CN118916821A - Optical fiber vibration detection method and system - Google Patents

Abstract

The present application provides a method and system for detecting optical fiber vibration, the method comprising: obtaining a vibration signal from an optical fiber sensor; performing feature extraction on the vibration signal to obtain vibration features, the vibration features including vibration amplitude, vibration frequency and phase; inputting the vibration features into a pre-trained regression model to obtain the vibration intensity of the vibration signal; inputting the vibration features into a pre-trained neural network model to obtain a category label corresponding to the vibration signal; obtaining a vibration score of the vibration signal based on the vibration intensity, the category label, the vibration frequency and the position information of the optical fiber sensor; and determining that the vibration signal is an abnormal vibration signal when the vibration score is greater than a preset score threshold. Using the method provided in the embodiment of the present application, it is possible to accurately determine whether a vibration signal is abnormal through multi-dimensional features.

Description

Optical fiber vibration detection method and system

Technical Field

The present application relates to the field of signal detection technology, and in particular to an optical fiber vibration detection method and system.

Background Art

In the fields of industrial production, infrastructure maintenance, etc., vibration signal analysis is an important technical means to monitor the status of equipment, predict the occurrence of failures, etc. In recent years, with the development of fiber optic sensing technology, fiber optic vibration detection has been widely used in vibration signal monitoring due to its advantages such as high sensitivity, long-distance transmission capability and anti-electromagnetic interference.

Traditional vibration signal processing methods usually rely on only a single feature (such as vibration intensity or frequency) to determine the status of the device. Although this method is simple, it cannot fully reflect the true status of the device by relying on only one feature, which can easily lead to misjudgment or missed judgment.

Summary of the invention

The technical problem to be solved by this application is to provide a method and system for optical fiber vibration detection, which can accurately determine whether the vibration signal is abnormal. The specific scheme is as follows:

An optical fiber vibration detection method, the method comprising:

Acquire vibration signals from optical fiber sensors;

Extracting features from the vibration signal to obtain vibration features, wherein the vibration features include vibration amplitude, vibration frequency and phase;

Inputting the vibration feature into a pre-trained regression model to obtain the vibration intensity of the vibration signal;

Inputting the vibration feature into a pre-trained neural network model to obtain a category label corresponding to the vibration signal;

Obtaining a vibration score of the vibration signal according to the vibration intensity, the category label, the vibration frequency, and the position information of the optical fiber sensor;

When the vibration score is greater than a preset score threshold, the vibration signal is determined to be an abnormal vibration signal.

In the above method, optionally, the training process of the regression model includes:

Acquire a first training sample set and an initial regression model to be trained; the first training sample set includes a plurality of first training samples and a vibration intensity corresponding to each first training sample; the first training sample includes a vibration amplitude, a vibration frequency and a phase of a historical vibration signal;

Determining a first target training sample currently used for training from each of the first training samples in the first training sample set;

Inputting the first target training sample into the initial regression model to obtain a first recognition result output by the initial regression model;

Calculate a first loss function value of the initial regression model based on a preset mean square error loss function, the vibration intensity corresponding to the first target training sample, and the first recognition result;

Updating the model parameters of the initial regression model using the first loss function value;

If the initial regression model after updating the model parameters does not meet the preset first training completion condition, returning to the step of determining a first target training sample currently used for training from each of the first training samples in the first training sample set;

When the initial regression model after updating the model parameters satisfies the preset first training completion condition, the initial regression model that satisfies the first training completion condition is used as the trained regression model.

