Building low-voltage cable fire hazard monitoring and positioning system based on multimode monitoring

Abstract

The invention relates to a building low-voltage cable fire hazard monitoring and positioning system based on multimode monitoring, and belongs to the technical field of cable hazard monitoring. The system comprises a multi-mode monitoring unit and an intelligent early warning host, wherein the intelligent early warning host is in communication connection with the multi-mode monitoring unit; the multi-mode monitoring unit is pre-buried in the cable and is used for synchronously acquiring multi-dimensional state parameters of the cable; and the intelligent early warning host is used for receiving and processing data of the multimode monitoring unit, and outputting graded early warning information including hidden danger types and positions through adaptive dynamic threshold analysis, multi-parameter correlation diagnosis and BIM fusion positioning. According to the invention, early perception of hidden dangers is realized through a multi-mode monitoring system, electrical distance is converted into three-dimensional space coordinates by using a BIM fusion technology to realize accurate positioning, interference and real hidden dangers are effectively distinguished by using a self-adaptive threshold and multi-level association early warning mechanism, and early warning accuracy and disposal efficiency are significantly improved.

Description

Technical Field

[0001] This invention belongs to the field of cable hazard monitoring technology, specifically relating to a building low-voltage cable fire hazard monitoring and location system based on multi-mode monitoring. Background Technology

[0002] Low-voltage power distribution cables in buildings are critical facilities at the end of the power system. During long-term operation, they are prone to abnormal temperature rise due to overload, insulation aging, loose contact points, etc., which is one of the main hidden dangers of electrical fires.

[0003] Traditional monitoring methods (such as infrared thermometry and point-type temperature detectors) have limitations, including discontinuous monitoring, difficulty in location, and inability to warn of early, hidden faults. While distributed fiber optic temperature measurement technology can achieve continuous temperature monitoring, it typically only provides linear distance alarms, failing to directly correlate with the actual location within the building structure. Furthermore, the single temperature parameter is susceptible to environmental interference, leading to false alarms or missed alarms. In addition, the lack of simultaneous perception and correlation analysis of multi-dimensional information such as insulation status and load current makes it difficult to accurately determine the cause and development stage of potential hazards, hindering precise intervention by maintenance personnel. Summary of the Invention

[0004] To address the aforementioned problems in existing technologies, this invention provides a building low-voltage cable fire hazard monitoring and location system based on multi-mode monitoring. The system achieves early detection of hazards through a multi-mode monitoring system, utilizes BIM fusion technology to convert electrical distances into three-dimensional spatial coordinates for precise location, and employs an adaptive threshold and multi-level correlation early warning mechanism to effectively distinguish between interference and real hazards, significantly improving the accuracy of early warning and the efficiency of response.

[0005] The objective of this invention can be achieved through the following technical solutions: The first aspect of this disclosure provides a building low-voltage cable fire hazard monitoring and location system based on multi-mode monitoring, including a multi-mode monitoring unit and an intelligent early warning host, wherein the intelligent early warning host is communicatively connected to the multi-mode monitoring unit; The multi-mode monitoring unit is embedded in the cable to synchronously collect multi-dimensional state parameters of the cable, including temperature distribution data, dielectric loss data, real-time load current data, and time-domain reflection signal. The intelligent early warning host is used to receive and process data from the multi-mode monitoring unit, and output graded early warning information containing the type and location of hidden dangers through adaptive dynamic threshold analysis, multi-parameter correlation diagnosis and BIM fusion positioning.

[0006] The intelligent early warning host includes a multi-level early warning and false alarm suppression module. This module is used to monitor and issue early warnings of fire hazards by executing early warning logic rules, which include: The system judges single-parameter anomalies based on dynamic early warning thresholds. When the abnormal state continues to exceed the short-term observation window, a low-level early warning is triggered. The single parameters include temperature, dielectric loss factor and load current. Based on the low-level warning start timer, if the abnormal state continues for longer than the long confirmation window, parameter correlation analysis is triggered; otherwise, the low-level warning is lifted. The parameter correlation analysis includes: Based on preset physical association rules, it is determined whether the correlation between the three parameters of temperature, dielectric loss factor and load current is successful. If the correlation is successful, an advanced warning is triggered; otherwise, an intermediate warning is triggered. Advanced warnings include hazard diagnosis results, BIM 3D positioning information and hazard patterns, while intermediate warnings include single-parameter anomaly monitoring reminders.

[0007] As a preferred embodiment of the present invention, the multi-mode monitoring unit includes a distributed temperature sensing module and an insulation status monitoring module; The distributed temperature sensing module is used to acquire temperature distribution data in the continuous space of the cable based on the Raman scattering principle; the insulation condition monitoring module is used to collect the dielectric loss value of the cable insulation layer in real time based on the dielectric loss factor measurement technology.

