Method and system for detecting leakage position of heat distribution pipeline based on automatic feature extraction

Abstract

The invention relates to the technical field of distributed optical fiber sensing temperature measurement, and provides a method and a system for detecting a leakage position of a heat distribution pipeline based on automatic feature extraction. The detection method comprises the following steps: acquiring training data, establishing a classification model, training the classification model by adopting the training data, acquiring a real-time signal, and determining a data type of the real-time signal according to the real-time signal and the trained classification model; and determining the leakage position of the heat distribution pipeline in response to the fact that the data type of the real-time signal is one type of data. The detection method effectively solves the problem that the pipeline leakage position information cannot be accurately obtained based on Raman scattering distributed optical fiber heat distribution pipeline temperature information, manual judgment is reduced, meanwhile, the accuracy rate is increased, and the method has great significance in real-time monitoring of the whole road section and the whole life cycle of the heat distribution pipeline.

Description

Technical Field

[0001] This application relates to the technical field of distributed optical fiber sensing temperature measurement, and particularly to a method and system for detecting the leakage position of a thermal pipeline with automatic feature extraction. Background Art

[0002] According to the technical characteristics of optical fiber distributed temperature measurement technology, such as long monitoring distance, high positioning accuracy, and strong anti-interference ability, it can realize the comprehensive detection of the temperature of thermal pipelines, providing strong support for the basic operation data of the pipe network for thermal pipeline operation enterprises. Since the thermal pipeline data is relatively complex, the peak areas extracted may also be the pipe jacking positions, the leakage positions of thermal pipelines, and the connection positions of thermal pipelines. When the pipeline is laid through areas such as rivers and roads that cannot be directly excavated, during the laying of optical fibers, equipment such as jacks and pipe jacking machines are used to gradually push the pipeline into the soil layer to complete the laying of the pipeline. In this way, the optical fiber is laid above the pipeline, which will cause the temperature of the thermal data in this area to be continuously high, and the data temperature at the leakage position is also high. For the connection position, in order to ensure the sealing performance, more connection materials will be constructed at the connection, resulting in a higher temperature.

[0003] However, in the existing detection methods, during the detection process of the leakage position, it is easy to confuse the pipe jacking position, the connection position of the thermal pipe, and the leakage position, resulting in misjudgment of the leakage position and affecting the detection accuracy.

[0004] Therefore, there is an urgent need for a detection method that can accurately determine the leakage position. Summary of the Invention

[0005] This application provides a method and system for detecting the leakage position of a thermal pipeline with automatic feature extraction to solve the technical problem that the existing thermal pipeline detection methods cannot accurately detect the leakage position.

[0006] The first aspect of this application provides a method for detecting the leakage position of a thermal pipeline with automatic feature extraction, which is applied to a thermal pipeline and includes: obtaining training data; where the training data includes acquisition signals, and the acquisition signals include leakage position data, pipe jacking position data, and connecting pipe position data. The leakage position data, pipe jacking position data, and connecting pipe position data all include temperature parameters and position parameters; establishing a classification model; training the classification model with the training data; where the input of the classification model is the acquisition signal, and the output of the classification model is the data type, and the data type includes type I data and type II data. Type I data includes leakage position data; type II data includes pipe jacking position data and connecting pipe position data; acquiring real-time signals; determining the data type of the real-time signal according to the real-time signal and the trained classification model; in response to the data type of the real-time signal being type I data, determining the leakage position of the thermal pipeline.

[0007] In a feasible implementation, the acquired signal includes a backward Stokes optical signal and an anti-Stokes scattered optical signal; training the classification model using training data includes: processing the backward Stokes optical signal and the anti-Stokes scattered optical signal; training the classification model using the processed backward Stokes optical signal and anti-Stokes scattered optical signal.

