A distributed optical fiber based compressed air energy storage underground chamber air tightness monitoring system

Abstract

The present application relates to compressed air energy storage safety monitoring technical field, specifically to a kind of compressed air energy storage underground chamber air tightness monitoring system based on distributed optical fiber, including distributed optical fiber sensing unit, signal transmitting and receiving unit, data processing and early warning unit;Distributed optical fiber sensing unit is made of multiple high-pressure resistant sensing optical fibers, optical fiber is laid in grid on the surface of chamber, forms full-coverage monitoring area, each area is provided with only identification code;Signal transmitting and receiving unit includes DAS host and DTS host, DAS host emits narrow pulse laser to optical fiber and receives Rayleigh scattering signal, DTS host real-time acquisition optical fiber axial temperature distribution data;Data processing and early warning unit is built-in signal analysis module, leakage identification module, positioning calculation module and early warning trigger module, for the synchronous processing and intelligent response of multi-source sensing data.The system realizes chamber full-space dead angle-free monitoring, solves the problem of insufficient coverage of point monitoring.

Description

Technical Field

[0001] This invention relates to the field of compressed air energy storage safety monitoring technology, specifically to a compressed air energy storage underground chamber air tightness monitoring system based on distributed optical fiber. Background Technology

[0002] Compressed air energy storage, as one of the key technologies for large-scale energy storage, plays an important role in peak shaving and valley filling of the power system and grid integration and consumption of renewable energy. One of its core facilities, the underground chamber, needs to withstand working pressures of up to 18 MPa or even higher to ensure the efficient and safe operation of the energy storage system. Under this high-pressure environment, the airtightness of the underground chamber becomes the core indicator for ensuring the overall efficiency and operational safety of the system. Any tiny gas leak may lead to a decrease in energy storage efficiency or even cause a safety accident. Therefore, real-time and accurate monitoring and early warning of the airtightness of the underground chamber are particularly important.

[0003] Currently, the main limitations of methods for monitoring the airtightness of compressed air energy storage underground chambers are as follows:

[0004] Limitations of manual inspection: Traditional manual inspection methods are difficult to achieve continuous and real-time monitoring, and are easily affected by human factors, leading to missed or false detections, and cannot meet the high safety requirements under high pressure environments.

[0005] Point sensors have limited coverage: Although point sensors can detect gas leaks at specific locations, their coverage is limited, making it difficult to monitor the entire space, which is especially evident in underground chambers with complex geometries.

[0006] The negative pressure wave method has insufficient sensitivity: As a commonly used pipeline leak detection technology, the negative pressure wave method has low sensitivity in the identification of micro-leakage, which is difficult to meet the monitoring needs of high-pressure chambers for micro-leakage.

[0007] Adaptability issues of existing distributed optical fiber technology: Although distributed optical fiber sensing technology is widely used in pipeline monitoring and other fields due to its advantages such as resistance to electromagnetic interference, high pressure resistance, and end-to-end sensing, existing technologies are mostly designed for linear pipeline structures and do not fully consider the complex spatial structure and high-pressure conditions of underground chambers. This leads to problems such as low positioning accuracy, susceptibility to environmental interference, and delayed early warning in practical applications, making it difficult to meet the high-precision and high-reliability requirements for airtightness monitoring in compressed air energy storage underground chambers. Specifically, the shortcomings of existing distributed optical fiber technology in underground chamber applications are mainly reflected in the following aspects:

[0008] Insufficient positioning accuracy: Due to the complex spatial structure of underground chambers and the limited fiber optic deployment methods, the positioning accuracy of leak points is not high, which is difficult to meet the actual engineering requirements.

[0009] Severe environmental interference: Interference signals generated by environmental factors such as geological vibration and equipment operation are easily confused with actual leakage signals, affecting the accuracy of monitoring results;

[0010] Inadequate early warning mechanisms: Existing technologies focus primarily on leak detection, but lack effective early warning mechanisms, failing to issue timely alerts in the early stages of leaks, thus increasing safety risks.

