Electrical fire monitoring and early warning method and device based on distributed optical fibers

Abstract

The invention discloses an electrical fire monitoring and early warning method and device based on distributed optical fibers. The method comprises the following steps: S1, deploying equipment; s2, preprocessing the data; s3, feature level fusion is carried out; s4, performing decision-making level fusion; s5, AI classification and identification; and S6, grading early warning and disposal. Through deep integration of DTS and DAS sensing systems and by use of a multi-parameter space-time correlation fusion algorithm and an AI identification model, collaborative analysis and intelligent diagnosis are performed on temperature and sound wave characteristics of an early electrical fire in an electric power facility, so that early and accurate fire risk early warning and fault positioning are realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electrical safety monitoring and fire early warning, in particular, to an electrical fire monitoring and early warning method and device based on distributed optical fiber. BACKGROUND

[0002] Electrical fire is one of the common disasters in power systems, industrial facilities and building power supply and distribution systems, which is usually caused by gradual development of hidden troubles such as loose cable joints, insulation aging and partial discharge. The existing electrical fire monitoring technology mainly includes: (1) point temperature detector: using discrete temperature sensors to monitor the temperature of specific points, but it has the problems of large monitoring blind area and slow response speed, and it is difficult to capture the overheating phenomenon in hidden parts; (2) traditional smoke detector: not sensitive to early overheating of electrical equipment, and only alarms when fire forms a visible flame or generates a large amount of smoke, which lags behind in early warning; (3) infrared thermal imager: can realize non-contact temperature monitoring, but the price is expensive, and it cannot effectively monitor hidden parts and is easily affected by obstacles.

[0003] Distributed optical fiber sensing technology utilizes the characteristics of optical fiber as both a sensing element and a signal transmission medium, and realizes the measurement of temperature, strain, vibration and other physical quantities distributed along the optical fiber by analyzing the changes of backscattered light (such as Raman, Brillouin and Rayleigh scattering) generated when laser is transmitted in the optical fiber. The specific technical branches include: DTS (Distributed Temperature Sensing): based on Raman scattering effect, real-time monitoring of temperature field distribution, temperature measurement accuracy can reach ±5℃; DAS (Distributed Acoustic Sensor): based on phase-sensitive optical time domain reflectometry of Rayleigh scattering, it captures the characteristic vibration signals caused by abnormal discharge or short circuit of the cable to identify potential electrical faults.

[0004] The prior art such as Chinese patent application CN119763921A discloses an intelligent early warning cable system, which comprises: a magnesium oxide mineral insulated fireproof cable; a DAS distributed acoustic sensor connected to a single-mode transmission optical fiber unit for detecting vibration information of the magnesium oxide mineral insulated cable; a DTS distributed temperature sensor connected to a multi-mode transmission optical fiber unit for detecting temperature information of the magnesium oxide mineral insulated fireproof cable; a data analysis module connected to the DAS distributed acoustic sensor and the DTS distributed temperature sensor, and determining the risk condition according to the vibration information detected by the DAS distributed acoustic sensor and the temperature information detected by the DTS distributed temperature sensor; a data warning module connected to the data analysis module, and performing corresponding warning measures according to the risk condition determined by the data analysis module.

[0005] The prior art takes the sensing function as an inherent component of a specific cable (magnesium oxide mineral insulated fireproof cable), the adaptability is determined by the specific cable integrated therein, and the system function expansion is limited. The sensing optical fiber is pre-embedded inside during cable manufacturing, and the deployment mode is fixed, which cannot be optimized for existing specific risk points.

[0006] In addition, the current prior art DAS and DTS are relatively independent and connect different types of optical fibers, and the "data analysis module" only discloses the use of vibration and temperature information to determine risks, without disclosing specific fusion algorithms or intelligent diagnosis methods. At the same time, there is no clear grading mechanism: only "corresponding warning measures are taken", and the precision, classification and operability of the warning are not specifically disclosed. SUMMARY

[0007] To solve the above problems, the present application provides an electrical fire monitoring and warning method and device based on distributed optical fiber, which integrates DTS and DAS sensing systems in depth, and uses multi-parameter space-time correlation fusion algorithm and AI recognition model to cooperatively analyze and intelligently diagnose the temperature and acoustic characteristics of early electrical fires in power facilities, thereby realizing early and accurate fire risk warning and fault positioning.

