifically includes: Intelligent electrical parameter sensing units are configured at each signal relay. The intelligent electrical parameter sensing unit uses the level change of the signal relay drive control signal as a synchronous trigger signal. At the level change moment, it triggers the acquisition of the current waveform of the coil current and records the current waveform of the entire process from the trigger moment until the coil current returns to stability, thereby obtaining the action current waveform detection signal. The current change rate is detected point by point for the action current waveform detection signal. The action current waveform detection signal is divided into transient motion detection segment and steady-state contact detection segment, with the time when the current change rate is lower than the current change rate threshold and the stability condition is met as the dividing point. In the transient motion detection section, the armature engagement time, contact bounce count, bounce duration, and current dip depth are measured to obtain transient motion electrical parameters. In the steady-state contact detection section, the contact resistance value, average coil operating current, and operating current ripple coefficient are measured to obtain steady-state contact electrical parameters. The transient motion electrical parameters and steady-state contact electrical parameters are combined to obtain the electrical operating parameters of the signal relay.

[0027] In railway signaling systems, signal relays are uniformly scheduled by a microcomputer control system. When the microcomputer system drives the relay to operate, it applies a drive control signal with a clear level transition characteristic to the relay coil circuit. This drive control signal is synchronously introduced into the trigger input terminal of the intelligent electrical parameter sensing unit. Once the acquisition unit detects a level transition, it immediately starts high-frequency continuous sampling of the coil current to ensure that transient characteristics such as contact bounce are completely captured.

[0028] The data acquisition continues until the coil current stabilizes, fully covering the entire process from the start of relay operation to the completion of the electrical process. The recorded time-domain current value sequence is the operating current waveform detection signal. Compared with fixed-period timing sampling, trigger-based acquisition uses level transitions as a precise time reference, eliminating the risk of timing misalignment between the acquisition window and the operation process, and ensuring that each frame of acquired data contains complete information from the start to the end of the operation.

[0029] After obtaining the operating current waveform detection signal, the system divides the current difference between adjacent sampling points by the sampling time interval to calculate the current change rate sequence point by point. The pre-calibrated current change rate threshold is determined during the factory calibration stage based on the coil circuit parameters of the signal relay model. Its value is slightly higher than the residual change rate of the coil current caused by power supply ripple in steady state, ensuring that the true steady-state stage meets the condition of being below the threshold. The specific criterion for judging the stability condition is: the continuous duration of the current change rate being below the threshold exceeds the preset stability judgment time. This stability judgment time is determined based on the coil circuit time constant and is no less than three times the coil time constant to cover the exponential convergence phase of the coil current. After the initial closure, the contacts typically experience multiple mechanical bounces. Each bounce causes the contacts to briefly open, resulting in a change in the magnetic reluctance of the magnetic circuit and a brief increase in the rate of change of current exceeding the threshold. When the system detects this situation, it resets the stabilization timer and uses the earliest moment when the dual conditions are met again after the bounce ends as the final dividing point. This ensures that the dividing point is located after all contact bounce processes have ended. Based on this dividing point, the action current waveform detection signal is divided into a transient motion detection segment and a steady-state contact detection segment.

[0030] The electrical parameters reflecting the mechanical state of the armature during the transient motion detection section are detected as follows: Armature engagement time is defined as the time span from the trigger moment to the boundary point, reflecting the total time required for the armature to move from the released state to full engagement. Its variation is directly related to mechanical degradation factors such as decreased spring stiffness, increased air gap in the iron core, and mechanical jamming. The number of contact bounces is determined by statistically analyzing the number of independent events where the current change rate repeatedly exceeds a threshold during the transient motion detection section. Each threshold-exceeding event corresponds to a complete bounce where the contact opens and closes again. The bounce duration is the cumulative value of the durations of each threshold-exceeding event. As the wear of the micro-protrusions on the contact surface deepens, the number of bounces and the duration exhibit a regular evolution, making it a sensitive indicator for assessing the mechanical wear state of the contact. The current sag depth is the difference between the minimum current waveform during transient motion and the expected undisturbed current value based on the coil circuit parameters. This sag is caused by the dynamic change in the air gap during armature movement, leading to a change in the effective inductance of the coil. The sag depth is closely related to the health state of the iron core magnetic circuit.

[0031] The electrical parameter detection method within the steady-state contact detection section is as follows: The detection of contact resistance requires a dedicated contact circuit measuring device. After the steady-state contact detection section begins, the measuring device injects a low-frequency detection excitation signal with stable amplitude into the contact circuit. Simultaneously, the contact voltage drop detection value and the contact circuit current detection value are collected by the contact voltage acquisition circuit and the contact circuit current acquisition circuit, respectively. The contact resistance value for this operation is obtained by dividing the contact voltage drop detection value by the contact circuit current detection value. Electrical isolation is used between the measuring device and the main signal circuit, ensuring that the detection excitation does not affect the normal operation of the main signal circuit. The average coil operating current is taken as the arithmetic mean of the coil current sampling values ​​within the steady-state contact detection section. This average value is affected by both the coil winding resistance and insulation resistance. When coil insulation aging leads to increased leakage current, the average operating current shows a regular upward trend, making it a direct measure of insulation aging and degradation. The operating current ripple coefficient is taken as the ratio of the root mean square error to the mean value of the current sampling value sequence within the steady-state contact detection section. The ripple component originates from changes in the coil's ferromagnetic saturation characteristics and AC interference in the contact circuit. A monotonically increasing coefficient indicates the degradation of the coil's ferromagnetic properties. After each action detection, the transient motion electrical parameters and steady-state contact electrical parameters are combined into complete electrical action parameters, which are then appended sequentially to the long-sequence electrical quantity detection record of the relay for use in subsequent steps.

[0032] This invention employs a full-process acquisition mechanism based on the transition of the coil control signal level to the trigger source, achieving high-precision synchronous waveform recording for each relay operation and eliminating timing misalignment issues in timed sampling. The dual segmentation criteria of current change rate and stable duration accurately distinguish between the two physical stages of transient mechanical motion and steady-state electrical contact. This ensures that the characterization parameters of the two failure mechanisms—contact mechanical wear and coil electrical aging—are not overlapping, each possessing a clear physical failure correspondence. This significantly improves the physical fidelity of the underlying detection data, providing a reliable foundation for subsequent high-precision lifetime prediction.