[0008] In a feasible implementation, processing the backward Stokes optical signal and the anti-Stokes scattered optical signal includes: preprocessing the backward Stokes optical signal and the anti-Stokes scattered optical signal, and extracting first data; wherein, the first data is one-dimensional data corresponding to temperature parameters and position parameters; decomposing and performing low-frequency processing on the first data to obtain second data; determining abnormal data in the second data where the temperature is higher than the temperature threshold; determining third data; wherein, the third data is the original data in the first data corresponding to the abnormal data.

[0009] In a feasible implementation, training the classification model using the processed backward Stokes optical signal and the anti-Stokes scattered optical signal includes: performing dimensionality reduction processing on the third data; classifying the dimensionally reduced third data using a clustering algorithm.

[0010] In a feasible implementation, the dimensionality reduction processing includes PCA dimensionality reduction processing; performing dimensionality reduction processing on the third data includes: creating a temperature vector based on the third data; normalizing the temperature vector; determining the covariance matrix of multiple normalized temperature vectors; calculating the eigenvalues and eigenvectors of the covariance matrix; determining new basis vectors based on the eigenvalues and eigenvectors; obtaining the dimensionally reduced third data based on the new basis vectors and the third data.

[0011] In a feasible implementation, determining abnormal data in the second data where the temperature is higher than the temperature threshold includes: obtaining the temperature mean and standard deviation in the second data; determining the temperature threshold based on the temperature mean and standard deviation; wherein, the temperature threshold is the sum of the temperature mean and the standard deviation; determining the abnormal data based on the temperature threshold; wherein, the temperature in the abnormal data is the peak temperature higher than the temperature threshold.

[0012] In a feasible implementation, decomposing and performing low-frequency processing on the first data to obtain second data includes: decomposing the first data using a multi-frequency decomposition technique; performing low-frequency processing on the decomposed first data to obtain second data; wherein, the low-frequency processing uses a Haar wavelet function.

[0013] In a feasible implementation, the preprocessing includes one or more of filtering, averaging, and calibration.

[0014] The leakage location detection method of the thermal pipeline with automatic feature extraction provided in the first aspect of the present application can effectively solve the problem that the temperature information of the distributed optical fiber thermal pipeline based on Raman scattering cannot accurately obtain the pipeline leakage location information, reduce manual judgment while increasing the accuracy rate, and has important significance for the real-time monitoring of the entire section and the entire life cycle of the thermal pipeline.

[0015] The leakage location detection system of the thermal pipeline with automatic feature extraction provided in the second aspect of the present application adopts the leakage location detection method of the thermal pipeline with automatic feature extraction provided in the first aspect. The leakage location detection system of the thermal pipeline with automatic feature extraction includes: an acquisition module configured to acquire training data; wherein, the training data includes acquisition signals, and the acquisition signals include leakage location data, jacking pipe location data, and connecting pipe location data. The leakage location data, jacking pipe location data, and connecting pipe location data all include temperature parameters and location parameters; a building module configured to build a classification model; a training module configured to train the classification model using the training data; wherein, the input of the classification model is the acquisition signal, and the output of the classification model is the data type, and the data type includes type I data and type II data. Type I data includes leakage location data; type II data includes jacking pipe location data and connecting pipe location data; an acquisition module configured to acquire real-time signals; a determination module configured to determine the data type of the real-time signal according to the real-time signal and the trained classification model; the determination module is further configured to, in response to the data type of the real-time signal being type I data, determine the leakage location of the thermal pipeline.

[0016] In a feasible implementation manner, the acquisition signal includes a backward Stokes optical signal and an anti-Stokes scattered optical signal; the training module is further configured to process the backward Stokes optical signal and the anti-Stokes scattered optical signal; and use the processed backward Stokes optical signal and anti-Stokes scattered optical signal to train the classification model.