[0008] An electrical fire monitoring and warning method based on distributed optical fiber, comprising the following steps: Step S1: device deployment: install the integrated DTS / DAS monitoring host in the cable tunnel entrance equipment room, along the cable supports on both sides of the cable tunnel, and lay the special high-temperature resistant armored sensing optical cable parallel to the original power cable, and adopt a continuous wave-shaped path; Step S2: data preprocessing, data denoising, data alignment and data normalization processing are performed on the DTS and DAS original data; Step S3: feature-level fusion: extract core parameters representing abnormalities from the preprocessed data, which are divided into two types of features: temperature and vibration; Step S4: decision-level fusion: based on the feature parameters, calculate the fire risk index through a weighted decision model; Step S5: AI classification and identification, build a classification model based on machine learning to accurately identify the type of anomaly and the type of disaster: Step S6: graded warning and disposal: according to the risk index and the degree of feature anomaly, trigger different levels of warning and disposal measures.

[0009] Preferably, the extraction of temperature anomaly features in step S3 includes calculating the temperature rise rate, calculating the temperature gradient, and extracting the absolute temperature value.

[0010] Preferably, the extraction of vibration anomaly features in step S3 includes calculating the energy value of a specific frequency band, counting the frequency of vibration events, and calculating the duration of vibration signals.

[0011] Preferably, step S4 includes the following steps: Step S4.1, determining the weighting coefficients; Step S4.2, calculating the spatiotemporal correlation score and analyzing the spatiotemporal matching degree between temperature anomalies and vibration anomalies; Step S4.3, calculating the risk index: Risk Index = ×Temperature rating+ ×Vibration Score+ × Spatiotemporal correlation score, It is the weight of temperature rating. It is the vibration score weight. It is the vacuum correlation score weight.

[0012] Preferably, step S5 includes the following steps: Step S5.1: Construct a sample library, collect historical abnormal event data, and label the corresponding feature parameters and disaster types; Step S5.2: Establish a feature pattern library; Step S5.3: Train a classification model using machine learning algorithms, train the model with labeled samples, and optimize the model parameters until the recognition accuracy meets the requirements; Step S5.4: Predict the disaster type, input the real-time features into the trained model, and output the disaster type and recognition confidence.

[0013] Preferably, the warnings in step S6 include Level 1 warnings, Level 2 warnings, and Level 3 warnings.

[0014] An electrical fire monitoring and early warning device for the above-mentioned electrical fire monitoring and early warning method includes a sensing optical cable assembly, a signal acquisition unit, a data analysis module, and a data early warning module; The sensing optical cable assembly uses a special high-temperature resistant double-layer armored sensing optical cable, which is laid in parallel along the monitored power cable or arranged around key nodes of electrical equipment. The signal acquisition unit includes a DAS distributed acoustic wave sensor and a DTS distributed temperature sensor; The data analysis module is equipped with a multi-parameter fusion analysis algorithm and an AI classifier to achieve simultaneous analysis and feature extraction of temperature and vibration signals. The data early warning module includes a multi-level early warning mechanism and a human-computer interaction interface.

[0015] Preferably, the sensing optical cable core is made of high-temperature resistant fiber, and the sheath adopts a double-layer armored structure, with the inner layer being a stainless steel flexible tube and the outer layer being a polytetrafluoroethylene sheath.

[0016] Preferably, an anti-electromagnetic interference shielding layer is embedded in the sensing optical cable.