[0008] As a preferred embodiment of the present invention, the multi-mode monitoring unit further includes an electrical load monitoring module and a positioning signal acquisition module; The electrical load monitoring module is used to monitor the real-time load current of the cable; the positioning signal acquisition module, based on time-domain reflection technology and a feature reflector set at a preset position on the cable, is used to inject positioning signals into the cable to acquire time-domain reflection signals and obtain a description of the location of the abnormal point.

[0009] As a preferred embodiment of the present invention, the step of obtaining the location description of the anomaly point includes the following steps: Standardized feature reflectors are pre-positioned at predetermined locations on the cable to establish an absolute position reference system; nanosecond-level fast-edge pulses are injected into the cable through a high-frequency coupling circuit, and its reflected signals are received synchronously. The reflected signal is amplified and then digitally sampled and combined by an ultra-high-speed analog-to-digital converter to form a complete time-domain reflection waveform. All characteristic reflection peaks are identified and matched on the waveform, and the actual physical location of each characteristic reflection peak is determined according to the location of the characteristic reflector. The time difference between the abnormal reflection peak and the nearest upstream and downstream characteristic reflection peaks is measured to obtain the distance of the abnormal point from the upstream and downstream characteristic reflectors.

[0010] As a preferred embodiment of the present invention, the intelligent early warning host further includes a BIM fusion positioning module and an adaptive threshold analysis module; The BIM fusion positioning module is used to store a building information model containing the digital cable laying path, and to match the abnormal point location description obtained by the positioning signal acquisition module with the path in the building information model to obtain the visualized location information of the potential hazard point in the three-dimensional space of the building.

[0011] As a preferred embodiment of the present invention, obtaining the visualized location information of the potential hazard point in the three-dimensional space of the building includes the following steps: Based on cable laying drawings or laser scanning data, a cable path model is created in the building information model. The three-dimensional coordinates of the pre-set feature reflectors in the cable path model are bound to their unique identification codes. Based on the known electrical distance of the feature reflectors, a "physical location-electrical distance" mapping table is established. Receive the description of the location of the anomaly point, retrieve the three-dimensional coordinates of the upstream and downstream feature reflectors in the cable path model from the mapping table based on the identification code of the feature reflector in the description, and extract the cable path segment connecting the upstream and downstream feature reflectors. The total spatial length between upstream and downstream feature reflectors in the cable path model is obtained by accumulating the Euclidean distances between each node of the cable path segment. The normalization scaling factor of the anomaly point in the interval between the upstream and downstream feature reflectors is calculated based on the electrical distance between the anomaly point and the feature reflector. The selected feature reflector starts by accumulating segmented distances along the node sequence. When the accumulated distance reaches or exceeds the target distance, the node interval is located. Linear interpolation is performed on the straight line segment formed by the nodes to calculate the three-dimensional coordinates of the potential hazard point. The calculated three-dimensional coordinates are then converted to the view coordinate system of the BIM visualization engine and highlighted on the BIM three-dimensional model interface.

[0012] As a preferred technical solution of the present invention, the adaptive threshold analysis module has a built-in machine learning model trained with historical normal operation data, which is used to dynamically obtain the temperature warning threshold based on the real-time ambient temperature and load current. The dynamically acquired temperature warning threshold includes the following steps: Offline model training: A large number of training data samples are collected synchronously during the normal operation of the cable; each sample includes ambient temperature, real-time load current, historical moving average, time features, and actual temperature at each monitoring point; after cleaning and normalizing the data samples, a dataset is constructed according to spatial points; gradient boosting regression tree or random forest is selected for training, and the training objective is set to minimize the error between the model's predicted temperature and the measured normal temperature; after training, the model performance is verified using data that was not used in the training. Online real-time prediction: Obtain the latest ambient temperature and real-time load current, calculate their historical moving average, and combine them with the current time characteristics to form a real-time input feature vector; input the real-time input feature vector into a trained machine learning model to obtain the predicted temperature of each monitoring point along the entire cable under the current operating conditions, and calculate the temperature warning threshold based on a dynamic adjustment formula.

[0013] As a preferred embodiment of the present invention, the expression of the dynamic adjustment formula is: T threshold (i) = T base (i) + α * σ(i) + β; In the formula, T threshold (i) is the warning threshold at point i, T base (i) represents the predicted temperature at point i, α is the sensitivity coefficient, σ(i) is the standard deviation of the uncertainty in the model's prediction at point i, and β is the safety margin constant.

[0014] As a preferred embodiment of the present invention, the preset of the physical association rule includes the following steps: Construction of transaction dataset: Based on prior knowledge, define the pattern characteristics of each hidden danger on the monitoring parameters, collect the parameters of the hidden danger status, combine the parameter status snapshots within the time window into the premise item set of the event transaction, the parameter status snapshot includes the pattern characteristics transformed by the monitoring parameters, and label the hidden danger type as the conclusion item for each event transaction to obtain the transaction database; Association rule mining: Set minimum support and minimum confidence, use association rule mining algorithms to scan the entire transaction database, find all frequently occurring itemsets, and generate all association rules from each frequent itemset; Rule filtering: Strongly correlated rules are filtered out from all association rules by calculating confidence scores, and rules that violate physical mechanisms are removed through review by domain experts.