[0017] The leakage location detection system of the thermal pipeline with automatic feature extraction provided in the second aspect of the present application adopts the leakage location detection method of the thermal pipeline with automatic feature extraction provided in the first aspect. Therefore, the beneficial technical effects thereof can be referred to the first aspect, and will not be elaborated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the present application, the drawings required for the embodiments will be briefly introduced below. Obviously, for those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0019] Figure 1 is a schematic structural diagram of an optical fiber laid for a thermal pipeline provided by an embodiment of the present application; Figure 2It is a schematic flow chart of a method for detecting the leakage position of a thermal pipeline with automatic feature extraction provided by an embodiment of the present application; Figure 3 It is a spectrogram of a data processing process provided by an embodiment of the present application; Figure 4 It is a structural block diagram of a controller provided by an embodiment of the present application.

[0020] Illustration marks: Controller; 101 - Acquisition module; 102 - Establishment module; 103 - Training module; 104 - Acquisition module; 105 - Determination module; 20 - Optical fiber. Detailed implementation manners

[0021] Next, the technical solutions in the embodiments of the present application will be clearly described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments of the present application, other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0022] Hereinafter, terms such as "first" and "second" are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, features defined with "first", "second", etc. may explicitly or implicitly include one or more of such features. In the description of the present application, unless otherwise stated, the meaning of "a plurality" is two or more.

[0023] In addition, in the present application, orientation terms such as "upper", "lower", "inner", "outer", etc. are defined relative to the orientation of the components shown in the accompanying drawings. It should be understood that these directional terms are relative concepts, which are used for relative description and clarification, and they can change accordingly with the change of the orientation of the components placed in the accompanying drawings.

[0024] Figure 1 It is a schematic structural diagram of an optical fiber laid for a thermal pipeline provided by an embodiment of the present application.

[0025] Refer to Figure 1 , Figure 1 In (a) of [reference], it is a schematic structural diagram of an optical fiber laid for a thermal pipeline in a normal state; refer to Figure 1 As shown in (a) of [reference], in a state of being flat, non-connected, and non-jacked, the optical fiber 20 is laid below the thermal pipeline. Figure 1 In (b) of [reference], it is a schematic structural diagram of an optical fiber laid for a thermal pipeline in a jacked state; refer to Figure 1 As shown in (b) of [reference], in a jacked state, the optical fiber 20 is laid above the thermal pipeline. Figure 1Figure (c) is a schematic structural diagram of laying optical fibers in the connected state of a heat pipeline. Refer to Figure 1 As shown in Figure (c), in the connected state, the optical fiber 20 is laid under the heat pipeline.

[0026] Among them, at the pipe jacking position and the connection position of the heat pipeline, due to the influence of pipe jacking and connection processes, the local temperature of the heat pipeline will increase. And this increase in temperature in this case is not caused by leakage. Therefore, during the detection of the heat pipeline, the pipe jacking position and the connection position need to be excluded. However, in the existing detection methods for detecting leakage positions, it is easy to confuse the pipe jacking position, the heat pipe connection position and the leakage position, resulting in misjudgment of the leakage position and affecting the detection accuracy.

[0027] To solve the technical problem of low detection accuracy, the embodiment of the present application provides a method for detecting the leakage position of a heat pipeline with automatic feature extraction, which is applied to the detection of heat pipelines. This detection method is simple and easy to implement, can realize all-weather real-time detection of heat pipelines, accurately judge the leakage position, effectively improve the detection accuracy, and provide strong support for the basic operation data of the pipe network in the field of heat pipeline operation.

[0028] Refer to Figure 2 , a method for detecting the leakage position of a heat pipeline with automatic feature extraction provided by the embodiment of the present application. This detection method can be implemented by the following steps S100 to step S600.

[0029] Step S100: Obtain training data.

[0030] Among them, the training data includes collected signals, and the collected signals include leakage position data, pipe jacking position data and connecting pipe position data. The leakage position data, pipe jacking position data and connecting pipe position data all include corresponding temperature parameters and position parameters.