[0017] Beneficial effects Compared with the prior art, the present invention has the following significant advantages: (1) Universality and low-cost deployment: It can be applied to various existing power facilities without the need to replace them with specific cables, which greatly reduces the cost and difficulty of intelligent transformation; (2) Early warning and accurate diagnosis: Through the spatiotemporal correlation analysis of temperature and vibration, early hidden dangers such as partial discharge can be identified, and the fault type can be distinguished through AI model, realizing the leap from "abnormal alarm" to "fault diagnosis"; (3) High reliability and low false alarm: Multi-parameter fusion and three-level early warning mechanism effectively filter environmental interference (such as a single vibration may come from construction), which greatly reduces the false alarm rate of the system; (4) Flexible and targeted deployment: The optical cable laying method can be optimized according to the actual risk distribution on site, and the monitoring sensitivity of key parts (such as cable joints) can be improved; (5) Strong environmental adaptability: The special sensor optical cable design makes it stable in harsh electrical environments such as strong electromagnetic fields and high temperatures. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating an electrical fire monitoring and early warning method based on distributed optical fiber according to the present invention; Figure 2 yes Figure 1 The flowchart shown is of the multi-parameter fusion analysis algorithm used in the electrical fire monitoring and early warning method. Figure 3 This is a framework diagram illustrating an electrical fire monitoring and early warning device based on distributed optical fiber according to the present invention; Figure 4 yes Figure 3 The diagram shows the structure of the sensing optical cable used in the electrical fire monitoring and early warning device. Detailed Implementation

[0019] The embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative and not restrictive.

[0020] Figure 1 A flowchart illustrating an embodiment of the electrical fire monitoring and early warning method based on distributed optical fiber according to the present invention is shown. Electrical... Figure 1 As shown, the electrical fire monitoring and early warning method based on distributed optical fiber provided in this embodiment includes the following steps: Step S1: Equipment Deployment: Install the integrated DTS / DAS monitoring host in the equipment room at the entrance of the cable tunnel. Along the cable supports on both sides of the cable tunnel, lay the special high-temperature armored sensing optical cable parallel to and close to the original power cable, and use a continuous wavy path to increase contact. Step S2: Data preprocessing, performing data noise reduction, data alignment and data normalization on the raw DTS and DAS data; Data denoising includes removing environmental noise from raw temperature data from distributed fiber optic thermometry (DTS) and raw vibration data from distributed acoustic sensors (DAS) using filtering algorithms (such as moving average filtering and wavelet denoising); data alignment involves adjusting the denoised DTS and DAS data to the same length and sampling frequency based on timestamps or spatial location information to ensure spatiotemporal matching; and data normalization involves using standardization (mean 0, variance 1) or normalization (0-1 interval) methods to unify the dimensions of temperature and vibration data and eliminate the impact of numerical range differences on the analysis.

[0021] Step S3: Feature-level fusion: Extract the core parameters representing anomalies from the preprocessed data, which are divided into two categories: temperature and vibration features; Among them, the extraction of temperature anomaly features includes calculating the temperature rise rate (the change in temperature per unit time (°C / s), calculating the temperature gradient (the temperature difference between adjacent measuring points in the spatial dimension (°C / m), and extracting the absolute temperature value (the highest temperature or the temperature of key measuring points during the monitoring period); the extraction of vibration anomaly features includes (1) calculating the energy value of a specific frequency band (e.g., 12kHz), performing frequency domain analysis on the vibration signal, and extracting the energy proportion of the target frequency band; (2) counting the frequency of vibration events, setting a vibration threshold, and counting the number of abnormal vibration events exceeding the threshold; (3) calculating the duration of the vibration signal, the duration of a single abnormal vibration event, and the cumulative duration; feature standardization, mapping all extracted feature parameters to a scoring range of 0-100 points, which facilitates subsequent weighted calculation.

[0022] Step S4: Decision-level fusion: Calculate the fire risk index based on feature parameters using a weighted decision model; These include: Step S4.1: Determine the weighting coefficients: Based on the type of electrical equipment (e.g., cables, switches, transformers), set... (Temperature score weighting) (Vibration score weighting) (Spatiotemporal correlation scoring weight); Step S4.2: Calculate the spatiotemporal correlation score and analyze the spatiotemporal matching degree between temperature anomalies and vibration anomalies (such as the overlap of anomaly locations and the time difference of anomaly occurrence), and map it to a score of 0-100. Step S4.3, Calculate the risk index: Risk index = ×Temperature rating+ ×Vibration Score+ × Spatiotemporal correlation score.