[0033] Preferably, the process of constructing the electrical measurement association topology includes: Using signal relays along the line as measurement nodes, a channel identifier is assigned to each measurement node. The channel identifier includes the station code, cabinet number, and slot identifier, and the signal relay corresponding to the measurement node is uniquely calibrated. The topographic features of the terrain area where each measurement node is located are used as the local geometric scale and assigned to the corresponding measurement node. Based on the geographical coordinates of each measurement node, measurement association channels are established for measurement node pairs whose spatial distance is less than the preset neighborhood radius. The conduction strength calibration value of the measurement association channel is jointly calibrated by the spatial distance of the measurement node pairs and the type of terrain bearing structure traversed. The channel identifier, local geometric scale of each measurement node and the conduction strength calibration value of each measurement association channel are combined to form an electrical measurement association topology covering the entire line.

[0034] Electrical measurement association topology is a product of abstracting the spatial distribution and physical conduction relationships of signal relays along a railway line into a mathematical structure. Each signal relay along the line is considered a measurement node, carrying two key attributes: channel identification and local geometric scale. The channel identification is a unique identifier for the signal relay across the entire line, composed of three levels of coding: the first level is the station code of the station to which the relay belongs, typically a combination of the station name abbreviation and the line code; the second level is the cabinet number of the cabinet where the relay is located, corresponding to the physical cabinet position within the interlocking room; and the third level is the slot identifier of the relay within the cabinet, precisely locating its specific installation position. These three levels of coding together constitute a complete channel identification that unambiguously and uniquely identifies the relay in the railway signaling system, ensuring the accuracy of subsequent data tracing. The local geometric scale is an attribute that binds the geometric parameters of the terrain microenvironment where the measurement node is located to the node's attributes. The specific content comes from terrain features: for measurement nodes in bridge sections, the local geometric scale includes the calculated span and bridge deck width of the bridge; for measurement nodes in tunnel sections, it includes the clearance height and cross-sectional area of ​​the tunnel. Local geometric scales enable each measurement node to carry spatial morphological information of its physical microenvironment, providing geometric input for subsequent conduction attenuation characteristic calibration.

[0035] The establishment of measurement association channels is based on the geographical coordinates of each measurement node. The three-dimensional spatial distance between each pair of measurement nodes along the entire line is calculated. When the spatial distance between two nodes is less than a preset neighborhood radius, a measurement association channel is established between them, indicating that there is a quantifiable physical environment transmission path between the node pairs. The preset neighborhood radius is determined based on the typical effective vibration propagation distance of the railway structure, typically aiming to establish channels between each measurement node and at least three neighboring nodes on average. The transmission intensity calibration value of the measurement association channel quantifies the strength of the physical transmission of environmental stress between the nodes at both ends of the channel. Its calibration is determined based on two independent factors: the smaller the spatial distance, the smaller the attenuation of environmental stress from one end to the other, and the larger the transmission intensity calibration value; the type of terrain bearing structure determines the connectivity of physical transmission. Between two nodes in the same structural unit (such as within the same bridge span or the same tunnel section), the transmission path is continuous and complete, and the transmission intensity calibration value is higher; when the transmission path crosses expansion joints, abrupt boundary changes in structural type, or discontinuous terrain sections, the transmission connectivity decreases, and the transmission intensity calibration value is correspondingly lower. The conduction intensity calibration value is normalized and ranges from greater than zero to no greater than 1. The larger the value, the stronger the physical conduction correlation.

[0036] By integrating the channel identifiers, local geometric scale attributes, and conduction strength calibration values ​​of all measurement nodes, an electrical measurement association topology covering the entire line is formed. Physically, this topology represents an environmental stress transmission network established between the locations of all signal relays along the entire line, based on the actual terrain structure and three-dimensional spatial relationships, through solid elastic conduction and fluid thermal diffusion.

[0037] The electrical measurement correlation topology established in this invention breaks away from the previous model of analyzing each signal relay as an isolated individual, and incorporates all relays along the line into a group monitoring architecture with spatial correlation attributes. Channel identification ensures accurate identification and data traceability of all equipment along the line, local geometric scale embeds the physical characteristics of the terrain microenvironment into the measurement nodes, and conduction strength calibration values ​​quantify the true physical correlation strength between equipment. This provides an indispensable spatial topological foundation for conduction attenuation characteristic calibration and multi-point differential compensation detection, and is a key prerequisite for realizing group-based environmentally driven degradation perception.

[0038] Specifically, the conduction intensity calibration value is determined by multiplying the distance attenuation coefficient and the structural connectivity coefficient: the distance attenuation coefficient is calculated based on the ratio of the three-dimensional spatial distance between the two measurement nodes to the reference propagation distance corresponding to the terrain type, using a preset attenuation index; the larger the ratio, the smaller the distance attenuation coefficient. The structural connectivity coefficient is determined in three levels based on the connectivity status of the terrain bearing structure along the measurement-related channel path: the first level is taken when the path is uninterrupted within the same structural unit, the second level is taken when the path passes through an expansion joint, and the third level is taken when the path passes through abrupt boundary of structural type or discontinuous section of terrain; the first level is greater than the second level, which is greater than the third level. The distance attenuation coefficient is multiplied by the structural connectivity coefficient and normalized to a range greater than 0 and not greater than 1 to obtain the conduction intensity calibration value of the measurement-related channel.

[0039] The distance attenuation coefficient is calibrated using the benchmark propagation distance corresponding to the terrain type as a reference. The benchmark propagation distance represents the reference scale for the effective transmission of vibration or thermal stress under that terrain type. Experimental data validated that for bridge sections, twice the average span of adjacent piers is used as the benchmark propagation distance for vibration; for tunnel sections, 50 times the square root of the tunnel cross-sectional area is used as the benchmark propagation distance for thermal stress; and for roadbed sections, the smaller of the two values ​​is used as the benchmark. The normalized distance is obtained by dividing the three-dimensional spatial distance between two measurement nodes by the benchmark propagation distance. The distance attenuation coefficient is equal to the reciprocal of the normalized distance and truncated to no greater than 1.