[0011] In summary, for the airtightness monitoring of compressed air energy storage underground chambers, there is an urgent need to develop a new monitoring system that can adapt to high-pressure conditions, achieve full-space coverage, high-precision positioning, and early warning. Summary of the Invention

[0012] To address the aforementioned problems, this invention provides a distributed optical fiber-based airtightness monitoring system for underground compressed air storage chambers. This system enables full-space monitoring without blind spots within the chamber, thus solving the problem of insufficient coverage in point-based monitoring.

[0013] The technical solution of the present invention is as follows:

[0014] A system for monitoring the air tightness of an underground compressed air storage chamber based on distributed optical fiber includes a distributed optical fiber sensing unit, a signal transmitting and receiving unit, and a data processing and early warning unit.

[0015] The distributed optical fiber sensing unit consists of multiple high-voltage resistant sensing optical fibers, which are arranged in a grid pattern on the surface of the chamber to form a full-coverage monitoring area. Each area has a unique identification code.

[0016] The signal transmitting and receiving unit includes a DAS host and a DTS host. The DAS host transmits narrow pulse laser to the optical fiber and receives Rayleigh scattering signals, while the DTS host collects optical fiber axial temperature distribution data in real time.

[0017] The data processing and early warning unit has built-in signal analysis, leakage identification, location calculation and early warning triggering modules for synchronous processing and intelligent response of multi-source sensor data.

[0018] Furthermore, the grid layout parameters of the sensing optical fiber are adjustable to adapt to different monitoring accuracy requirements and form a dual-end monitoring link.

[0019] Furthermore, the data processing and early warning unit has a built-in leak sound signal template library, which contains characteristic parameters of compressed air leaks, and the template library can be dynamically updated based on historical data and on-site calibration.

[0020] A method for early warning of air tightness in underground compressed air storage chambers based on distributed optical fiber, and an air tightness monitoring system for underground compressed air storage chambers based on distributed optical fiber, includes the following steps:

[0021] S1: System initialization, parameter calibration of DAS and DTS main units, acquisition of acoustic baseline signal and temperature baseline distribution under leak-free conditions in the chamber, and setting of normal fluctuation threshold range;

[0022] S2: Real-time data acquisition, synchronously acquiring Rayleigh scattering signals and temperature distribution data of the fiber array, with a sampling frequency of 2-4kHz and a data transmission delay of ≤1 second;

[0023] S3: Abnormal signal identification. Real-time data is compared with baseline data. When the acoustic signal characteristics match the template library by ≥85%, or the local temperature change is ≥0.1℃ and the duration is ≥3 seconds, it is marked as an abnormal signal.

[0024] S4: Leakage location and quantification. The abnormal area is determined by the distance code corresponding to the abnormal signal, the coordinates of the abnormal point are determined, and then the leakage amount Q is estimated based on the rate of change of temperature gradient.

[0025] S5: Tiered early warning triggering. Multiple levels of early warning can be set based on the estimated leakage amount or leakage diffusion rate. The early warning signal will simultaneously trigger audible and visual alarms and data upload.

[0026] Furthermore, in step S1, a standard database is established through "dual 168-hour" baseline acquisition, covering normal signal characteristics under different operating conditions, providing a benchmark for anomaly identification.

[0027] Furthermore, in step S3, a noise removal algorithm is used to eliminate environmental interference signals such as geological vibration and equipment operation by analyzing the frequency filtering and duration comparison of different events.

[0028] Furthermore, in step S4, determining the coordinates of the anomaly points includes: initially locating the anomaly region through the distance encoding corresponding to the anomaly signal, and accurately locating the coordinates of the anomaly points based on the intelligent localization algorithm of the CNN-LSTM fusion network.

[0029] Furthermore, the CNN-LSTM fusion network intelligent localization algorithm includes data preprocessing, CNN feature extraction, LSTM temporal processing, fully connected output, and training optimization steps.

[0030] Furthermore, in step S4, the leakage amount estimation formula is as follows:

[0031] Q = k × ΔT × A

[0032] Where: k is a correction factor related to the chamber pressure;

[0033] ΔT is the maximum temperature change value;

[0034] A represents the area of ​​the temperature anomaly region.