[0031] At the beginning of executing step S100, a detection system for the leakage position of a heat pipeline with automatic feature extraction can be built first, and the optical fiber is used as a sensor to transmit optical signals. Specifically, the optical fiber is laid on the outer wall surface of the heat pipeline, and the optical fiber is in the same extending direction as the heat pipeline. A silicon core pipe can be sleeved outside the optical fiber, and the outer wall of the silicon core pipe fits on the outer wall surface of the heat pipeline. In this way, the position on the optical fiber can correspond to the position on the heat pipeline. By detecting and calculating the returned signal in the optical fiber, the leakage position on the heat pipeline corresponding to the optical fiber can be accurately judged.

[0032] Specifically, during the laying process, at non-pipe-jacking and non-connection positions, it is preferred to lay the optical fiber on the side of the heat pipe away from the ground. Since the optical fiber is relatively sensitive to external influences, with such an arrangement, the optical fiber is far from the ground, avoiding the influence of factors such as ground vibration on the working state of the optical fiber and effectively ensuring the accuracy and precision of the detection method. During the laying process, the silicon core tube sleeved outside the optical fiber fits on the outer wall surface of the heat pipe, and then the backfill soil is compacted, thereby ensuring the stability of the optical fiber and the heat pipe. After setting up the heat pipe leakage position detection system for automatic feature extraction, detection light can be sent to the optical fiber laid on the outer wall surface of the pipe. Among them, the detection light carries a detection optical signal.

[0033] Step S200: Establish a classification model.

[0034] Among them, the classification module can be a model with classification functions.

[0035] Step S300: Train the classification model using training data.

[0036] Among them, the input of the classification model is the acquired signal, and the output of the classification model is the data type. The data type includes type I data and type II data. Type I data includes leakage position data; type II data includes pipe-jacking position data and connecting pipe position data. In other words, when the classification model outputs type I data, it means that the acquired signal is a leakage signal. When the classification model outputs type II data, it means that the acquired signal is a non-leakage signal.

[0037] By training the classification model with training data, the data type of the real-time signal acquired subsequently can be determined using the classification model.

[0038] Specifically, the acquired signal can include the backward Stokes optical signal and the anti-Stokes scattered optical signal. Step S300 can be implemented by the following step S301 and step S302.

[0039] Step S301: Process the backward Stokes optical signal and the anti-Stokes scattered optical signal.

[0040] Specifically, there are usually noise signals generated by external interference in the received backward Stokes optical signal and anti-Stokes scattered optical signal. The purpose of processing the backward Stokes optical signal and the anti-Stokes scattered optical signal is to reduce the influence of noise. Among them, there are various processing methods, and this application does not make specific limitations on the processing method.

[0041] In some feasible implementation manners, step S301 can be implemented by the following steps S3011 to S3014.

[0042] Step S3011: Preprocess the backward Stokes optical signal and the anti-Stokes scattered optical signal, and extract the first data.

[0043] Among them, the first data is one-dimensional data corresponding to the temperature parameter and the position parameter. That is to say, the first data contains the temperature information of each position of the optical fiber along the thermal pipeline, and there is noise interference in this temperature information.

[0044] The purpose of preprocessing is to filter out noise signals, effectively ensure the detection accuracy, and avoid misjudgment.

[0045] Specifically, the preprocessing method can include one or more of filtering, averaging, and calibration. Of course, in other implementation manners, the preprocessing manner can also be other manners.

[0046] See Figure 3 , Figure 3 In (a) of [reference], it is the spectrogram of the first data. Figure 3 In (b) of [reference], it is the spectrogram of the second data. Figure 3 In (c) of [reference], it is the spectrogram of the third data. Among them, the abscissa is the length, and at this time, the optical fiber can be a 12 Km optical fiber detection section.

[0047] See Figure 3 As shown in (a) of [reference], in the first data, there are many temperature fluctuation situations. These temperature fluctuation situations include thermal pipeline leakage, connection points, and local high temperatures at the pipe jacking points. And there are multiple noise interferences in the first data.