[0015] The second aspect of this disclosure provides a method for monitoring and locating fire hazards in building low-voltage cables based on multi-mode monitoring, applied to a building low-voltage cable fire hazard monitoring and location system based on multi-mode monitoring as described above, including the following steps: S1. Synchronously and in real time acquire the temperature distribution data, insulation dielectric loss factor, real-time load current data, and time domain reflection signal of the entire cable. S2. Based on the real-time ambient temperature and load current, dynamically calculate the adaptive temperature warning threshold at each point along the entire cable under the current operating conditions using a pre-trained machine learning model. S3. Compare the real-time monitored temperature, dielectric loss factor and load current data with the corresponding early warning thresholds to determine the abnormal state of a single parameter. S4. If an abnormal state occurs, start the timer; if the abnormal state continues for longer than the short observation window, analyze the correlation between the three parameters of temperature, dielectric loss factor and current, and determine whether the combination conforms to the preset physical correlation rules in order to identify the type of hidden danger. S5. If only a single parameter is abnormal, a medium-level warning is triggered; if multiple parameters are associated and matched successfully, and this state continues for more than the long-term confirmation window, a fire hazard is confirmed and a high-level warning is triggered. S6. Output the hazard diagnosis type. At the same time, based on the electrical distance of the abnormal point relative to the preset feature reflector, the path matching and spatial interpolation calculation are performed in the building information model through the BIM fusion positioning module, and the three-dimensional spatial coordinates of the hazard point are output.

[0016] The beneficial effects of this invention are as follows: This invention first achieves continuous spatial sensing of cable temperature across the entire cable length by pre-embedding distributed optical fibers based on Raman scattering principles. Simultaneously, it collects and integrates three key parameters—temperature, dielectric loss, and real-time load current—to construct a multi-dimensional monitoring system. Secondly, it deeply integrates time-domain reflectometry (TDAR) positioning technology with Building Information Modeling (BIM). By pre-setting feature reflectors at cable joints and bends, the BIM-integrated positioning module intelligently matches and spatially interpolates the electrical distance obtained from TDAR positioning with the cable's digital laying path in the 3D building model, transforming the abstract "cable length" into intuitive "building spatial coordinates," thus solving the problems of unintuitive linear alarm positioning and difficulty in finding the correct location. Finally, it introduces adaptive dynamic threshold analysis based on a machine learning model, which dynamically adjusts the warning threshold according to real-time ambient temperature and load current, avoiding false alarms due to fixed thresholds caused by environmental or load fluctuations. Furthermore, it establishes a multi-level warning logic based on multi-parameter association rules and time window verification. A higher-level alarm is triggered only when the three parameters—temperature, dielectric loss, and current—meet the preset physical logic and the abnormal state persists beyond the confirmation time window, simultaneously outputting a hazard type diagnosis. This filters out transient interference, ensuring the accuracy and guidance of the alarm. Attached Figure Description

[0017] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.

[0018] Figure 1 This is a schematic diagram of a building low-voltage cable fire hazard monitoring and location system based on multi-mode monitoring, provided in an embodiment of the present invention. Figure 2 This is a schematic diagram of the structure of the multi-mode monitoring unit provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of the intelligent early warning host provided in an embodiment of the present invention; Figure 4 This is a schematic diagram illustrating the steps of a method for monitoring and locating fire hazards in building low-voltage cables based on multi-mode monitoring, provided in an embodiment of the present invention. Detailed Implementation

[0019] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided.

[0020] This embodiment provides a building low-voltage cable fire hazard monitoring and location system based on multi-mode monitoring, such as... Figure 1 As shown, it includes a multi-mode monitoring unit and an intelligent early warning host, and the intelligent early warning host is communicatively connected to the multi-mode monitoring unit.

[0021] The multi-mode monitoring unit is embedded in the cable to synchronously collect multi-dimensional state parameters of the cable, including temperature distribution data, dielectric loss data, real-time load current data, and time-domain reflection signal.

[0022] like Figure 2 As shown, the multi-mode monitoring unit includes at least: A distributed temperature sensing module, based on the Raman scattering principle, is used to acquire temperature distribution data over a continuous space within a cable. The acquisition of this temperature distribution data includes the following steps: Laser pulses are injected into a cable with a pre-embedded sensing fiber using an optical fiber coupler. Backscattered light is collected and wavelength division multiplexed to separate Stokes and anti-Stokes light. The two separated optical signals are converted into electrical signals and digitized at high speed. The time of flight is used to determine the location of the scattering point. The temperature of each spatial point is calculated by combining the dual intensity ratio method with calibration constants.