[0035] Furthermore, in step S5, the multi-level early warning includes four levels: when the leakage amount Q < 0.01 Furthermore, a gray warning is issued when the leakage diffusion velocity v < 0.1 m / s, and when 0.01 ≤ Q < 0.05 A yellow alert is issued when 0.1 ≤ v < 0.3 m / s, and when 0.01 ≤ Q < 0.05. An orange alert is issued when 0.1 ≤ v < 0.3 m / s, and when Q ≥ 0.05 A red alert is issued when v ≥ 0.3 m / s.

[0036] The beneficial effects of this invention are as follows:

[0037] 1. The present invention discloses an airtightness monitoring system for underground compressed air storage chambers based on distributed optical fibers. This airtightness monitoring system for underground compressed air storage chambers based on distributed optical fibers has comprehensive monitoring coverage: the grid-like optical fiber array realizes full-space monitoring without blind spots in the chamber, adapts to complex geometric structures, and solves the problem of insufficient coverage of point-based monitoring.

[0038] 2. The present invention discloses an airtightness monitoring system for underground compressed air storage chambers based on distributed optical fiber. This airtightness monitoring system for underground compressed air storage chambers based on distributed optical fiber has high identification accuracy and reliability: it integrates DAS and DTS dual-dimensional data, achieves a positioning accuracy of within 0.5 meters, and the interference elimination algorithm makes the identification accuracy ≥95%, which can effectively identify micro-leakage.

[0039] 3. The present invention discloses an airtightness monitoring system for underground compressed air storage chambers based on distributed optical fiber. This airtightness monitoring system for underground compressed air storage chambers based on distributed optical fiber has strong adaptability to operating conditions: the high-pressure resistant optical fiber and elastic buffer design are suitable for high-pressure environments above 18 MPa, and have high long-term operational stability.

[0040] 4. The present invention discloses an airtightness monitoring system for underground compressed air storage chambers based on distributed optical fiber. This system has outstanding early warning timeliness: the delay from signal acquisition to early warning triggering is ≤3 seconds, and the graded early warning mechanism provides sufficient time for emergency response and reduces safety risks. Attached Figure Description

[0041] Figure 1 This is a structural diagram of an underground compressed air storage chamber air tightness monitoring system based on distributed optical fiber, according to an embodiment of the present invention.

[0042] Figure 2 This is a schematic diagram of the fiber optic network layout of an underground compressed air storage chamber for monitoring the air tightness of a compressed air storage chamber based on distributed optical fiber, according to an embodiment of the present invention.

[0043] Figure 3This is a flowchart of an airtightness early warning method for an underground compressed air energy storage chamber based on distributed optical fiber, according to an embodiment of the present invention.

[0044] in, Figure 1 The system consists of three units: a sensing unit that adapts to the chamber, a signal unit that provides two-dimensional data acquisition, and a processing unit that enables intelligent analysis and early warning.

[0045] Figure 2 The layout of optical fibers on each surface of the chamber and the connection of each unit are shown. ① is the circumferentially distributed optical fiber, ② is the longitudinally distributed optical fiber, ③ is the grid formed by the longitudinal and circumferential optical fibers, ④ is the data acquisition host, ⑤ is the communication module, and ⑥ is the data processing and early warning system.

[0046] Figure 3 It presents the entire process from initialization to early warning triggering, reflecting the core logic of dual data fusion and hierarchical response. Detailed Implementation

[0047] The present invention will now be described in further detail with reference to specific embodiments and accompanying drawings. Similar elements in different embodiments are referred to by associated similar element reference numerals. In the following embodiments, many details are described to facilitate a better understanding of this application. However, those skilled in the art will readily recognize that some features may be omitted in different situations, or may be replaced by other elements, materials, or methods. In some cases, certain operations related to this application are not shown or described in the specification. This is to avoid obscuring the core parts of this application with excessive description. For those skilled in the art, detailed description of these related operations is not necessary; they can fully understand the related operations based on the description in the specification and general technical knowledge in the art.