[0048] Specifically, the first data can be extracted at preset time intervals. Among them, the preset time can be 0.5 min, 1 min, or 1.5 min. Of course, it can also be other times.

[0049] The first data can also be extracted at a preset frequency. Among them, the preset frequency can be once every 1 min, twice every 1 min, once every 1.5 min, once every 3 min, etc. Of course, it can also be other frequencies.

[0050] Step S3012: Decompose and perform low-frequency processing on the first data to obtain the second data.

[0051] Among them, the purpose of decomposing and performing low-frequency processing on the first data is to facilitate subsequent calculations.

[0052] Specifically, step S3012 can be implemented by the following step S3012a and step S3012b.

[0053] S3012a: Decompose the first data using multi-frequency decomposition technology.

[0054] The multi-frequency decomposition technique is an efficient, reliable, and easy-to-implement signal processing technique that decomposes a one-dimensional signal into individual frequency signals using different methods. Specifically in this application, the multi-frequency decomposition technique can decompose the first data into multiple sub-data. In this way, subsequent processing of the sub-data can reduce the complexity of processing the first data and effectively simplify the calculation process.

[0055] S3012b: Perform low-frequency processing on the decomposed first data to obtain second data.

[0056] Among them, the low-frequency processing uses the Haar wavelet function.

[0057] Perform low-frequency processing on multiple sub-data to further filter out the influence of high-frequency noise on the detection accuracy.

[0058] Step S3013: Determine the abnormal data in the second data whose temperature is higher than the temperature threshold.

[0059] Specifically, step S3013 can be implemented by the following steps S3013a to step S3013c.

[0060] S3013a: Obtain the temperature mean and standard deviation in the second data.

[0061] According to the temperature values corresponding to each coordinate point in the second data, calculate the temperature mean and standard deviation of this detection section through the temperature values.

[0062] S3013b: Determine the temperature threshold according to the temperature mean and standard deviation.

[0063] Among them, the temperature threshold is the sum of the temperature mean and standard deviation.

[0064] S3013c: Determine the abnormal data according to the temperature threshold.

[0065] Among them, the temperature in the abnormal data is the peak temperature higher than the temperature threshold.

[0066] After calculating the temperature threshold, the temperature threshold can be used as a reference point, retaining the data higher than the temperature threshold and excluding the data lower than the temperature threshold.

[0067] See Figure 3In (b), the temperature threshold can be 50°C. Therefore, data above 50°C are retained, and data below 50°C are eliminated. Among them, the second data obtained by processing is relatively smooth, and more temperature fluctuations are eliminated. In the obtained second data, there are five abnormal temperature peaks, namely the first peak A, the second peak B, the third peak C, the fourth peak D and the fifth peak E. The specific temperatures are 90°C at the first peak A, 61°C at the second peak B, 90°C at the third peak C, 96°C at the fourth peak D and 63°C at the fifth peak E. The corresponding positions of these five peaks are L A , L B , L C , L D , L E Thus, the abnormal data is the peak temperature where the temperature is higher than the temperature threshold. In other words, Figure 3 In the example (b), the abnormal data are (L A , 90℃)、(L B , 61℃), (L C , 90℃)、(L D , 96℃), (L E , 63℃).

[0068] Then, it is necessary to further identify the data of the leakage location among these five abnormal data.

[0069] Step S3014: Determine the third data.

[0070] The third data is the original data corresponding to the abnormal data in the first data. Then, the positions of the five peaks can be determined in the first data by comparing with the first data.

[0071] The third data is the original data of the abnormal data. Figure 3 In (c), peak A1 is the original data corresponding to the first peak A in the first data, peak B1 is the original data corresponding to the first peak B in the first data, peak C1 is the original data corresponding to the first peak C in the first data, peak D1 is the original data corresponding to the first peak D in the first data, and peak E1 is the original data corresponding to the first peak E in the first data. Figure 3 (c) Figure 3 The data other than peak A1, peak B1, peak C1, peak D1 and peak E1 in (a) are eliminated, thereby obtaining the third data.