[0023] The insulation condition monitoring module, based on dielectric loss factor measurement technology, is used to acquire the dielectric loss value of the cable insulation layer in real time; the acquisition of the dielectric loss value of the cable insulation layer includes the following steps: The non-invasive synchronous coupling of the operating voltage signal and leakage current signal enables high-precision synchronous sampling and digitization of the two signals. The phase difference between the two fundamental signals is accurately extracted using digital signal processing algorithms (such as zero-crossing method and correlation method). The dielectric loss factor is calculated according to the formula tanδ = tan(90°-|φ|), where φ is the phase difference and δ is the dielectric loss angle. Finally, temperature compensation and system calibration are performed on the measurement results to output a stable and reliable tanδ value.

[0024] The electrical load monitoring module is used to monitor the real-time load current of the cable. Specifically, it uses a non-invasive current sensor (such as a Rogowski coil) to safely acquire the current signal, amplifies, filters, and performs high-resolution synchronous sampling on the signal to ensure time alignment. It extracts the true effective value of the current, harmonics, unbalance, and other core characteristics through digital signal processing (integration, FFT, RMS calculation).

[0025] The positioning signal acquisition module, based on time-domain reflectometry and a feature reflector set at a preset cable position, is used to inject positioning signals into the cable to acquire time-domain reflectometry signals and obtain descriptions of anomaly locations.

[0026] Obtaining the location description of an anomaly point includes the following steps: Standardized characteristic reflectors (impedance discontinuity elements) are pre-positioned at predetermined locations on the cable (such as joints, bends, and branch points) to establish an absolute position reference system; nanosecond-level fast-edge pulses are injected into the cable through a high-frequency coupling circuit, and its reflected signals are received synchronously.

[0027] The reflected signal is amplified and then digitally sampled and combined using an ultra-high-speed analog-to-digital converter to form a complete time-domain reflection waveform. All characteristic reflection peaks (characteristic reflection peaks generated by all preset characteristic reflectors) are identified and matched on the waveform. The actual physical location corresponding to each characteristic reflection peak is determined based on the location of the characteristic reflector (e.g., "Connector A", 50 meters away from the host).

[0028] The time difference between the abnormal reflection peak (when a hidden danger such as local overheating occurs, the cable impedance at this point will change slightly, and an additional, smaller abnormal reflection peak will be generated between two adjacent known characteristic reflection peaks on the waveform) and the nearest upstream and downstream characteristic reflection peaks is measured to obtain the distance of the abnormal point from the upstream and downstream characteristic reflectors. The specific distance calculation formula is ΔL = v*Δt / 2, where v is the propagation speed of electromagnetic waves in the cable, and Δt is the time difference between the abnormal reflection peak and the upstream or downstream characteristic reflection peak. The final description of the abnormal point location is as follows: the abnormal point is located between characteristic reflector A (50 meters away) 7.3 meters downstream and characteristic reflector B (65 meters away) 7.7 meters upstream.

[0029] The intelligent early warning host is used to receive and process data from the multi-mode monitoring unit, and outputs graded early warning information containing the type and location of potential hazards through adaptive dynamic threshold analysis, multi-parameter correlation diagnosis, and BIM fusion positioning. Figure 3 As shown, the intelligent early warning host includes a BIM fusion positioning module, an adaptive threshold analysis module, and a multi-level early warning and false alarm suppression module.

[0030] The BIM fusion positioning module is used to store a building information model containing the digital cable laying path, and to match the abnormal point location description obtained by the positioning signal acquisition module with the path in the building information model to obtain the visualized location information of the potential hazard point in the three-dimensional space of the building.

[0031] Obtaining the visualized location information of potential hazards in the three-dimensional space of the building includes the following steps: Based on cable laying drawings or laser scanning data, a cable path model is created in the building information model. The cable path model consists of a series of continuous nodes with three-dimensional coordinates, accurately reflecting the actual route of the cable in the building space.

[0032] Bind the three-dimensional coordinates of the preset feature reflector in the cable path model to its unique identifier, and establish a "physical location-electrical distance" mapping table based on the known electrical distance of the feature reflector (e.g., "50 meters from the host").

[0033] The system receives the location description of the anomaly point, retrieves the three-dimensional coordinates of the upstream and downstream feature reflectors in the cable path model from the mapping table based on the identifier code of the feature reflector in the description, and extracts the cable path segment connecting the upstream and downstream feature reflectors (this path segment is usually defined by a series of more detailed node coordinate sequences, which accurately describes the actual spatial trajectory of the cable between the two points, such as possibly including bends and climbs).