[0040] The structural connectivity coefficient is determined based on the path connectivity status. The first coefficient is used when there are no structural interruptions within the same bridge span or tunnel section, indicating a continuous and complete physical transmission path with minimal environmental stress transmission loss. The second coefficient is used when the path passes through bridge expansion joints, where elastic wave transmission loss is determined on-site through vibration excitation transmission rate measurements based on the joint width and the type of filling material. The third coefficient is used when the path passes through abrupt structural type boundaries between the bridge and roadbed or between the roadbed and tunnel, where structural stiffness and material properties change abruptly, resulting in the lowest vibration energy transmission rate. The specific values ​​for each coefficient are determined through on-site multi-frequency vibration excitation tests before the line is put into operation and updated quarterly based on structural condition monitoring data. The transmission intensity calibration value of the measurement-related channel is obtained by multiplying the distance attenuation coefficient by the structural connectivity coefficient and normalizing.

[0041] The two-factor quantification method decomposes the determination of conduction intensity calibration values ​​into two physically independent effects: asymptotic attenuation due to distance and abrupt interruption of structural connectivity. This accurately characterizes the continuous impact of spatial distance on energy dissipation and the abrupt blocking effect of expansion joints and structural abrupt boundaries on conduction. Compared to empirical estimations relying solely on spatial distance, this method eliminates the bias of treating conduction intensity equally when the path crosses an expansion joint versus when the path is within the same bridge span. This results in higher physical realism in the conduction modeling of electrical measurement correlation topology, providing a more accurate spatial weighting basis for subsequent multi-point differential compensation detection.

[0042] Preferably, the process of obtaining the conduction attenuation calibration parameters specifically includes: For mechanical vibration propagation with solid structure as medium, the vibration energy attenuation characteristics of the solid propagation path between adjacent measurement node pairs are physically calibrated based on the local geometric scale of each measurement node in the electrical measurement association topology, so as to obtain the vibration transmission attenuation calibration value of each measurement node pair. For thermal diffusion propagation using fluid air as the medium, the attenuation characteristics of heat transfer between adjacent measurement node pairs are physically calibrated based on the cross-sectional area and connection length of the enclosed space where each measurement node is located, and the thermal transfer attenuation calibration value of each measurement node pair is obtained. The vibration transmission attenuation calibration value and the thermal transmission attenuation calibration value correspond to independent physical propagation processes, forming an orthogonal conduction attenuation relationship; the signal relay measurement node pair is used as the calibration node pair, and the vibration transmission attenuation calibration value and the thermal transmission attenuation calibration value are used as corresponding components to form conduction attenuation calibration parameters.

[0043] The physical calibration process for vibration transmission attenuation calibration values ​​is as follows: Mechanical vibration propagates in the solid structure of a railway in the form of elastic stress waves. The vibration energy decreases with the propagation distance and the elastic attenuation characteristics of the solid medium. The specific process for vibration transmission attenuation calibration is as follows: using the local geometric scale of each measurement node as a spatial path reference, and simultaneously introducing the material property parameters of the solid structure traversed by the measurement-related channel as joint calibration inputs. The material property parameters include the measured values ​​of the elastic modulus and damping coefficient of the solid structure. The elastic modulus reflects the stiffness characteristics of the structure in the direction of elastic wave propagation; the higher the elastic modulus, the faster the vibration propagation speed and the lower the attenuation coefficient per unit distance. The damping coefficient reflects the structure's ability to dissipate vibration energy; the larger the damping coefficient, the faster the vibration attenuation. The above two material property parameters are obtained by applying a known static load to the solid structure and detecting the strain response. For bridge structures, they can be obtained based on the dynamic performance test data in the periodic bridge inspection report; for roadbed and tunnel lining structures, they are determined according to material design specifications and measured data from similar projects.

[0044] After obtaining the local geometric scale and material property parameters, for each adjacent measurement node pair with an established measurement association channel in the electrical measurement association topology, a calibration vibration exciter is installed at one measurement node, applying a sweeping vibration excitation with a frequency range covering 1 Hz to 200 Hz. A vibration receiving sensor is installed at the other measurement node, recording the vibration acceleration amplitude at each frequency. Using the known applied vibration acceleration amplitude at the exciter as a benchmark, the ratio of the average vibration acceleration amplitude at the receiver within a preset frequency range to the benchmark value is calculated to obtain the vibration transmission attenuation calibration value for that node pair. This physical calibration process comprehensively considers the path geometric length (characterized by the local geometric scale) and the influence of materials on vibration propagation (characterized by the elastic modulus and damping coefficient). Compared to estimating attenuation solely based on geometric scale, this significantly improves the physical accuracy of the vibration transmission attenuation calibration value. For measurement node pairs located between different structural types (bridge structures, tunnel linings, and roadbed soil and rock), calibration is performed according to the elastic modulus and damping coefficient of their corresponding structural types. Material property parameters must not be mixed between different structural types.

[0045] The physical calibration process for heat transfer attenuation calibration values: Heat diffusion occurs in enclosed spaces using air as the fluid medium, and the intensity of heat transfer is constrained by the cross-sectional area and connection length of the enclosed space. For adjacent measurement node pairs within the same tunnel section, a temporary heat source device with known power is placed at one measurement node, and a temperature sensor is placed at the other. The ratio of the temperature rise at the receiving end to the temperature rise at the heat source end during the energized period is recorded as the heat transfer attenuation calibration value for that node pair. For adjacent measurement node pairs in open areas, heat dissipates rapidly into the atmosphere in the open space, and direct heat transfer between node pairs is negligible. The heat transfer attenuation calibration value is taken as the minimum value close to zero, indicating that the heat conduction channel is not connected.

[0046] The establishment of orthogonal conduction attenuation relationships is based on the confirmation of the physical independence of the two propagation mechanisms. After calibration, the orthogonality of the vibration transmission attenuation calibration values ​​and the thermal transmission attenuation calibration values ​​of each adjacent measurement node pair is verified: whether the receiver temperature changes when vibration excitation is applied, and whether the receiver vibration acceleration increases abnormally when a heat source is applied. If there is no significant cross-response, the two types of calibration values ​​are determined to satisfy the orthogonal independence condition. Using the vibration transmission attenuation calibration value as the vibration dimension component and the thermal transmission attenuation calibration value as the thermal dimension component, each calibration node pair is combined to form a two-dimensional conduction attenuation calibration parameter.