In the above method, optionally, the training process of the neural network model includes:

Acquire a second training sample set and an initial neural network model to be trained; the second training sample set includes a plurality of second training samples and a classification label of each second training sample; the second training sample includes the vibration amplitude, vibration frequency and phase of the historical vibration signal, and the classification label includes one of a normal operation vibration type, an equipment failure vibration type, a man-made activity vibration type and a natural activity vibration type;

Determining a second target training sample currently used for training from each of the second training samples in the second training sample set;

Inputting a second target training sample into the initial neural network model to obtain a second recognition result output by the initial neural network model;

Calculate a second loss function value of the initial neural network model based on a preset cross entropy loss function, the classification label of the second target training sample, and the second recognition result;

Updating the model parameters of the initial neural network model using the second loss function value;

If the initial neural network model after updating the model parameters does not meet the preset second training completion condition, returning to the step of determining a second target training sample currently used for training from each of the second training samples in the second training sample set;

When the initial neural network model after updating the model parameters satisfies the preset second training completion condition, the initial neural network model that satisfies the second training completion condition is used as the trained neural network model.

The above method, optionally, determining that the vibration signal is an abnormal vibration signal further includes:

Obtaining an average value of the historical scoring thresholds corresponding to each historical vibration signal;

Calculate a target threshold value according to the scoring threshold value and the average value;

The current scoring threshold is replaced by the target threshold so as to use the target threshold as a new scoring threshold.

The above method, optionally, after determining that the vibration signal is an abnormal vibration signal, further comprises:

Obtaining the facility importance parameter corresponding to the optical fiber sensor and the cumulative number of occurrences of the vibration signal corresponding to the category label;

Determine the resource scheduling priority corresponding to the vibration signal according to the vibration intensity, the facility importance parameter corresponding to the optical fiber sensor, and the cumulative number of occurrences of the vibration signal corresponding to the category label;

A maintenance work order is generated according to the resource scheduling priority corresponding to the vibration signal, and the maintenance work order is output.

An optical fiber vibration detection system, comprising:

A first acquisition unit, used for acquiring a vibration signal from the optical fiber sensor;

A feature extraction unit, used to extract features from the vibration signal to obtain vibration features, wherein the vibration features include vibration amplitude, vibration frequency and phase;

A first execution unit, configured to input the vibration feature into a pre-trained regression model to obtain the vibration intensity of the vibration signal;

A second execution unit is used to input the vibration feature into a pre-trained neural network model to obtain a category label corresponding to the vibration signal;

A third execution unit, configured to obtain a vibration score of the vibration signal according to the vibration intensity, the category label, the vibration frequency, and the position information of the optical fiber sensor;

The first determining unit is configured to determine that the vibration signal is an abnormal vibration signal when the vibration score is greater than a preset score threshold.

In the above system, optionally, the first execution unit includes:

A first acquisition subunit is used to acquire a first training sample set and an initial regression model to be trained; the first training sample set includes a plurality of first training samples and a vibration intensity corresponding to each first training sample; the first training sample includes a vibration amplitude, a vibration frequency and a phase of a historical vibration signal;

A first determining subunit, configured to determine a first target training sample currently used for training in the first training sample set;

A first execution subunit is used to input a first target training sample into an initial regression model to obtain a first recognition result output by the initial regression model;

A first calculation subunit, configured to calculate a first loss function value of the initial regression model based on a preset mean square error loss function, the vibration intensity corresponding to the first target training sample, and the first recognition result;

A first updating subunit, configured to update the model parameters of the initial regression model using the first loss function value;

a second execution subunit, configured to return to trigger the first determination subunit to determine a first target training sample currently used for training from each of the first training samples in the first training sample set when the initial regression model after updating the model parameters does not meet the preset first training completion condition;

The third execution subunit is used to use the initial regression model that meets the first training completion condition as the trained regression model when the initial regression model after updating the model parameters meets the preset first training completion condition.