[0034] The total spatial length between upstream and downstream feature reflectors in the cable path model is obtained by summing the Euclidean distances between each node in the cable path segment. The normalized scaling factor (k = L) of the anomaly point within the upstream and downstream feature reflector interval is calculated based on the electrical distance between the anomaly point and the feature reflector. TDR / D TDR L TDR D is the electrical distance from the anomaly point to the characteristic reflector. TDR (where the electrical distance is between the characteristic reflectors), thus mapping the electrical distance scale to the spatial distance scale.

[0035] The selected feature reflector starts by accumulating segmented distances along the node sequence. When the accumulated distance reaches or just exceeds the target distance (the electrical distance between the anomaly point and the initial feature reflector), the node interval is located. Linear interpolation is performed on the straight line segment formed by the nodes to calculate the three-dimensional coordinates of the hidden danger point. The calculated three-dimensional coordinates are then converted to the view coordinate system of the BIM visualization engine and highlighted on the BIM three-dimensional model interface.

[0036] For example: the anomaly point is located at characteristic reflector A (electrical distance L). A Downstream L TDR It is located at a point between A and B (the electrical distance from B is L).B Then the electrical distance D from A to B TDR =L B -L A (); Obtain the three-dimensional coordinates of A and B from the BIM model, and calculate the spatial length L along the cable path from A to B. BIM (By accumulating the Euclidean distances between adjacent nodes); calculate the electrical distance ratio: k electric = L TDR / D TDR ; Calculate the spatial distance L from the anomaly point to point A. physical = k electric *L BIM Starting from point A, along the cable path node sequence in the BIM model, accumulate the distances between adjacent nodes until the accumulated value reaches or exceeds L. physical When the accumulated value just exceeds L physical At that time, it locates the location between two adjacent nodes.

[0037] The adaptive threshold analysis module has a built-in machine learning model trained on historical normal operation data, which is used to dynamically obtain the temperature warning threshold based on real-time ambient temperature and load current.

[0038] The dynamically acquired temperature warning threshold includes the following steps: Offline model training: A large number of training data samples are collected synchronously during the normal operation of the cable; each sample includes ambient temperature, real-time load current, historical moving average (the average load current over a previous period), time features (time of day, day of the week, season), and actual temperature at each monitoring point; after cleaning and normalizing the data samples, a dataset is constructed according to spatial points.

[0039] Choose gradient boosting regression tree or random forest for training, set minimizing the error (such as mean square error) between the model's predicted temperature and the measured normal temperature as the training objective, and use data that was not used in the training to verify the model's performance after training.

[0040] Understandably, this embodiment selects gradient boosting regression tree or random forest. These ensemble tree models can automatically capture the complex nonlinear relationships and interactions between ambient temperature, load current (and its history) and cable temperature (e.g., the same load will produce a higher temperature rise at high ambient temperatures).

[0041] Online real-time prediction: Obtain the latest ambient temperature and real-time load current, calculate their historical moving average, and combine them with the current time characteristics to form a real-time input feature vector; input the real-time input feature vector into a trained machine learning model to obtain the predicted temperature of each monitoring point along the entire cable under the current operating conditions, and calculate the temperature warning threshold based on a dynamic adjustment formula.

[0042] The expression for the dynamic adjustment formula is as follows: T threshold (i) = T base (i) + α * σ(i) + β; In the formula, T threshold (i) is the warning threshold at point i, T base (i) represents the predicted temperature at point i; α is the sensitivity coefficient (e.g., set to 2 or 3) used to control the confidence interval width. σ(i) is the standard deviation of the uncertainty predicted by the model at point i (a measure of the uncertainty of this spatial point in the model's prediction, i.e., the standard deviation of the prediction error at this point on the validation set); β is the safety margin constant, specifically a fixed temperature increment (e.g., 2℃ or 5℃) used to compensate for additional risks not considered by the model or to provide additional safety buffers.

[0043] The multi-level early warning and false alarm suppression module is used to monitor and issue early warnings of fire hazards by executing early warning logic rules, wherein the early warning logic rules include: The system determines single-parameter anomalies based on dynamic early warning thresholds. When an abnormal state persists beyond a short observation window, a low-level early warning is triggered. The single parameters include temperature, dielectric loss factor, and load current. Single-parameter anomalies are determined by their respective early warning thresholds. When the threshold is exceeded, the cable condition represented by the parameter becomes abnormal, and the system will initiate a low-level early warning and subsequent analysis processes (such as timing and correlation analysis).

[0044] Based on the low-level warning start timer, parameter correlation analysis is triggered when the abnormal state continues for longer than the long-term confirmation window; otherwise, the low-level warning is lifted. It can be understood that by setting a time window mechanism of short-term observation window and long-term confirmation window, instantaneous interference and continuous abnormality are distinguished, thereby eliminating instantaneous abnormality and issuing warnings for continuous abnormality.