[0048] Furthermore, the features, operations, or characteristics described in the specification can be combined in any suitable manner to form various embodiments. At the same time, the steps or actions in the method description can be rearranged or adjusted in a manner obvious to those skilled in the art. Therefore, the various orders in the specification and drawings are only for the clear description of a particular embodiment and do not imply a necessary order, unless otherwise stated that a particular order must be followed.

[0049] like Figure 1 The airtightness monitoring system for compressed air energy storage underground chambers shown is composed of three parts: a distributed optical fiber sensing unit, a signal transmission and reception unit, and a data processing and early warning unit.

[0050] The distributed fiber optic sensing unit uses high-voltage resistant quartz sensing fibers, which are arranged in a grid pattern to form full-space coverage. The elastic buffer layer can prevent damage to the fibers from the deformation of the chamber. The grid coding design enables accurate location of leakage points. Specifically, the distributed fiber optic sensing unit consists of multiple high-voltage resistant sensing fibers. The fibers are arranged in a grid pattern along the vault, side walls and bottom plate of the chamber to form a scale-like monitoring area. Each monitoring area is identified by a unique distance code. An elastic buffer layer is set between the fiber and the chamber wall.

[0051] The signal transmitting and receiving unit includes a DAS host and a DTS host. The DAS host transmits narrow pulse laser to the optical fiber and receives Rayleigh scattering signals. The DTS host collects optical fiber axial temperature distribution data in real time. The DAS host captures leakage sound signals based on the Rayleigh scattering principle. The DTS host achieves temperature monitoring with an accuracy of 0.1℃. The synchronous operation of the two hosts ensures the time consistency of multi-source data.

[0052] The data processing and early warning unit integrates industrial computers and intelligent algorithms. It eliminates interference through signal fusion analysis, establishes a correlation model between leakage characteristics and early warning levels, and supports local alarms and remote data uploads. It has built-in signal analysis modules, leakage identification modules, location calculation modules, and early warning triggering modules, which can realize synchronous processing and intelligent response of multi-source sensor data.

[0053] Furthermore, the mesh layout of the sensing optical fibers, such as... Figure 2 As shown, two optical fibers are used, one circumferentially along the chamber and the other longitudinally along the chamber. The chamber is divided into grids by the longitudinal and circumferential optical fibers. The grid coding rule starts from the small mileage of the chamber, first coding in the circumferential direction. Each ring starts from the arch grid and is coded clockwise, then moves towards the large mileage direction to the next ring to continue coding.

[0054] Furthermore, the grid layout parameters of the sensing optical fiber are adjustable. The pitch to chamber perimeter ratio can be set to 1:3 to 1:8 according to the monitoring accuracy requirements, and the spacing between adjacent optical fibers is 0.3-1.0 meters. Both ends of the optical fiber are connected to the ground signal cabinet to form a dual-end monitoring link.

[0055] Furthermore, the data processing and early warning unit has a built-in leak sound signal template library, which includes characteristic frequencies (50-500Hz), amplitude variation ranges and duration threshold parameters of compressed air leaks at different pressure levels. The template library can be dynamically updated through historical data and on-site calibration.

[0056] like Figure 3 As shown, an early warning method for a compressed air storage underground chamber air tightness monitoring system based on distributed optical fiber is described, with the following specific steps:

[0057] S1. System initialization: Perform parameter calibration on the DAS and DTS main units, collect acoustic baseline signals and temperature baseline distribution under leak-free conditions in the chamber, and set the normal fluctuation threshold range;

[0058] S2. Real-time data acquisition: Synchronously acquire Rayleigh scattering signals and temperature distribution data of the fiber array, with a sampling frequency of 2-4kHz and a data transmission delay of ≤1 second;

[0059] S3. Abnormal signal identification: Compare real-time data with baseline data. When the acoustic signal characteristics match the template library by ≥85%, the local temperature change is ≥0.1℃ and the duration is ≥3 seconds, it is marked as an abnormal signal.

[0060] S4. Leakage localization and quantification: The abnormal area is determined by the distance code corresponding to the abnormal signal, the coordinates of the abnormal point are determined by the intelligent localization algorithm of CNN-LSTM fusion network, and then the leakage amount (Q) is estimated based on the temperature gradient change rate.