[0047] When determining the calibration range of the conduction attenuation calibration parameters, the calibration range strictly corresponds to the range of adjacent measurement node pairs with established measurement association channels in the electrical measurement association topology; the two are the same set. For measurement node pairs whose spatial distance exceeds the preset neighborhood radius and for which no measurement association channel was established during the construction of the electrical measurement association topology, direct conduction attenuation characteristic calibration is not performed. The influence of environmental stress transmission between these long-distance node pairs is indirectly reflected through the step-by-step transmission detection of each adjacent calibrated conduction link in the electrical measurement association topology. They do not need to be calibrated separately, so as to avoid unnecessary calibration workload and maintain the consistency between the calibration system and the topology connection system.

[0048] This invention overcomes the systematic errors caused by neglecting material parameters when estimating attenuation characteristics solely based on geometric dimensions by jointly introducing the elastic modulus and damping coefficient of the solid structure as physical inputs during vibration transmission attenuation calibration. This ensures that the vibration transmission attenuation calibration value accurately reflects the combined effects of path length and material dissipation simultaneously. The strict consistency design between the calibration range and the topology connection range eliminates the logical contradiction between the calibration range and the neighborhood-limited topology range for all nodes, ensuring that the calibration workload matches the actual application range of subsequent compensation testing, avoiding the waste of ineffective calibration resources, and simultaneously giving the entire conduction attenuation calibration parameter system a high degree of internal self-consistency.

[0049] Preferably, the process of obtaining electrical degradation parameters includes: By utilizing the conduction strength calibration value and attenuation calibration value of the conduction attenuation calibration parameter of each measurement association channel in the electrical measurement association topology, multi-point weighted differential compensation detection is performed on the influence of the superposition of physical quantities of the neighboring environment on each measurement node, so as to obtain the multi-physics field distribution detection value of each measurement node at the current time. Based on the multi-physics field distribution detection values ​​and the pre-calibrated environmental physical quantity and electrical quantity response calibration relationship, the common-mode environmental electrical drift caused by terrain load and climate at each measurement node is detected; the electrical action parameters of each measurement node at the current moment are used as the detection values ​​to be compensated, and the common-mode environmental electrical drift is used as the interference compensation amount. Differential interference suppression is performed on the detection values ​​to be compensated to remove the common-mode environmental electrical interference from the electrical action parameters, and further electrical degradation parameters reflecting the accumulation of irreversible damage inside the relay are obtained.

[0050] The goal of multi-point weighted differential compensation detection is to estimate the vibration and thermodynamic field strengths at the current location of the signal relay, providing input for subsequent common-mode drift calculation. Specifically, it consists of two sub-steps: applying baseline constraints and multi-point weighted detection.

[0051] The implementation of the benchmark constraint application sub-step is as follows: For each anchor sensor node where the environmental benchmark detection value has been acquired at the current moment, the measured vibration acceleration value, measured temperature value, and measured humidity value are directly assigned as the multi-physics distribution detection value of that node, serving as a known deterministic benchmark point. These benchmark points act as fixed references for environmental state estimation in space, preventing the estimation results of non-anchor nodes from drifting in unreasonable directions due to error accumulation, thus constraining the estimation space to the boundary of the measured physical quantities.

[0052] The multi-point weighted detection sub-step is executed one by one for measurement nodes in the electrical measurement association topology that do not have sensors installed. For the target measurement node, the current multiphysics distribution detection values ​​of all neighboring measurement nodes that have established a measurement association channel with it are extracted from the electrical measurement association topology. The multiphysics distribution detection values ​​of each neighboring node are weighted in two dimensions: vibration field and thermal field. The weighting coefficient consists of the product of two terms: the first term is the conduction intensity calibration value of the measurement association channel between the neighboring node and the target node, reflecting the degree of physical connection between the two nodes; the second term is the attenuation calibration value of the node in the corresponding physical field dimension in the conduction attenuation calibration parameter, reflecting the energy attenuation ratio of the corresponding physical field during the propagation from the neighboring node to the target node. The contributions of all neighboring nodes are multiplied and weighted, and then divided by the sum of the weighting coefficients of all neighboring nodes to obtain the current vibration field intensity distribution detection value and thermal field intensity distribution detection value of the target node, which together constitute the multiphysics distribution detection value of the measurement node. The calculation process takes into account both spatial distance (reflected by the conduction intensity calibration value) and physical field propagation attenuation (reflected by the attenuation calibration value), which physically correspond to the comprehensive influence of environmental stress propagating in the terrain structure and superimposed on the target node.

[0053] The spatial extrapolation of multiphysics distribution detection values ​​adopts a hierarchical advancement strategy. Before performing weighted extrapolation, all measurement nodes are first sorted in ascending order of the number of hops in the measurement association channels between them and the nearest anchor measurement node, forming an extrapolation priority sequence, with lower hop counts indicating higher priority. During extrapolation, nodes are processed sequentially from high to low priority. All measurement nodes within the same hop count level use the multiphysics distribution detection values ​​of the nodes in the previous hop count level and the direct anchor node as input, without relying on the extrapolation results of nodes in the same or higher hop count levels, thus avoiding recursive dependencies. For non-anchor measurement nodes with a hop count of 1, there is at least one anchor measurement node in its neighborhood, and the input is known, so weighted extrapolation can be performed directly. For non-anchor measurement nodes with a hop count of 2 or higher, the extrapolation values ​​of the required neighboring nodes are determined after the extrapolation of the previous level is completed, and sequential extrapolation can be performed. Once all non-anchor point measurement nodes along the entire line have completed their calculations, the maximum absolute value of the difference between the multiphysics distribution detection value obtained in this calculation and the calculated value from the previous time step is used as the convergence detection value. If this convergence detection value does not exceed the preset convergence detection threshold, the calculation is considered converged and the calculation result is valid. If it exceeds the preset convergence detection threshold, the calculated value from the previous time step is retained and the calculation for that time step is marked as abnormal for subsequent data quality management steps. The preset convergence detection threshold is determined based on the measurement accuracy requirements of each type of sensor, and is usually set to one-thousandth of the range of each physical quantity as the upper limit.