[0023] The final risk index is normalized to 0-100 points, with higher scores indicating higher risk.

[0024] Step S5: AI classification and recognition, building a classification model based on machine learning to accurately identify anomaly types and disaster categories: These include: Step S5.1: Build a sample library, collect historical abnormal event data, and label the corresponding feature parameters and disaster types (such as eight categories including ice cover, wildfire, and dancing). Step S5.2: Establish a feature pattern library, including a temperature anomaly pattern library (summarizing typical temperature characteristics such as cable joint overheating, load overload, and insulation aging), an acoustic anomaly pattern library (summarizing typical vibration / acoustic characteristics such as partial discharge, arc flash, and cable breakdown), and a multi-parameter correlation pattern library (establishing a mapping relationship between temperature-vibration spatiotemporal correlation characteristics and fault types). Step S5.3: The classification model is trained using machine learning algorithms such as random forest and neural network. The model is trained with labeled samples and the model parameters are optimized until the recognition accuracy meets the requirements (covering more than 90% of risk scenarios). Step S5.4: Disaster type prediction. Input the real-time features into the trained model and output the disaster type and recognition confidence. Step S6: Tiered early warning and response: Based on the risk index and the degree of abnormality of the characteristics, trigger different levels of early warning and response measures.

[0025] exist Figure 2 The details are shown in the middle. Figure 1 A flowchart of the multi-parameter fusion analysis algorithm used in the electrical fire monitoring and early warning method. Regarding the graded early warning and response in step S6, the early warning levels include: Level 1 (Observation Level): Triggering condition: slight anomaly of a single parameter (e.g., slow temperature increase, isolated vibration event), risk index < 30 points; Response measures: the system automatically records abnormal data, marks observation points, and continuously tracks data changes; Level 2 (Alert Level): Triggering condition: significant anomaly of a single parameter (e.g., temperature exceeding the threshold, vibration frequency doubling) or slight anomaly of two parameters, risk index 30-70 points; Response measures: the system issues an audible / visual / SMS alarm, notifying maintenance personnel to conduct on-site verification and strengthen inspections; Level 3 (Action Level): Triggering condition: significant anomalies in both temperature and vibration parameters with high spatiotemporal correlation, risk index ≥ 70 points; Response measures: triggering the highest level alarm, linking the fire protection system (e.g., activating fire extinguishing devices, cutting off faulty circuits), and immediately notifying emergency response personnel.

[0026] Figure 3 This is a framework diagram of an electrical fire monitoring and early warning device based on distributed optical fiber according to the present invention. Figure 3As shown, the electrical fire monitoring and early warning device based on distributed optical fiber provided in this embodiment includes a sensing optical cable assembly, a signal acquisition unit, a data analysis module, and a data early warning module. The sensing optical cable assembly uses a special high-temperature resistant double-armored sensing optical cable, laid parallel to the monitored power cable or arranged around key nodes of electrical equipment (such as joints and switch contacts), to sense the cable discharge sound vibration spectrum and the temperature signals of the cable and its environment, and transmit them to the optical fiber. The signal acquisition unit includes a DAS distributed acoustic wave sensor and a DTS distributed temperature sensor, used to collect the information sensed by the sensing optical cable assembly. The data analysis module is equipped with a multi-parameter fusion analysis algorithm and an AI classifier, used to analyze signal spectrum characteristics and temperature anomaly characteristics, and combine anomaly information to analyze and locate the distance information of the anomaly point, achieving synchronous analysis and feature extraction of temperature and vibration signals. The data early warning module includes a multi-level early warning mechanism and a human-machine interface, used to determine whether the abnormal fault information characteristics meet the early warning alarm process and report it.