[0045] The parameter correlation analysis includes: Based on preset physical association rules, the system determines whether the correlation between the three parameters of temperature, dielectric loss factor, and load current is successful. If the correlation is successful, a high-level warning is triggered; otherwise, a medium-level warning is triggered. High-level warnings include hazard diagnosis results, BIM 3D positioning information, and hazard patterns. Medium-level warnings include single-parameter anomaly monitoring reminders, such as "The cable temperature at point X is consistently high; inspection is recommended."

[0046] The preset of the physical association rules includes the following steps: Construction of transaction dataset: Based on prior knowledge, define the pattern characteristics of each hidden danger on the monitoring parameters, including overload, insulation aging, poor contact and external heat source influence, etc. Collect parameters of hidden danger status, combine the parameter status snapshots within the time window into the premise item set of event transactions. The parameter status snapshots include the pattern characteristics (such as high temperature rise rate) transformed from the monitoring parameters (temperature, current, dielectric loss), and label the hidden danger type as the conclusion item for each event transaction to obtain the transaction database.

[0047] Understandably, each potential hazard exhibits a pattern of characteristics in the monitoring parameters, such as: overload hazard: current continuously exceeds the safe current carrying capacity, temperature is positively correlated with the square of current and the temperature rise rate is fast (minute-level response); insulation aging hazard: dielectric loss continuously increases and exceeds the threshold, temperature is positively correlated with dielectric loss value; poor contact (such as loose joints): local hot spot temperature is much higher than adjacent sections, current is not obviously abnormal; external heat source influence: there is a heat source near the cable, causing the cable temperature to rise, but the current and dielectric loss are normal.

[0048] Association rule mining: Setting minimum support and minimum confidence, an association rule mining algorithm (such as the Apriori algorithm or FP-growth algorithm) is used to scan the entire transaction database to find all frequently occurring itemsets. "Frequent" here means that the itemset (e.g., {load rate > 130%, vulnerability type - overload}) appears more than the preset minimum support threshold in all historical events. All association rules are generated from each frequent itemset.

[0049] Rule filtering: Strongly correlated rules are filtered out from all association rules by calculating confidence scores, and rules that violate physical mechanisms are removed through review by domain experts.

[0050] This embodiment also provides a method for monitoring and locating fire hazards in building low-voltage cables based on multi-mode monitoring, such as... Figure 4 The above includes the following steps: S1. Synchronously and in real time acquire the temperature distribution data, insulation dielectric loss factor, real-time load current data, and time domain reflection signal of the entire cable. S2. Based on the real-time ambient temperature and load current, dynamically calculate the adaptive temperature warning threshold at each point along the entire cable under the current operating conditions using a pre-trained machine learning model. S3. Compare the real-time monitored temperature, dielectric loss factor and load current data with the corresponding early warning thresholds to determine the abnormal state of a single parameter. S4. If an abnormal state occurs, start the timer; if the abnormal state continues for longer than the short observation window, analyze the correlation between the three parameters of temperature, dielectric loss factor and current, and determine whether the combination conforms to the preset physical correlation rules in order to identify the type of hidden danger. S5. If only a single parameter is abnormal, a medium-level warning is triggered; if multiple parameters are associated and matched successfully, and this state continues for more than the long-term confirmation window, a fire hazard is confirmed and a high-level warning is triggered. S6. Output the hazard diagnosis type. At the same time, based on the electrical distance of the abnormal point relative to the preset feature reflector, the path matching and spatial interpolation calculation are performed in the building information model through the BIM fusion positioning module, and the three-dimensional spatial coordinates of the hazard point are output.

[0051] This invention first achieves continuous spatial sensing of the cable's temperature along its entire length by pre-embedding distributed optical fibers based on the Raman scattering principle. Simultaneously, it collects and integrates three key parameters—temperature, dielectric loss, and real-time load current—to construct a multi-dimensional monitoring system. This not only captures abnormal temperature rises indicating overload but also detects early, latent insulation aging trends through continuous changes in dielectric loss, enabling earlier and more comprehensive early warnings of fire hazards. Secondly, it deeply integrates Time Domain Reflectometry (TDR) positioning technology with Building Information Modeling (BIM). By pre-setting feature reflectors at cable joints and bends, the BIM-integrated positioning module intelligently matches and spatially interpolates the electrical distance obtained from TDR positioning with the cable's digital laying path in the 3D building model. Ultimately, it transforms the abstract "cable length" into intuitive "building spatial coordinates," solving the problems of unintuitive linear alarm positioning and difficult location, greatly improving operation and maintenance efficiency. Finally, an adaptive dynamic threshold analysis based on a machine learning model is introduced, which can dynamically adjust the warning threshold according to the real-time ambient temperature and load current. This allows the system to "understand" the normal performance of the cable under the current operating conditions, avoiding false alarms caused by fixed thresholds due to environmental or load fluctuations. At the same time, a multi-level warning logic based on multi-parameter association rules and time window verification is established. The system only triggers a high-level alarm when the three parameters of temperature, dielectric loss, and current meet the preset physical logic (such as overload and aging models) and the abnormal state continues for more than the confirmation time window. It also outputs the diagnosis of the hidden danger type, thereby filtering out instantaneous interference and ensuring the accuracy and guidance of the alarm.