[0054] The detection of common-mode environmental electrical drift relies on a pre-calibrated relationship between environmental physical quantities and electrical quantity responses. This relationship describes the magnitude of reversible shift in various electrical operating parameters of the relay (including armature engagement time, contact resistance, average coil current, and operating current ripple coefficient) when the vibration or thermal field intensity at the location of the signal relay changes by a unit amount; in other words, the response coefficient of each electrical parameter to changes in environmental physical quantities. This response calibration relationship is established during the initial commissioning of newly installed signal relays by applying vibration excitation of known amplitude and temperature / humidity changes to the same model of relay in a controlled laboratory environment, simultaneously recording the shift of each electrical operating parameter. The response coefficient of each electrical parameter is obtained through linear fitting of multiple sets of excitation intensity and corresponding shift data.

[0055] During service, whenever a signal relay at a measurement node is re-energized after maintenance or replacement, the system performs online response calibration on the initial measured data of the new relay to update the response coefficient at that node, ensuring the timeliness of the calibration relationship. The vibration and thermal field intensities from the multiphysics distribution detection values ​​of each measurement node are substituted into the response calibration relationship, multiplied by the corresponding response coefficients, and their contributions to the vibration and thermal field dimensions are summed to obtain the common-mode environmental electrical drift of each measurement node at the current moment. The common-mode environmental electrical drift is a reversible electrical parameter shift caused by current terrain loads and climatic conditions, occurring simultaneously in multiple relays within the region. Its characteristic is that it automatically recovers after the environmental stress is eliminated and does not represent permanent damage to the relay's internal components.

[0056] The differential interference suppression step uses the components of the electrical action parameters at the current moment of each measurement node as the minus and the components of the corresponding common-mode environmental electrical drift as the subtractors. Differential operations are performed component by component to remove common-mode environmental interference components from the measured electrical action parameters one by one. The residual vector obtained from the differential operation is the electrical degradation parameter. The change of this parameter reflects the actual accumulation process of irreversible physical damage such as mechanical wear of the relay contacts, plating erosion, and coil insulation aging. It is unaffected by seasonal temperature fluctuations and train vibration disturbances, exhibits a monotonically increasing degradation trend, and provides a high signal-to-noise ratio degradation signal input for subsequent life prediction.

[0057] Before being used for life prediction, the electrical degradation parameter vector needs to be converted into a single electrical degradation scalar. This conversion is achieved through component normalization followed by weighted synthesis. For each component in the electrical degradation parameter vector, the component detection value for each action is divided by the corresponding reference value, using the product technical specification failure detection threshold or the initial commissioning baseline detection value as a reference. This yields the dimensionless normalized degradation value for each component. When the normalized degradation value of a component exceeds 1, that component has reached its corresponding failure boundary. The synthesis weights of each component are determined using historical degradation data before system commissioning, based on the historical correlation of each component with the relay's final failure. The normalized degradation value of each component is multiplied by its corresponding weight, summed, and then divided by the sum of all weights to obtain the electrical degradation scalar for the current action. This scalar value is greater than zero and typically not greater than 1; a value closer to 1 indicates a closer proximity to the failure boundary. The preset electrical degradation failure detection threshold is set as the composite value corresponding to the electrical degradation scalar reaching the product design failure boundary, which is usually determined based on the maximum degradation amount of each component allowed in the product technical specifications. The weight of each component and the failure threshold are calibrated through product type testing and historical failure cases before the system is put into operation, and are periodically verified during service based on the data of newly put into operation relays at replacement nodes.

[0058] This invention combines spatially measured correlation topological conduction strength with orthogonal conduction attenuation calibration parameters for multi-point weighted differential compensation detection. This achieves precise decoupling of the two-dimensional environmental interference of vibration and thermodynamic fields at the physical measurement level. Furthermore, by suppressing differential interference, reversible environmental responses are removed from electrical action parameters, extracting electrical degradation parameters reflecting irreversible damage accumulation. This method enables the system to stably output accurate degradation state characterization quantities even during periods of severe vibration disturbance, such as in adverse weather or peak traffic periods. This eliminates false degradation alarms caused by terrain loads and climatic factors, significantly improving the reliability of railway signaling system life prediction.

[0059] Preferably, the process of obtaining the remaining life prediction results of the signal relay includes: The electrical degradation parameters of the signal relay are detected to obtain a sequence of individual electrical degradation values ​​for the signal relay. The historical common-mode environmental electrical drift of the neighboring measurement nodes connected to the target signal relay via the measurement association channel is extracted from the electrical measurement association topology. The conduction strength calibration value of the measurement association channel of each neighboring measurement node is used as a coefficient to weight and merge the historical common-mode environmental electrical drift to obtain the common-mode electrical attenuation parameter of the neighboring measurement nodes. The change rate of the common-mode electrical attenuation parameter is detected. When the change rate of the common-mode electrical attenuation parameter continuously exceeds a preset rate detection threshold, it is determined that there is environmental degradation in the area where the target signal relay is located, and multi-node electrical quantity correlation detection is initiated. In the multi-node electrical quantity correlation detection, the change rate of the neighboring common-mode electrical attenuation parameter is used as an environmental acceleration detection factor, which is combined with the initial lifetime prediction detection value obtained from the sequence of individual electrical degradation detection values. The initial lifetime prediction detection value is then reduced and corrected to obtain a lifetime prediction result corrected by multi-node joint electrical quantity detection. Based on the lifetime prediction result, graded early warning information for the signal relay is output.

[0060] The establishment and maintenance of the individual electrical degradation detection value sequence is as follows: After each signal relay completes its operation and outputs the electrical degradation parameter, the system appends the parameter value along with the operation timestamp to the relay's historical sequence, forming a strictly ordered sequence of individual electrical degradation detection values. Each data point in the sequence includes the operation time and the electrical degradation parameter value at that time, comprehensively recording the internal damage accumulation state of the relay after each operation since its commissioning. This sequence continuously grows with operating time and serves as the direct data source for subsequent degradation rate fitting and initial life prediction detection.