[0027] It should be noted that, although the distributed optical fiber-based electrical fire monitoring and early warning device provided in this embodiment can be used for implementation Figures 1-2 The electrical fire monitoring and early warning methods shown include, for example, Figures 1-2 The sensing cable in step S1 of the method shown can be the sensing optical cable assembly, the data acquisition in step S2 uses the signal acquisition unit, the data analysis and machine learning process in steps S3-S5 uses the data analysis module (AI classifier), and the graded early warning and handling in step S6 uses the data early warning module. However, those skilled in the art will understand that the electrical fire monitoring and early warning device can also implement other early warning methods based on the distributed optical fiber mode.

[0028] Figure 4 yes Figure 3 The diagram shows the structure of the sensing optical cable used in the electrical fire monitoring and early warning device.

[0029] The sensing optical cable used in this invention can be a commercially available ordinary sensing optical cable, but it is best to meet certain design requirements, such as using a special high-temperature resistant armored sensing optical cable, and laying the special high-temperature resistant armored sensing optical cable parallel and close to the existing power cable along the cable supports on both sides of the cable tunnel (e.g.) Figure 4 As shown, the sensing fiber bundle is laid parallel to and close to the existing power cable C through cable connector J, and a continuous wavy path (or a continuous spiral path) is used to increase the contact between the two.

[0030] In a preferred embodiment, the sensing optical cable is specially designed to meet the demanding environmental requirements of electrical fire monitoring: Fiber core: High-temperature resistant fiber core is used, and the coating layer is specially treated to ensure long-term stable operation in an environment of 200℃.

[0031] Sheath: Double-layer armored structure, with an inner stainless steel hose and an outer polytetrafluoroethylene sheath, providing good mechanical protection and corrosion resistance.

[0032] Anti-interference design: An anti-electromagnetic interference shielding layer is embedded in the optical cable to ensure that signal acquisition is not affected in strong electromagnetic field environments.

[0033] This design enables the pipeline fatigue monitoring system to achieve stable monitoring in high-temperature environments above 400°C.

[0034] The preferred embodiment of the present invention will be described in detail below using the deployment of the system of the present invention in an existing 110kV cable tunnel as an example. The cable tunnel already contains conventional power cables, and the purpose of this study is to install the fire monitoring system of the present invention on top of these cables.

[0035] Step 1: System Deployment.

[0036] The integrated DTS / DAS monitoring host described in this invention is selected and installed in the equipment room at the entrance of the cable tunnel.

[0037] Along the cable supports on both sides of the cable tunnel, special high-temperature resistant armored sensing optical cables are laid parallel and close to the existing power cables, and a continuous wavy path is used to increase contact.

[0038] Connect the laid optical cable to the monitoring host and test the connection.

[0039] Step 2: System Operation and Early Warning Examples The system continues to operate. One day, the data analysis module detected the following synchronization signal: DTS data: The temperature at connector No. 15 was detected to rise continuously from 45°C at a rate of 2°C / minute.

[0040] DAS data: Continuous vibration signals with frequencies concentrated in the 1-2kHz range were detected synchronously at the same location, and their spectral characteristics matched those of the partial discharge model library.

[0041] Fusion analysis: The algorithm confirmed that the temperature anomaly and vibration anomaly occurred synchronously in time and were completely located in space. After calculation by the AI ​​classifier, the fault was determined to be "poor contact caused by loose cable joints and accompanied by discharge", which is of high risk level.

[0042] Warning Output: The system immediately triggers a Level 3 (Action-level) warning, displaying an alarm on the monitoring center screen, precisely locating "Joint No. 15 in the eastern section of the cable tunnel," and showing the diagnostic results. Simultaneously, maintenance personnel are notified via SMS.

[0043] Step 3: Processing Verification Upon receiving the alert, maintenance personnel quickly arrived at the scene and inspected connector No. 15, finding loose crimp bolts and burn marks on the connector. After tightening, both temperature and vibration signals returned to normal. This entire process, from system detection of the anomaly to on-site handling, was completed before the fire occurred, demonstrating the effectiveness of this invention.