[0052] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A building low-voltage cable fire hazard monitoring and location system based on multi-mode monitoring, characterized in that: It includes a multi-mode monitoring unit and an intelligent early warning host, wherein the intelligent early warning host is communicatively connected to the multi-mode monitoring unit; The multi-mode monitoring unit is embedded in the cable to synchronously collect multi-dimensional state parameters of the cable, including temperature distribution data, dielectric loss data, real-time load current data, and time-domain reflection signal. The intelligent early warning host is used to receive and process data from the multi-mode monitoring unit, and output graded early warning information containing the type and location of hidden dangers through adaptive dynamic threshold analysis, multi-parameter correlation diagnosis and BIM fusion positioning. The intelligent early warning host includes a multi-level early warning and false alarm suppression module. This module is used to monitor and issue early warnings of fire hazards by executing early warning logic rules, which include: The system judges single-parameter anomalies based on dynamic early warning thresholds. When the abnormal state continues to exceed the short-term observation window, a low-level early warning is triggered. The single parameters include temperature, dielectric loss factor and load current. Based on the low-level warning start timer, if the abnormal state continues for longer than the long confirmation window, parameter correlation analysis is triggered; otherwise, the low-level warning is lifted. The parameter correlation analysis includes: Based on preset physical association rules, it is determined whether the correlation between the three parameters of temperature, dielectric loss factor and load current is successful. If the correlation is successful, an advanced warning is triggered; otherwise, an intermediate warning is triggered. Advanced warnings include hazard diagnosis results, BIM 3D positioning information and hazard patterns, while intermediate warnings include single-parameter anomaly monitoring reminders.

2. The building low-voltage cable fire hazard monitoring and location system based on multi-mode monitoring according to claim 1, characterized in that: The multi-mode monitoring unit includes a distributed temperature sensing module and an insulation status monitoring module; The distributed temperature sensing module is used to acquire temperature distribution data in the continuous space of the cable based on the Raman scattering principle; the insulation condition monitoring module is used to collect the dielectric loss value of the cable insulation layer in real time based on the dielectric loss factor measurement technology.

3. The building low-voltage cable fire hazard monitoring and location system based on multi-mode monitoring according to claim 2, characterized in that: The multi-mode monitoring unit also includes an electrical load monitoring module and a positioning signal acquisition module; The electrical load monitoring module is used to monitor the real-time load current of the cable; the positioning signal acquisition module, based on time-domain reflection technology and a feature reflector set at a preset position on the cable, is used to inject positioning signals into the cable to acquire time-domain reflection signals and obtain a description of the location of the abnormal point.

4. The building low-voltage cable fire hazard monitoring and location system based on multi-mode monitoring according to claim 3, characterized in that: The process of obtaining the location description of the anomaly point includes the following steps: Standardized feature reflectors are pre-positioned at predetermined locations on the cable to establish an absolute position reference system; nanosecond-level fast-edge pulses are injected into the cable through a high-frequency coupling circuit, and its reflected signals are received synchronously. The reflected signal is amplified and then digitally sampled and combined by an ultra-high-speed analog-to-digital converter to form a complete time-domain reflection waveform. All characteristic reflection peaks are identified and matched on the waveform, and the actual physical location of each characteristic reflection peak is determined according to the location of the characteristic reflector. The time difference between the abnormal reflection peak and the nearest upstream and downstream characteristic reflection peaks is measured to obtain the distance of the abnormal point from the upstream and downstream characteristic reflectors.

5. The building low-voltage cable fire hazard monitoring and location system based on multi-mode monitoring according to claim 1, characterized in that: The intelligent early warning host also includes a BIM fusion positioning module and an adaptive threshold analysis module; The BIM fusion positioning module is used to store a building information model containing the digital cable laying path, and to match the abnormal point location description obtained by the positioning signal acquisition module with the path in the building information model to obtain the visualized location information of the potential hazard point in the three-dimensional space of the building.