[0061] In the electrical measurement association topology, all neighboring measurement nodes that have established a measurement association channel with the target signal relay have stored a sequence of common-mode environmental electrical drift detection values ​​for historical time periods. For the same historical time period, the common-mode environmental electrical drift detection values ​​of each neighboring measurement node at that time are extracted. Using the conduction strength calibration value of the measurement association channel between each neighboring measurement node and the target node as a weighting coefficient, the common-mode drift detection values ​​of each neighboring node are weighted, summed, and then divided by the sum of all weights to obtain the neighborhood weighted common-mode drift value for that time period. The above operation is repeated for all historical time periods to obtain a sequence of neighborhood common-mode electrical attenuation parameters arranged in time sequence. The neighboring node with the larger the conduction strength calibration value contributes more weight to the common-mode parameters of the target node, which physically reflects that relays that are spatially closer and structurally more connected have stronger similarities in their environmental states.

[0062] The determination of environmentally driven degradation is achieved by detecting the rate of change of the common-mode electrical attenuation parameter sequence in the neighborhood. The unit time increment of the common-mode electrical attenuation parameter in the neighborhood within a preset sliding time window is taken as the current rate of change detection value. When the rate of change detection value exceeds the preset rate detection threshold in several consecutive sliding windows, it is determined that there is environmentally driven degradation in the micro-topography area where the target signal relay is located. That is, the comprehensive environmental stress in the area is systematically increasing over time, and the degradation rate of the relay group in the area faces an additional risk of acceleration beyond the normal level, triggering the multi-node electrical quantity correlation detection program.

[0063] The initial life prediction detection value is obtained by linearly extrapolating the degradation trend of the individual electrical degradation detection value sequence. A linear degradation rate is fitted to the individual electrical degradation detection value sequence within the most recent preset statistical time window, representing the average increase in electrical degradation per unit time. The preset electrical degradation failure detection threshold is determined based on the maximum permissible contact resistance and maximum permissible coil leakage current specified in the signal relay product technical specifications. Exceeding this threshold means the relay can no longer reliably perform its safe operation function. The initial life prediction detection value, in time dimension, is obtained by dividing the difference between the preset electrical degradation failure detection threshold and the current electrical degradation by the fitted degradation rate. This value characterizes how long it will take for the electrical degradation to reach the failure judgment boundary if degradation continues at the current rate.

[0064] The dimensionless environmental acceleration detection factor is calculated as follows: The average rate of change of the common-mode electrical attenuation parameter at each measurement node within a reference section of the roadbed, where the terrain is flat and the impact of vibration and temperature / humidity is minimal, is used as the baseline rate of change. This baseline value represents the standard environmental stress level experienced by the relay group under normal environmental conditions. The dimensionless environmental acceleration detection factor is obtained by dividing the current rate of change of the common-mode electrical attenuation parameter in the neighborhood of the target signal relay by the aforementioned baseline rate of change. This factor is a pure ratio, eliminating the dimensional mismatch between the environmental acceleration factor and the lifetime time dimension. When the environmental acceleration detection factor equals 1, the degree of environmental acceleration in the target area is consistent with that in the reference section, and no reduction is needed. When it is greater than 1, the initial lifetime prediction detection value is divided by the environmental acceleration detection factor to obtain the reduced lifetime prediction result after correction by multi-node joint electrical quantity detection. The reduced lifetime prediction result is shorter than the initial value, reflecting the additional consumption of relay lifetime due to environmental stress exceeding normal levels in the target area.

[0065] The lifespan prediction results are compared with the preset graded early warning thresholds in sequence. Based on the remaining lifespan, the system is divided into four early warning levels: normal operation, observation, planned maintenance, and emergency response. Each level corresponds to different operation and maintenance response measures. For the planned maintenance level and below, a window period replacement plan is arranged. For the emergency response level, the emergency spare parts replacement procedure is immediately initiated, thus forming the final graded early warning information output for each signal relay.

[0066] This invention incorporates the rate of change of the common-mode electrical attenuation parameter in the neighborhood into the lifetime prediction correction mechanism. This allows the system to detect the accelerated degradation of equipment in the region as a whole before the degradation parameters of a single signal relay reach the alarm threshold. This advances the warning window by weeks to months, providing maintenance departments with ample time to prepare for planned maintenance. The normalization design of the dimensionless environmental acceleration detection factor ensures the dimensional consistency of the reduction calculation, making lifetime prediction results from different terrain conditions comparable. This comprehensively improves the railway signaling system's ability to proactively detect degradation driven by hidden environmental factors and enhances the accuracy of maintenance scheduling.

[0067] Example 2:

[0068] This invention provides a long-term lifetime prediction method for railway signal relays, comprising: Obtain the geographical coordinates, terrain features, and environmental anchor point data of signal relays along the railway line; When the signal relay operates, the current waveform of the entire process is collected triggered by the jump in coil level, and the transient motion stage and steady-state contact stage are divided based on the current change rate threshold to obtain the electrical action parameters. The channel identifier and local geometric scale of the signal relay are assigned to each measurement node in the electrical measurement association topology; based on the terrain features, the conduction attenuation characteristics of mechanical vibration medium and thermal diffusion fluid are calibrated respectively, and an orthogonal conduction attenuation relationship is established to obtain the conduction attenuation calibration parameters; Using environmental anchor point data as physical reference constraints, the electrical action parameters are detected by multi-point differential compensation using electrical measurement correlation topology and conduction attenuation calibration parameters. The multi-physics distribution detection values ​​of each measurement node are obtained, and the common-mode environmental electrical drift affected by terrain load and climate is removed from the electrical action parameters to obtain electrical degradation parameters. Based on electrical degradation parameters and common-mode electrical attenuation parameters of neighboring measurement nodes in the electrical measurement correlation topology, multi-node electrical quantity correlation detection is performed, and the life prediction results of each signal relay are output.

[0069] Furthermore, in the transient motion detection segment, the detected contact bounce events are classified into three degradation mechanism types based on the duration of the first bounce, the decreasing pattern of the duration of each bounce, and the attenuation pattern of the peak amplitude of the bounce: spring elasticity decay type, contact surface contamination type, and contact micro-welding risk type. A high number of bounces with a slow, proportional decrease in duration is classified as spring elasticity decay type; a high number of bounces with approximately equal durations is classified as contact surface contamination type; and an abnormally long first bounce duration with very few subsequent bounces is classified as contact micro-welding risk type. A bounce mechanism type label is added to the transient motion electrical parameters, and independent long-time sequence detection records of characteristic parameters are maintained for each of the three types.