In the above system, optionally, the second execution unit includes:

a second acquisition subunit, configured to acquire a second training sample set and an initial neural network model to be trained; the second training sample set includes a plurality of second training samples and a classification label of each second training sample; the second training sample includes a vibration amplitude, a vibration frequency and a phase of a historical vibration signal, and the classification label includes one of a normal operation vibration type, an equipment failure vibration type, a man-made activity vibration type and a natural activity vibration type;

A second determining subunit, configured to determine a second target training sample currently used for training in the second training sample set;

A fourth execution subunit, used for inputting a second target training sample into the initial neural network model to obtain a second recognition result output by the initial neural network model;

A second calculation subunit, configured to calculate a second loss function value of the initial neural network model based on a preset cross entropy loss function, the classification label of the second target training sample, and the second recognition result;

A second updating subunit, used to update the model parameters of the initial neural network model using the second loss function value;

a fifth execution subunit, configured to return to trigger the second determination subunit to determine a second target training sample currently used for training from each of the second training samples in the second training sample set when the initial neural network model after updating the model parameters does not meet the preset second training completion condition;

The sixth execution subunit is used to use the initial neural network model that meets the second training completion condition as the trained neural network model when the initial neural network model after updating the model parameters meets the preset second training completion condition.

The above system may optionally further include:

A second acquisition unit is used to acquire an average value of the historical scoring thresholds corresponding to each historical vibration signal;

A calculation unit, configured to calculate a target threshold value according to the scoring threshold value and the average value;

A replacing unit is used to replace the current scoring threshold with a target threshold so as to use the target threshold as a new scoring threshold.

The above system may optionally further include:

A third acquisition unit is used to acquire the facility importance parameter corresponding to the optical fiber sensor and the cumulative number of occurrences of the vibration signal corresponding to the category label;

A second determining unit is used to determine the resource scheduling priority corresponding to the vibration signal according to the vibration intensity, the facility importance parameter corresponding to the optical fiber sensor, and the cumulative number of occurrences of the vibration signal corresponding to the category label;

An output unit is used to generate a maintenance work order according to the resource scheduling priority corresponding to the vibration signal, and output the maintenance work order.

Based on the above-mentioned implementation of the present application, an optical fiber vibration detection method and system are provided, the method comprising: obtaining a vibration signal from an optical fiber sensor; performing feature extraction on the vibration signal to obtain vibration features, the vibration features including vibration amplitude, vibration frequency and phase; inputting the vibration features into a pre-trained regression model to obtain the vibration intensity of the vibration signal; inputting the vibration features into a pre-trained neural network model to obtain the category label corresponding to the vibration signal; obtaining a vibration score of the vibration signal based on the vibration intensity, the category label, the vibration frequency and the position information of the optical fiber sensor; and determining that the vibration signal is an abnormal vibration signal when the vibration score is greater than a preset score threshold. Using the method provided in the embodiment of the present application, it is possible to accurately determine whether a vibration signal is abnormal through multi-dimensional features.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

FIG1 is a method flow chart of an optical fiber vibration detection method provided by the present application;

FIG2 is a flow chart of a training process of a regression model provided by the present application;

FIG3 is a flow chart of a training process of a neural network model provided by the present application;

FIG4 is a schematic diagram of the structure of an optical fiber vibration detection system provided in the present application.

DETAILED DESCRIPTION

The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

In this application, the terms "comprises", "comprising" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or device comprising the element.

The embodiment of the present application provides a method for optical fiber vibration detection, which can be applied to electronic devices, and the electronic devices can be controllers, personal computers, servers, tablet devices, smart phones, smart wearable devices, etc. The method flow chart of the method is shown in FIG1, and specifically includes:

S101: Acquire a vibration signal from a fiber optic sensor.

In this embodiment, a light pulse can be emitted to the optical fiber, and then the optical signal in the optical fiber is collected by the optical fiber sensor, and then the optical signal is converted into an electrical signal, and then the electrical signal is converted into a digital signal, that is, a vibration signal.

S102: Extract features from the vibration signal to obtain vibration features, where the vibration features include vibration amplitude, vibration frequency and phase.

In this embodiment, the vibration signal can be first denoised, for example, the vibration signal can be filtered and averaged, and then the vibration signal is analyzed in the time domain to calculate the vibration amplitude; the vibration signal is analyzed in the frequency domain to obtain the vibration frequency, which can be the center frequency, peak frequency, etc.; then the vibration signal is analyzed in the time and frequency domains to obtain the phase of the vibration signal.