6. The building low-voltage cable fire hazard monitoring and location system based on multi-mode monitoring according to claim 5, characterized in that: The process of obtaining the visualized location information of potential hazards in the three-dimensional space of a building includes the following steps: Based on cable laying drawings or laser scanning data, a cable path model is created in the building information model. The three-dimensional coordinates of the pre-set feature reflectors in the cable path model are bound to their unique identification codes. Based on the known electrical distance of the feature reflectors, a "physical location-electrical distance" mapping table is established. Receive the description of the location of the anomaly point, retrieve the three-dimensional coordinates of the upstream and downstream feature reflectors in the cable path model from the mapping table based on the identification code of the feature reflector in the description, and extract the cable path segment connecting the upstream and downstream feature reflectors. The total spatial length between upstream and downstream feature reflectors in the cable path model is obtained by accumulating the Euclidean distances between each node of the cable path segment. The normalization scaling factor of the anomaly point in the interval between the upstream and downstream feature reflectors is calculated based on the electrical distance between the anomaly point and the feature reflector. The selected feature reflector starts by accumulating segmented distances along the node sequence. When the accumulated distance reaches or exceeds the target distance, the node interval is located. Linear interpolation is performed on the straight line segment formed by the nodes to calculate the three-dimensional coordinates of the potential hazard point. The calculated three-dimensional coordinates are then converted to the view coordinate system of the BIM visualization engine and highlighted on the BIM three-dimensional model interface.

7. The building low-voltage cable fire hazard monitoring and location system based on multi-mode monitoring according to claim 6, characterized in that: The adaptive threshold analysis module has a built-in machine learning model trained with historical normal operation data, which is used to dynamically obtain the temperature warning threshold based on real-time ambient temperature and load current. The dynamically acquired temperature warning threshold includes the following steps: Offline model training: A large number of training data samples are collected synchronously during the normal operation of the cable; each sample includes ambient temperature, real-time load current, historical moving average, time features, and actual temperature at each monitoring point; after cleaning and normalizing the data samples, a dataset is constructed according to spatial points; gradient boosting regression tree or random forest is selected for training, and the training objective is set to minimize the error between the model's predicted temperature and the measured normal temperature; after training, the model performance is verified using data that was not used in the training. Online real-time prediction: Obtain the latest ambient temperature and real-time load current, calculate their historical moving average, and combine them with the current time characteristics to form a real-time input feature vector; input the real-time input feature vector into a trained machine learning model to obtain the predicted temperature of each monitoring point along the entire cable under the current operating conditions, and calculate the temperature warning threshold based on a dynamic adjustment formula.

8. The building low-voltage cable fire hazard monitoring and location system based on multi-mode monitoring according to claim 7, characterized in that: The expression for the dynamic adjustment formula is: T threshold (i) = T base (i) + a * σ(i) + b; In the formula, T threshold (i) is the warning threshold at point i, T base (i) represents the predicted temperature at point i, α is the sensitivity coefficient, σ(i) is the standard deviation of the uncertainty in the model's prediction at point i, and β is the safety margin constant.

9. The building low-voltage cable fire hazard monitoring and location system based on multi-mode monitoring according to claim 1, characterized in that: The preset of the physical association rules includes the following steps: Construction of transaction dataset: Based on prior knowledge, define the pattern characteristics of each hidden danger on the monitoring parameters, collect the parameters of the hidden danger status, combine the parameter status snapshots within the time window into the premise item set of the event transaction, the parameter status snapshot includes the pattern characteristics transformed by the monitoring parameters, and label the hidden danger type as the conclusion item for each event transaction to obtain the transaction database; Association rule mining: Set minimum support and minimum confidence, use association rule mining algorithms to scan the entire transaction database, find all frequently occurring itemsets, and generate all association rules from each frequent itemset; Rule filtering: Strongly correlated rules are filtered out from all association rules by calculating confidence scores, and rules that violate physical mechanisms are removed through review by domain experts.

10. A method for monitoring and locating fire hazards in building low-voltage cables based on multi-mode monitoring, applied to a building low-voltage cable fire hazard monitoring and location system based on multi-mode monitoring as described in any one of claims 1-9, characterized in that: Includes the following steps: S1. Synchronously and in real time acquire the temperature distribution data, insulation dielectric loss factor, real-time load current data, and time domain reflection signal of the entire cable. S2. Based on the real-time ambient temperature and load current, dynamically calculate the adaptive temperature warning threshold at each point along the entire cable under the current operating conditions using a pre-trained machine learning model. S3. Compare the real-time monitored temperature, dielectric loss factor and load current data with the corresponding early warning thresholds to determine the abnormal state of a single parameter. S4. If an abnormal state occurs, start the timer; if the abnormal state continues for longer than the short observation window, analyze the correlation between the three parameters of temperature, dielectric loss factor and current, and determine whether the combination conforms to the preset physical correlation rules in order to identify the type of hidden danger. S5. If only a single parameter is abnormal, a medium-level warning will be triggered. If a multi-parameter correlation anomaly match is successful and this state continues for more than the long-term confirmation window, then a fire hazard is confirmed and an advanced warning is triggered. S6. Output the hazard diagnosis type. At the same time, based on the electrical distance of the abnormal point relative to the preset feature reflector, the path matching and spatial interpolation calculation are performed in the building information model through the BIM fusion positioning module, and the three-dimensional spatial coordinates of the hazard point are output.

Images