[0072] Step S302: training a classification model using the processed backward Stokes light signal and anti-Stokes scattered light signal.

[0073] Specifically, step S302 can be implemented by the following steps S3021 and S3022.

[0074] Step S3021: Perform dimensionality reduction on the third data.

[0075] In some feasible implementation manners, principal component analysis (PCA) can be adopted. PCA dimensionality reduction is to project the original data onto new dimensions, thereby extracting the main features of the data. Using smaller dimensions to contain most of the information in the original data can reduce the storage space of the data and the demand for computing resources, and effectively improve the speed and efficiency of data processing.

[0076] Specifically in this application, step S3021 can be implemented by the following steps S3021a to S3021f.

[0077] Step S3021a: Create a temperature vector according to the third data.

[0078] Take the temperature of each peak in each third data as a time-domain feature input and create a temperature vector.

[0079] Step S3021b: Standardize the temperature vector.

[0080] Standardize the temperature vector of each peak. Among them, the standardization processing method can be to make its mean 0 and variance 1.

[0081] Step S3021c: Determine the covariance matrix of multiple standardized temperature vectors.

[0082] Calculate the covariance matrix of the standardized data to understand the correlation between features.

[0083] Step S3021d: Calculate the eigenvalues and eigenvectors of the covariance matrix.

[0084] The eigenvector is the basis vector, and the eigenvalue reflects the variance size corresponding to the basis vector.

[0085] Step S3021e: Determine the new basis vectors according to the eigenvalues and eigenvectors.

[0086] The eigenvectors corresponding to the two largest eigenvalues can be selected, and these two eigenvectors are the new basis vectors after data dimensionality reduction.

[0087] Step S3021f: Obtain the third data after dimensionality reduction according to the new basis vectors and the third data.

[0088] These eigenvectors project the third data onto the selected principal components to obtain the data after dimensionality reduction.

[0089] Step S3022: Classify the third data after dimensionality reduction using a clustering algorithm.

[0090] Classify the third data after dimensionality reduction by the clustering method. The classification results include type-one data and type-two data. In this way, the leakage data in the third data can be determined by the clustering method.

[0091] Step S400: Collect real-time signals.

[0092] Step S500: Determine the data type of the real-time signal according to the real-time signal and the trained classification model.

[0093] During actual detection, input the collected real-time signal into the trained classification model, and classify the real-time signal through the classification model.

[0094] Step S600: In response to the real-time signal being of type-one data, determine the leakage location of the heat pipeline.

[0095] When it is determined that the real-time signal is of type-one data, determine the leakage location according to the position parameter in the real-time signal, so as to determine the leakage location of the heat pipeline by automatically extracting the features in the real-time signal.

[0096] The method for detecting the leakage location of a heat pipeline with automatic feature extraction provided by the embodiment of the present application effectively solves the problem that the leakage location information of the pipeline cannot be accurately obtained from the temperature information of the distributed optical fiber heat pipeline based on Raman scattering, reduces manual judgment while increasing the accuracy rate, and has important significance for the real-time monitoring of the entire section and the full life cycle of the heat pipeline.

[0097] Corresponding to the foregoing embodiment of the method for detecting the leakage location of a heat pipeline with automatic feature extraction, the present application also provides an embodiment of a system for detecting the leakage location of a heat pipeline with automatic feature extraction. The detection system includes a controller 10 and an optical fiber 20. The optical fiber 20 is laid on the outer wall surface of the heat pipeline. The optical fiber 20 is underground and is connected to the controller 10 located on the ground. The controller 10 may include an acquisition module 101, a building module 102, a training module 103, a collection module 104, and a determination module 105.