[0044] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for monitoring and early warning of electrical fires based on distributed optical fibers, characterized in that: Includes the following steps: Step S1: Equipment deployment: Install the integrated DTS / DAS monitoring host in the equipment room at the entrance of the cable tunnel. Along the cable supports on both sides of the cable tunnel, lay the special high-temperature armored sensing optical cable parallel to and close to the original power cable, using a continuous wavy path. Step S2: Data preprocessing, performing data noise reduction, data alignment and data normalization on the raw DTS and DAS data; Step S3: Feature-level fusion: Extract the core parameters representing anomalies from the preprocessed data, which are divided into two categories: temperature and vibration features; Step S4: Decision-level fusion: Calculate the fire risk index based on feature parameters using a weighted decision model; Step S5: AI classification and recognition, building a classification model based on machine learning to accurately identify anomaly types and disaster categories: Step S6: Tiered early warning and response: Based on the risk index and the degree of abnormality of the characteristics, trigger different levels of early warning and response measures.

2. The electrical fire monitoring and early warning method according to claim 1, characterized in that: The extraction of temperature anomaly features in step S3 includes calculating the temperature rise rate, calculating the temperature gradient, and extracting the absolute temperature value.

3. The electrical fire monitoring and early warning method according to claim 1, characterized in that: The extraction of vibration anomaly features in step S3 includes calculating the energy value of a specific frequency band, counting the frequency of vibration events, and calculating the duration of the vibration signal.

4. The electrical fire monitoring and early warning method according to claim 1, characterized in that: Step S4 includes the following steps: Step S4.1, determining the weighting coefficients; Step S4.2, calculating the spatiotemporal correlation score and analyzing the spatiotemporal matching degree between temperature anomalies and vibration anomalies; Step S4.3, calculating the risk index: Risk Index = ×Temperature rating+ ×Vibration Score+ × Spatiotemporal correlation score, It is the weight of temperature rating. It is the vibration score weight. It is the vacuum correlation score weight.

5. The electrical fire monitoring and early warning method according to claim 1, characterized in that: Step S5 includes the following steps: Step S5.1: Construct a sample library, collect historical abnormal event data, and label the corresponding feature parameters and disaster types; Step S5.2: Establish a feature pattern library; Step S5.3: Train a classification model using machine learning algorithms, train the model with labeled samples, and optimize the model parameters until the recognition accuracy meets the requirements; Step S5.4: Predict the disaster type, input the real-time features into the trained model, and output the disaster type and recognition confidence.

6. The electrical fire monitoring and early warning method according to claim 1, characterized in that: The warnings in step S6 include Level 1 warnings, Level 2 warnings, and Level 3 warnings.

7. An electrical fire monitoring and early warning device for use in the electrical fire monitoring and early warning method according to any one of claims 1 to 6, characterized in that: It includes a sensing optical cable assembly, a signal acquisition unit, a data analysis module, and a data early warning module.

8. The electrical fire monitoring and early warning device according to claim 7, characterized in that: The sensing optical cable assembly uses a special high-temperature resistant double-layer armored sensing optical cable, which is laid in parallel along the monitored power cable or arranged around key nodes of the electrical equipment. The signal acquisition unit includes a DAS distributed acoustic wave sensor and a DTS distributed temperature sensor. The data analysis module is equipped with a multi-parameter fusion analysis algorithm and an AI classifier to achieve simultaneous analysis and feature extraction of temperature and vibration signals. The data early warning module includes a multi-level early warning mechanism and a human-computer interaction interface.

9. The electrical fire monitoring and early warning device according to claim 8, characterized in that: The fiber core of the sensing optical cable is made of high-temperature resistant fiber core, and the sheath adopts a double-layer armor structure, with the inner layer being a stainless steel flexible tube and the outer layer being a polytetrafluoroethylene sheath.

10. The electrical fire monitoring and early warning device according to claim 8, characterized in that: An anti-electromagnetic interference shielding layer is embedded in the sensing optical cable.

Images