[0070] The physical causes and specific distinguishing characteristics of the three bounce patterns are as follows: In the spring elastic decay type, the contact spring constant gradually decreases with the increase of the cumulative number of actions, weakening the reset force after the armature is attracted, resulting in insufficient contact closure buffering. This mode is characterized by the initial bounce duration being within the normal reference range, but subsequent bounces exceeding the upper limit of the normal reference range. The duration of each bounce decreases slowly, and the peak amplitude of each bounce decreases approximately proportionally. The criterion is that the total number of subsequent bounces exceeds a preset threshold and the decay slope of each bounce duration is lower than a preset slope threshold, both of which must be satisfied simultaneously.

[0071] The surface contamination type corresponds to conductive foreign matter adhering to the contact surface of the contact point, forming tiny contact bridges that hinder stable contact closure. This mode is characterized by an extremely short initial bounce duration, followed by continuous, dense, small bounces. The duration of each bounce is approximately equal, and the bounce intervals are uniform. The criterion is that the bounce uniformity coefficient exceeds a preset value. The bounce uniformity coefficient is defined as the reciprocal of the ratio of the standard deviation of each bounce duration to the mean; a higher ratio indicates more uniform bounce.

[0072] The contact micro-welding risk type corresponds to the contact undergoing micro-fusion welding due to arc heat during the high current switching moment, causing the contacts to stick together and be forcibly pulled apart. This mode is characterized by an initial bounce duration exceeding twice the normal reference value, a bounce peak amplitude that is the highest among the three modes, and one or no subsequent bounces. The criterion is that the ratio of the initial bounce duration to the average of subsequent bounce durations exceeds a preset threshold. After each action detection, the bounce mechanism type label is appended to the transient motion electrical parameters and output together.

[0073] The three-type classification of contact bounce modes upgrades transient motion electrical parameters from simple bounce count indicators to diagnostic information carrying failure mechanism type labels. This enables the life prediction system to identify three degradation paths with significantly different physical progression rates and hazard levels: elasticity decay, surface contamination, and micro-welding risk. Shorter warning thresholds are set for contact micro-welding risk type, and longer replacement cycles are set for spring elasticity decay type, significantly improving the precision of contact failure mechanism diagnosis and the pertinence of maintenance decisions.

[0074] Furthermore, when performing multi-point differential compensation detection of vibration field components, the estimated train passage time of each measurement node section is obtained from the train dispatching system. A vibration pulse isolation window is established within the time range specified above, with the estimated train arrival time as the starting time and the preset isolation time required for the vibration acceleration to decay to within the background mean plus a preset multiple standard deviation after the train passes as the ending time. Within the isolation window, the vibration field distribution detection value at the start of the window is used to replace the real-time calculated value, and the update of the vibration field components is suspended. The thermal field components continue to undergo multi-point differential compensation detection within the isolation window without being affected. After the isolation window ends, normal compensation detection of the vibration field components resumes.

[0075] The peak value of wheel-rail impact vibration generated when a train passes far exceeds the daily background vibration level of the railway structure. This peak value is short in duration and high in intensity, constituting a pulsed disturbance superimposed on the background vibration. Including vibration data collected during train passage in multi-point differential compensation detection would cause a sudden overestimation of the vibration field intensity during the train's passage period. This would lead to a significant overestimation of the common-mode environmental electrical drift during this period, resulting in a false step change in the electrical degradation parameter sequence that is strictly synchronized with the train's running time, thus interfering with the stable detection of degradation trends.

[0076] The duration of the vibration pulse isolation window is determined according to the terrain type. In bridge sections, the vibration damping of reinforced concrete structures is low, and the vibration decay is slow, so the isolation time is 90 seconds after the train leaves. In roadbed sections, the soil damping is high, and the vibration decay is fast, so the isolation time is 45 seconds. In tunnel sections, due to the multiple reflection effects of vibration from the enclosed lining structure, the decay is the slowest, so the isolation time is 120 seconds. The above durations are determined statistically by recording the actual time required for the vibration acceleration to recover to within three times the background mean after a train passes in each typical terrain section, and are revised quarterly based on the measured decay time data after the system is put into operation. Within the isolation window, the system maintains the vibration field distribution detection value at the beginning of the window unchanged and continues to perform normal compensation detection of the thermal field component until the vibration isolation window ends and the vibration field is recalculated in real time.

[0077] The vibration pulse isolation time window mechanism actively isolates and excludes the periodic and predictable characteristics of train passing vibration pulses in railway scenarios from the vibration field input of multi-point differential compensation detection. This eliminates the periodic interference of high-density train passing vibrations on the estimation of common mode drift during peak traffic periods, so that the electrical degradation parameter sequence no longer exhibits false fluctuations closely related to train operation time on busy lines. This significantly improves the stability and reliability of degradation trend detection in high-frequency high-speed railways or urban rail transit scenarios.

[0078] 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.