[0098] The acquisition module 101 is configured to acquire training data; wherein, the training data includes acquisition signals, and the acquisition signals include leakage location data, pipe jacking location data, and connecting pipe location data. The leakage location data, pipe jacking location data, and connecting pipe location data all include temperature parameters and position parameters.

[0099] That is to say, the acquisition module 101 is used to execute step S100 in the embodiment of the method for detecting the leakage location of a heat pipeline with automatic feature extraction.

[0100] The building module 102 is configured to build a classification model.

[0101] That is, the establishment module 102 is used to execute step S200 in the embodiment of the method for detecting the leakage position of a thermal pipeline by automatic feature extraction.

[0102] The training module 103 trains a classification model using training data; wherein, the input of the classification model is the acquired signal, and the output of the classification model is the data type, and the data type includes type I data and type II data. Type I data includes leakage position data; type II data includes jacking pipe position data and connecting pipe position data.

[0103] That is, the training module 103 is used to execute step S300 in the embodiment of the method for detecting the leakage position of a thermal pipeline by automatic feature extraction.

[0104] In some feasible implementation manners, the acquired signal includes a backward Stokes optical signal and an anti-Stokes scattered optical signal; the training module 103 is further configured to process the backward Stokes optical signal and the anti-Stokes scattered optical signal; and the classification model is trained using the processed backward Stokes optical signal and anti-Stokes scattered optical signal.

[0105] The acquisition module 104 is configured to acquire real-time signals.

[0106] That is, the acquisition module 104 is used to execute step S400 in the embodiment of the method for detecting the leakage position of a thermal pipeline by automatic feature extraction.

[0107] The determination module 105 is configured to determine the data type of the real-time signal according to the real-time signal and the trained classification model.

[0108] That is, the determination module 105 is used to execute step S500 in the embodiment of the method for detecting the leakage position of a thermal pipeline by automatic feature extraction.

[0109] The determination module 105 is further configured to determine the leakage position of the thermal pipeline in response to the data type of the real-time signal being type I data.

[0110] That is, the determination module 105 is further used to execute step S600 in the embodiment of the method for detecting the leakage position of a thermal pipeline by automatic feature extraction.

[0111] Compared with the traditional system, this detection system does not require any hardware improvement and effectively saves the detection cost.

[0112] It should be noted that those skilled in the art will easily think of other implementation manners of this application after considering the specification and practicing the application disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application, and these variations, uses, or adaptations follow the general principles of this application and include common general knowledge or conventional technical means in the technical field not disclosed in this application.

[0113] It should be understood that the present application is not limited to the exact structures described above and shown in the drawings, and various modifications and changes can be made without departing from its scope. The true scope is indicated by the present application.

Claims

1. A method for detecting leakage position of a thermal pipeline by automatic feature extraction, applied in a thermal pipeline, characterized in that: include: Acquire training data; wherein the training data includes acquisition signals, the acquisition signals include leakage position data, jacking pipe position data and connecting pipe position data, and the leakage position data, the jacking pipe position data and the connecting pipe position data all include temperature parameters and position parameters; Build a classification model; The classification model is trained using the training data; wherein the input of the classification model is the acquisition signal, and the output of the classification model is a data type, the data type includes first-class data and second-class data, the first-class data includes the leakage position data; the second-class data includes the jacking pipe position data and the connecting pipe position data; Collect real-time signals; Determine the data type of the real-time signal according to the real-time signal and the trained classification model; In response to the data type of the real-time signal being the first type of data, a leakage location of the thermal pipeline is determined.

2. The method for detecting leakage position of thermal pipelines by automatic feature extraction according to claim 1 is characterized in that: The collected signals include back-Stokes light signals and anti-Stokes scattered light signals; Using the training data to train the classification model includes: Processing the backward Stokes light signal and the anti-Stokes scattered light signal; The classification model is trained using the processed backward Stokes light signal and the anti-Stokes scattered light signal.