S103: Inputting the vibration feature into a pre-trained regression model to obtain the vibration intensity of the vibration signal.

In this embodiment, the regression model may be a multiple linear regression model, a support vector machine regression model, a random forest regression model, or the like.

S104: Inputting the vibration feature into a pre-trained neural network model to obtain a category label corresponding to the vibration signal.

In this embodiment, the neural network model may include an input layer, a hidden layer, and an output layer. The neural network may be a fully connected neural network, a multi-layer perceptron, a convolutional neural network, etc.

S105: Obtain a vibration score of the vibration signal according to the vibration intensity, the category label, the vibration frequency, and the position information of the optical fiber sensor.

In this embodiment, the vibration score of the vibration signal can be calculated according to a preset vibration score formula, which is as follows:

Among them, O represents the vibration score; V represents the vibration intensity, represents the weight coefficient of vibration intensity, C represents the category label, represents the weight coefficient of the category label, F represents the vibration frequency, Represents the weight coefficient of the vibration frequency, H represents the current time, and I(H) is an indicator function. For a specific time period H, it returns 1 to indicate that special attention is required, otherwise it returns 0; represents the weight coefficient of the current time; L represents the location information of the optical fiber sensor, represents the weight coefficient of the position information; D represents the cumulative number of vibrations corresponding to the category label, Indicates the weight coefficient of the cumulative number of occurrences.

S106: When the vibration score is greater than a preset score threshold, determine that the vibration signal is an abnormal vibration signal.

In this embodiment, the scoring threshold is a fixed threshold or a dynamic threshold. The scoring threshold may be initially set to 1.

By applying the method provided by the embodiment of the present application, it is possible to comprehensively consider information from multiple dimensions, so that it is possible to more accurately determine whether the vibration signal is an abnormal vibration signal, and it is possible to combine the location information, the current time, and the cumulative number of vibrations corresponding to the category label, so as to dynamically adjust the decision rules. For example, in a specific time period, for certain geographical locations, even if the vibration intensity is low, it may be necessary to pay attention; conversely, in other time periods or locations, higher vibration intensities may not require immediate action. This method can better adapt to different application scenarios and changes in conditions. Through multi-dimensional data fusion, the robustness of the detection process can be enhanced and misjudgments caused by fluctuations in a single feature can be reduced. For example, even if the vibration intensity at a certain point in time is abnormal, if such vibrations often occur at that location (supported by historical data), the system can make decisions more cautiously to avoid frequent false alarms.

In an embodiment provided in the present application, based on the above solution, optionally, the training process of the regression model, as shown in FIG2 , includes:

S201: Acquire a first training sample set and an initial regression model to be trained; the first training sample set includes multiple first training samples and the vibration intensity corresponding to each first training sample; the first training sample includes the vibration amplitude, vibration frequency and phase of the historical vibration signal.

S202: Determine a first target training sample currently used for training from each of the first training samples in the first training sample set.

S203: Inputting the first target training sample into the initial regression model to obtain a first recognition result output by the initial regression model.

S204: Calculate a first loss function value of the initial regression model based on a preset mean square error loss function, the vibration intensity corresponding to the first target training sample, and the first recognition result.

S205: Using the first loss function value to update the model parameters of the initial regression model.

S206: Detect whether the initial regression model after updating the model parameters meets the preset first training completion condition; if not, execute S202; if yes, execute S207.

In this embodiment, the first training completion condition may be that the loss function of the initial regression model converges, the number of training times is greater than a preset first training times threshold, and the like.

S207: Using the initial regression model that meets the first training completion condition as a trained regression model.

In an embodiment provided in the present application, based on the above solution, optionally, the training process of the neural network model, as shown in FIG3 , includes:

S301: Obtain a second training sample set and an initial neural network model to be trained; the second training sample set includes multiple second training samples and a classification label for each second training sample; the second training sample includes the vibration amplitude, vibration frequency and phase of the historical vibration signal, and the classification label includes one of a normal operation vibration type, an equipment failure vibration type, a man-made activity vibration type and a natural activity vibration type.