[0061] S5. Tiered Early Warning Trigger: A three-tiered early warning system is set based on the leakage amount (Q) and leakage diffusion rate (v), with a yellow warning (0.01 ≤ Q < 0.05). Or 0.1≤v<0.3m / s), Orange alert (0.01≤Q<0.05) And 0.1≤v<0.3m / s), red alert (Q≥0.05) (or v≥0.3m / s), the warning signal simultaneously triggers an audible and visual alarm and uploads data.

[0062] Furthermore, in step S1, a standard database is established by collecting data from "dual 168-hour" baselines, covering the normal signal characteristics under different operating conditions, providing a benchmark for anomaly identification;

[0063] "168 hours": refers to a continuous 7-day time cycle (because 7 days × 24 hours / day = 168 hours); "Double": refers to two such 168-hour cycles, which are usually arranged under different operating conditions to maximize coverage of the system's possible normal operating conditions; "Double 168-hour" baseline acquisition means that, under the condition that there is no leakage in the compressed gas storage underground chamber, the monitoring system continuously collects a total of 336 hours (14 days) of acoustic and temperature data. These 14 days are divided into two phases, each lasting 7 days, and are arranged under different typical operating conditions as much as possible.

[0064] The "dual 168-hour" baseline acquisition can establish a comprehensive "normal state fingerprint database": the acoustic environment and temperature distribution inside the compressed gas storage chamber are not static; they are affected by various factors, such as: internal factors: pressure changes caused by different operating stages of energy storage (filling) and energy release (venting); external factors: geological background vibrations, diurnal and seasonal temperature differences, vibrations from other equipment operation, etc.; the long-term data acquisition of "dual 168 hours" can capture these periodic or non-periodic normal fluctuations, thereby establishing a rich database of "normal signal characteristics"; "dual 168 hours" Baseline acquisition improves the accuracy of anomaly identification: With this comprehensive baseline database covering various operating conditions, the data processing and early warning unit can more accurately distinguish between real leakage anomaly signals and normal environmental fluctuations in subsequent real-time monitoring; based on the large normal dataset, the system can statistically analyze the normal fluctuation range of acoustic signals and temperature data, thereby setting a more scientific and realistic "normal fluctuation threshold range" instead of setting a fixed value based on experience, avoiding false alarms (reporting normal fluctuations as leaks) and false alarms (failing to identify real minor leaks).

[0065] Furthermore, in step S3, a noise elimination algorithm is used, which analyzes the frequency filtering and duration comparison of different events to eliminate environmental interference signals such as geological vibration and equipment operation, with an interference signal identification accuracy of ≥90%.

[0066] Furthermore, in step S4, the leak point is located. First, a preliminary location is performed. Using the speed of light in the optical fiber (approximately 3 × 10⁸ m / s) and the echo time difference, the location of the anomaly point is calculated using the formula L = (c × t) / 2 (where L is the distance, c is the speed of light, and t is the time from laser emission to reception of the echo). Then, the coordinates of the anomaly point are further determined using the CNN-LSTM fusion network intelligent localization algorithm.

[0067] Furthermore, in step S4, the leakage estimation formula is: Q=k×ΔT×A, where k is a correction coefficient related to the chamber pressure (0.8-1.2), ΔT is the maximum temperature change value, and A is the area of ​​the temperature anomaly region.

[0068] Furthermore, the specific steps of the CNN-LSTM fusion network intelligent localization algorithm are as follows:

[0069] (1) Data preprocessing

[0070] The aim is to convert the raw vibration signals (continuous signals in the time domain) acquired by DAS into structured data suitable for network input through data preprocessing, and to remove environmental noise (such as low-frequency interference from geological vibrations and equipment vibrations). The specific steps are as follows:

[0071] ① Raw signal acquisition: The raw vibration signal is an offline time series, represented as Where N is the number of sampling points and the sampling frequency is... .

[0072] ② A filter is used to remove low-frequency noise. The filtered signal is: .

[0073] ③ Divide the continuous signal into frames of fixed length (window length L = 512 points, corresponding to time).

[0074] Each frame of signal serves as an input sample for the network.