3. The method for detecting leakage position of thermal pipelines by automatic feature extraction according to claim 2 is characterized in that: Processing the back-Stokes light signal and the anti-Stokes scattered light signal comprises: Preprocessing the back-Stokes light signal and the anti-Stokes scattered light signal, and extracting first data; wherein the first data is one-dimensional data corresponding to the temperature parameter and the position parameter; Decomposing and low-frequency processing the first data to obtain second data; Determining abnormal data in the second data whose temperature is higher than a temperature threshold; Determine third data; wherein the third data is original data in the first data corresponding to the abnormal data.

4. The method for detecting leakage position of thermal pipelines by automatic feature extraction according to claim 3 is characterized in that: Training the classification model using the processed backward Stokes light signal and the anti-Stokes scattered light signal comprises: Performing dimensionality reduction processing on the third data; The third data after dimension reduction is classified using a clustering algorithm.

5. The method for detecting leakage position of thermal pipelines by automatic feature extraction according to claim 4 is characterized in that: The dimensionality reduction processing includes PCA dimensionality reduction processing; performing dimensionality reduction processing on the third data includes: Creating a temperature vector according to the third data; normalizing the temperature vector; determining a covariance matrix of a plurality of normalized temperature vectors; Calculating the eigenvalues ​​and eigenvectors of the covariance matrix; Determine a new basis vector according to the eigenvalue and the eigenvector; The third data after dimension reduction is obtained according to the new basis vectors and the third data.

6. The method for detecting leakage position of thermal pipelines by automatic feature extraction according to claim 3 is characterized in that: Determining abnormal data in the second data whose temperature is higher than a temperature threshold includes: Obtaining a temperature mean and a standard deviation in the second data; Determining the temperature threshold according to the temperature mean and the standard deviation; wherein the temperature threshold is the sum of the temperature mean and the standard deviation; Abnormal data is determined according to the temperature threshold; wherein the temperature in the abnormal data is a peak temperature higher than the temperature threshold.

7. The method for detecting leakage position of thermal pipelines by automatic feature extraction according to claim 3 is characterized in that: Decomposing and low-frequency processing the first data to obtain second data includes: Decomposing the first data using a multi-frequency decomposition technique; The decomposed first data is subjected to low-frequency processing to obtain second data; wherein the low-frequency processing adopts Haar wavelet function.

8. The method for detecting leakage position of thermal pipelines by automatic feature extraction according to claim 3 is characterized in that: The preprocessing includes one or more of filtering, averaging, and calibration.

9. A thermal pipeline leakage position detection system using automatic feature extraction, using the thermal pipeline leakage position detection method using automatic feature extraction according to any one of claims 1 to 8, characterized in that: The thermal pipeline leakage position detection system with automatic feature extraction includes: An acquisition module is configured to acquire training data; wherein the training data includes an acquisition signal, the acquisition signal includes leakage position data, jacking pipe position data and connecting pipe position data, and the leakage position data, the jacking pipe position data and the connecting pipe position data all include temperature parameters and position parameters; A building module configured to build a classification model; A training module is configured to train the classification model using the training data; wherein the input of the classification model is the acquisition signal, and the output of the classification model is a data type, wherein the data type includes first-class data and second-class data, wherein the first-class data includes the leakage position data; and the second-class data includes the jacking pipe position data and the connecting pipe position data; An acquisition module, configured to acquire real-time signals; A determination module, configured to determine the data type of the real-time signal according to the real-time signal and the trained classification model; The determination module is further configured to determine a leakage location of the thermal pipeline in response to the data type of the real-time signal being type one data.

10. The thermal pipeline leakage location detection system according to claim 9, characterized in that: The collected signal includes a backward Stokes light signal and an anti-Stokes scattered light signal; the training module is further configured to process the backward Stokes light signal and the anti-Stokes scattered light signal; and the classification model is trained using the processed backward Stokes light signal and the anti-Stokes scattered light signal.

Images