Claims

1. A method for predicting the long-term lifespan of railway signal relays, characterized in that, include: Obtain the geographical coordinates, terrain features, and environmental anchor point data of signal relays along the railway line; When the signal relay operates, the current waveform of the entire process is collected triggered by the jump in coil level, and the transient motion stage and steady-state contact stage are divided based on the current change rate threshold to obtain the electrical action parameters. The process of obtaining electrical action parameters specifically includes: configuring intelligent electrical parameter sensing units at each signal relay; the intelligent electrical parameter sensing units use the level transition of the signal relay drive control signal as a synchronous trigger signal, triggering the acquisition of the coil current waveform at the level transition moment, recording the current waveform of the entire process from the trigger moment until the coil current returns to stability, and obtaining the action current waveform detection signal; detecting the current change rate point by point on the action current waveform detection signal, using a pre-calibrated current change rate threshold as a criterion, and using the moment when the current change rate is lower than the current change rate threshold and the stability condition is met as the dividing point, dividing the action current waveform detection signal into a transient motion detection segment and a steady-state contact detection segment; detecting the armature engagement time, contact bounce count, bounce duration, and current dip depth in the transient motion detection segment to obtain transient motion electrical parameters; detecting the contact resistance value, average coil operating current, and operating current ripple coefficient in the steady-state contact detection segment to obtain steady-state contact electrical parameters; and merging the transient motion electrical parameters and steady-state contact electrical parameters to obtain the electrical action parameters of the signal relay action; The channel identifier and local geometric scale of the signal relay are assigned to each measurement node in the electrical measurement association topology; based on the terrain features, the conduction attenuation characteristics of mechanical vibration medium and thermal diffusion fluid are calibrated respectively, and an orthogonal conduction attenuation relationship is established to obtain the conduction attenuation calibration parameters; Using environmental anchor point data as physical reference constraints, the electrical action parameters are detected by multi-point differential compensation using electrical measurement correlation topology and conduction attenuation calibration parameters. The multi-physics distribution detection values ​​of each measurement node are obtained, and the common-mode environmental electrical drift affected by terrain load and climate is removed from the electrical action parameters to obtain electrical degradation parameters. The process of obtaining electrical degradation parameters includes: using the conduction strength calibration value and conduction attenuation calibration value of each measurement association channel in the electrical measurement association topology, performing multi-point weighted differential compensation detection on the influence of the superposition of physical quantities of the neighboring environment on each measurement node, and obtaining the multi-physics field distribution detection value of each measurement node at the current time; based on the multi-physics field distribution detection value and the pre-calibrated environmental physical quantity and electrical quantity response calibration relationship, detecting the common-mode environmental electrical drift of each measurement node caused by terrain load and climate; using the electrical action parameters of each measurement node at the current time as the detection value to be compensated, and using the common-mode environmental electrical drift as the interference compensation amount, performing differential interference suppression on the detection value to be compensated, removing the common-mode environmental electrical interference from the electrical action parameters, and further obtaining electrical degradation parameters reflecting the accumulation of irreversible damage inside the relay; Based on electrical degradation parameters and common-mode electrical attenuation parameters of neighboring measurement nodes in the electrical measurement correlation topology, multi-node electrical quantity correlation detection is performed, and the life prediction results of each signal relay are output.

2. The long-term life prediction method for railway signal relays according to claim 1, characterized in that: Obtain the geographic coordinates, terrain features, and environmental anchor point data of relays along the railway line, specifically including: The longitude coordinates, latitude coordinates, and rail surface elevation of each signal relay are obtained to obtain the spatial position parameters of the signal relay, which are used as geographical coordinates. The load-bearing structures along the railway line are measured to distinguish between bridge sections, tunnel sections, cutting sections and roadbed sections. For bridge sections, the bridge span and pier spacing are measured and calculated. For tunnel sections, the clearance height and cross-sectional area are measured to obtain the geometric scale parameters of each section as topographic features. Multi-parameter intelligent sensing nodes are deployed at signal relays along the line. These nodes integrate vibration acceleration detection units, temperature detection units, and humidity detection units. They continuously collect vibration acceleration, temperature, and humidity data at the measurement points, and after local preprocessing, upload the data through a communication interface to obtain environmental benchmark detection values, which serve as environmental anchor point data.

3. The long-term life prediction method for railway signal relays according to claim 1, characterized in that: The process of constructing an electrical measurement association topology includes: Using signal relays along the line as measurement nodes, a channel identifier is assigned to each measurement node. The channel identifier includes the station code, cabinet number, and slot identifier, and the signal relay corresponding to the measurement node is uniquely calibrated. The topographic features of the terrain area where each measurement node is located are used as the local geometric scale and assigned to the corresponding measurement node. Based on the geographical coordinates of each measurement node, measurement association channels are established for measurement node pairs whose spatial distance is less than the preset neighborhood radius. The conduction strength calibration value of the measurement association channel is jointly calibrated by the spatial distance of the measurement node pairs and the type of terrain bearing structure traversed. The channel identifier, local geometric scale of each measurement node and the conduction strength calibration value of each measurement association channel are combined to form an electrical measurement association topology covering the entire line.

4. The long-term life prediction method for railway signal relays according to claim 1, characterized in that: The process of obtaining conduction attenuation calibration parameters specifically includes: For mechanical vibration propagation with solid structure as medium, the vibration energy attenuation characteristics of the solid propagation path between adjacent measurement node pairs are physically calibrated based on the local geometric scale of each measurement node in the electrical measurement association topology, so as to obtain the vibration transmission attenuation calibration value of each measurement node pair. For thermal diffusion propagation using fluid air as the medium, the attenuation characteristics of heat transfer between adjacent measurement node pairs are physically calibrated based on the cross-sectional area and connection length of the enclosed space where each measurement node is located, and the thermal transfer attenuation calibration value of each measurement node pair is obtained. The vibration transmission attenuation calibration value and the thermal transmission attenuation calibration value correspond to independent physical propagation processes, forming an orthogonal conduction attenuation relationship; the signal relay measurement node pair is used as the calibration node pair, and the vibration transmission attenuation calibration value and the thermal transmission attenuation calibration value are used as corresponding components to form conduction attenuation calibration parameters.

5. A long-term life prediction method for railway signal relays according to claim 1, characterized in that: The process of obtaining the remaining life prediction results for signal relays includes: The electrical degradation parameters of the signal relay are detected to obtain a sequence of individual electrical degradation values ​​for the signal relay. The historical common-mode environmental electrical drift of the neighboring measurement nodes connected to the target signal relay via the measurement association channel is extracted from the electrical measurement association topology. The conduction strength calibration value of the measurement association channel of each neighboring measurement node is used as a coefficient to weight and merge the historical common-mode environmental electrical drift to obtain the common-mode electrical attenuation parameter of the neighboring measurement nodes. The change rate of the common-mode electrical attenuation parameter is detected. When the change rate of the common-mode electrical attenuation parameter continuously exceeds a preset rate detection threshold, it is determined that there is environmental degradation in the area where the target signal relay is located, and multi-node electrical quantity correlation detection is initiated. In the multi-node electrical quantity correlation detection, the change rate of the neighboring common-mode electrical attenuation parameter is used as an environmental acceleration detection factor, which is combined with the initial lifetime prediction detection value obtained from the sequence of individual electrical degradation detection values. The initial lifetime prediction detection value is then reduced and corrected to obtain a lifetime prediction result corrected by multi-node joint electrical quantity detection. Based on the lifetime prediction result, graded early warning information for the signal relay is output.

Images