[0075]

[0076] Where S is the step size (S=L / 2 is used to avoid information loss). For frame index.

[0077] (2) CNN Feature Extraction Layer

[0078] The goal is to extract local features from each frame of signal using 1D convolution operations, such as the peak value of the leakage impact, waveform slope changes, and high-frequency harmonics. These features are directly related to the distance to the leakage point (the closer the distance, the larger the signal amplitude and the richer the high-frequency components). The specific steps are as follows:

[0079] ①1D convolutional layer

[0080] For input frames ,use convolution kernels , ( =3-5, adapting to the short-period characteristics of high-frequency signals, are convolved to output a feature map:

[0081]

[0082] in: Location index of the feature map ( ),

[0083] For the activation function, ReLU is used to suppress noise: ,

[0084] This is the bias of the k-th convolutional kernel.

[0085] ② Pooling layer

[0086] Max pooling (window size 2, stride 2) is performed on the feature map output by convolution to preserve local maxima (enhancing peak features of anomalous signals):

[0087]

[0088] ③CNN structure configuration

[0089] The output dimension is gradually compressed from 1×1024 to K3×128 (K3=64, i.e., 64 feature maps, each with 128 dimensions) by using 3 layers of convolution and stacked pooling.

[0090] (3) LSTM timing processing layer

[0091] This system processes the frame sequence features of CNN output to capture the temporal changes of anomalous signals (such as the start / end time of anomalies and the time difference of multipath reflected waves), thus solving the non-stationary problem of signals in underground chambers caused by reflection and scattering. Long Short-Term Memory (LSTM) is implemented through a "forget gate, input gate, cell state, and output gate."

[0092] ① Forgotten Gate (decides to discard historical information)

[0093]

[0094] in: This is the hidden state from the previous moment. The current input CNN features, , For weights and biases, This is the sigmoid function (outputs 0-1, controlling the forgetting rate).

[0095] ② Input gate (update cell state):

[0096]

[0097]

[0098] in: To update the ratio, The candidate cell state is represented by tanh output [-1,1], which controls the update amplitude.

[0099] ③ Cell state update:

[0100]

[0101] For element-wise multiplication, useful historical information is preserved and new information is added.

[0102] ④ Output gate (generates the current hidden state):

[0103]

[0104]

[0105] The hidden state at the current moment contains temporal features.

[0106] ⑤LSTM structure configuration

[0107] A 2-layer LSTM (128 hidden units) is used. The input is the feature sequence output by the CNN (length equal to the number of frames, e.g., 10 frames), and the output is the hidden state at the last time step. .

[0108] (4) Fully connected output layer (localization result regression)

[0109] The temporal features output by the LSTM are mapped to anomaly locations (distance L from the fiber optic start point, in meters).

[0110] ① Fully connected layer:

[0111] Hidden state The features are converted into high-dimensional features by a fully connected layer, and then the localization result is output through the output layer.

[0112]

[0113]

[0114] in: The distance to the predicted outlier. , , For weights and biases.

[0115] ②Loss function (regression task):

[0116] Mean squared error (MSE) is used to measure the difference between the predicted value and the true distance y:

[0117]

[0118] Where M is the number of samples, and the network parameters are optimized through backpropagation.

[0119] (5) Training and optimization

[0120] ① Optimizer

[0121] Using the Adam optimizer (adaptive learning rate), the parameter update formula is as follows:

[0122]

[0123] in: For network parameters, = 0.001 is the learning rate. , It is a moving average of first- and second-order momentum.

[0124] ② Training data

[0125] Tag data (known leak point distance + corresponding vibration signal) was collected through simulated abnormal experiments (leak sources with different apertures, pressures, and distances), and the data was divided into training set (80%), validation set (10%), and test set (10%).

[0126] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the embodiments of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0127] The above examples illustrate the present invention only to aid in understanding it and are not intended to limit the scope of the invention. Those skilled in the art can make various simple deductions, modifications, or substitutions based on the principles of this invention.

Claims

1. A method for early warning of air tightness in underground compressed air energy storage chambers based on distributed optical fibers, characterized in that: The method is implemented based on a distributed optical fiber-based compressed air storage underground chamber air tightness monitoring system, which includes a distributed optical fiber sensing unit, a signal transmission and reception unit, and a data processing and early warning unit. The distributed optical fiber sensing unit consists of multiple high-voltage resistant sensing optical fibers, which are arranged in a grid pattern on the surface of the chamber to form a full-coverage monitoring area. Each area has a unique identification code. The signal transmitting and receiving unit includes a DAS host and a DTS host. The DAS host transmits narrow pulse laser to the optical fiber and receives Rayleigh scattering signals, while the DTS host collects optical fiber axial temperature distribution data in real time. The data processing and early warning unit has built-in signal analysis module, leakage identification module, location calculation module and early warning triggering module, which are used for synchronous processing and intelligent response of multi-source sensor data; The method includes the following steps: S1: System initialization, parameter calibration of DAS and DTS main units, acquisition of acoustic baseline signal and temperature baseline distribution under leak-free conditions in the chamber, and setting of normal fluctuation threshold range; S2: Real-time data acquisition, synchronously acquiring Rayleigh scattering signals and temperature distribution data of the fiber array, with a sampling frequency of 2-4kHz and a data transmission delay of ≤1 second; S3: Abnormal signal identification. Real-time data is compared with baseline data. When the acoustic signal characteristics match the template library by ≥85%, or the local temperature change is ≥0.1℃ and the duration is ≥3 seconds, it is marked as an abnormal signal. S4: Leakage location and quantification. The abnormal area is determined by the distance code corresponding to the abnormal signal, the coordinates of the abnormal point are determined, and then the leakage amount Q is estimated based on the rate of change of temperature gradient. S5: Tiered early warning triggering. Multiple levels of early warning can be set based on the estimated leakage amount or leakage diffusion rate. The early warning signal will simultaneously trigger audible and visual alarms and data upload. in: In step S4, the leakage amount is estimated using the following formula: Q = k × ΔT × A Where: k is a correction factor related to the chamber pressure; ΔT is the maximum temperature change value; A represents the area of ​​the temperature anomaly region; In step S5, the multi-level early warning includes four levels: a gray warning when the leakage amount Q < 0.01 m³ / s and the leakage diffusion velocity v < 0.1 m / s; a yellow warning when 0.01 ≤ Q < 0.05 m³ / s or 0.1 ≤ v < 0.3 m / s; an orange warning when 0.01 ≤ Q < 0.05 m³ / s and 0.1 ≤ v < 0.3 m / s; and a red warning when Q ≥ 0.05 m³ / s or v ≥ 0.3 m / s.

2. The method for early warning of air tightness in underground compressed air storage chambers based on distributed optical fibers as described in claim 1, characterized in that, The mesh layout parameters of the sensing optical fiber are adjustable to adapt to different monitoring accuracy requirements and form a dual-end monitoring link.

3. The method for early warning of air tightness in underground compressed air storage chambers based on distributed optical fibers as described in claim 1, characterized in that, The data processing and early warning unit has a built-in leak sound signal template library, which contains characteristic parameters of compressed air leaks, and the template library can be dynamically updated based on historical data and on-site calibration.

4. The method for early warning of air tightness in underground compressed air storage chambers based on distributed optical fibers as described in claim 1, characterized in that, In step S3, a noise removal algorithm is used to eliminate environmental interference signals such as geological vibration and equipment operation by analyzing the frequency filtering and duration comparison of different events.

5. The method for early warning of air tightness in underground compressed air storage chambers based on distributed optical fibers as described in claim 1, characterized in that, In step S4, determining the coordinates of the anomaly points includes: initially locating the anomaly region by using the distance encoding corresponding to the anomaly signal, and accurately locating the coordinates of the anomaly points based on the intelligent localization algorithm of the CNN-LSTM fusion network.

6. The method for early warning of air tightness of underground compressed air energy storage chambers based on distributed optical fibers as described in claim 5, characterized in that, The CNN-LSTM fusion network intelligent localization algorithm includes data preprocessing, CNN feature extraction, LSTM temporal processing, fully connected output, and training optimization steps.

Images