S302: Determine a second target training sample currently used for training from each of the second training samples in the second training sample set.

S303: Inputting the second target training sample into the initial neural network model to obtain a second recognition result output by the initial neural network model.

S304: Based on a preset cross entropy loss function, the classification label of the second target training sample and the second recognition result, a second loss function value of the initial neural network model is calculated.

S305: Using the second loss function value to update the model parameters of the initial neural network model.

S306: Detect whether the initial neural network model after updating the model parameters meets the preset second training completion condition; if not, execute S302; if so, execute S307.

In this embodiment, the second training completion condition may be that the loss function of the initial neural network model converges, the prediction accuracy is greater than a preset accuracy threshold, the number of training times is greater than a preset first training times threshold, etc.

S307: Using the initial neural network model that meets the second training completion condition as a trained neural network model.

In an embodiment provided in the present application, based on the above solution, optionally, determining that the vibration signal is an abnormal vibration signal further includes:

Obtaining an average value of the historical scoring thresholds corresponding to each historical vibration signal;

Calculate a target threshold value according to the scoring threshold value and the average value;

The current scoring threshold is replaced by the target threshold so as to use the target threshold as a new scoring threshold.

In this embodiment, the process of calculating the target threshold value based on the scoring threshold value and the average value is as follows:

in, is the target threshold, is the average of all historical scoring thresholds, is the current scoring threshold, α is a smoothing factor, and can take values between (0, 1).

By applying the method provided in this embodiment and dynamically adjusting the threshold, it is possible to better adapt to environmental changes and reduce false positives and false negatives.

In an embodiment provided in the present application, based on the above solution, optionally, after determining that the vibration signal is an abnormal vibration signal, the method further includes:

Obtaining the facility importance parameter corresponding to the optical fiber sensor and the cumulative number of occurrences of the vibration signal corresponding to the category label;

Determine the resource scheduling priority corresponding to the vibration signal according to the vibration intensity, the facility importance parameter corresponding to the optical fiber sensor, and the cumulative number of occurrences of the vibration signal corresponding to the category label;

A maintenance work order is generated according to the resource scheduling priority corresponding to the vibration signal, and the maintenance work order is output.

Optionally, the facility importance parameter corresponding to the optical fiber sensor may be a parameter characterizing the importance of the facility monitored by the optical fiber sensor, and the facility importance parameter may be determined based on factors such as expert experience evaluation, the degree of impact after a facility failure, and user feedback.

In this embodiment, the process of determining the resource scheduling priority corresponding to the vibration signal is as follows:

In this embodiment, Facility importance level The priority weight of . is the priority weight of the cumulative number of occurrences D of the vibration signal corresponding to the category label, that is, the cumulative number of occurrences of the vibration corresponding to the category label. is the priority weight of the final category label O of the current vibration signal, is the priority weight of the vibration intensity V, The priority weight of the geographical distance L. The priority weight of the number of occurrences W within the time window.

By applying the method provided in the embodiment of the present application, by calculating the resource scheduling priority, it is possible to ensure that high-priority tasks are processed first. This helps the system to respond quickly and take timely measures to reduce the impact of failures when an emergency occurs. Resource scheduling priority can help to reasonably allocate limited resources. In the case of limited resources, giving priority to important tasks can maximize the utilization efficiency of resources and avoid waste of resources. By prioritizing different tasks, it can be ensured that key facilities and important tasks are given priority. This helps to maintain the stable operation of the system and reduce system failures caused by improper resource allocation. Automatically calculating resource scheduling priorities can reduce the need for manual intervention. The resource scheduling priority can be adjusted not only according to the current state, but also dynamically adjusted according to historical data and forecast information. This means that the system can adjust the resource allocation strategy in a timely manner according to changes in actual conditions. By recording and analyzing historical data, the priority calculation algorithm can be continuously optimized. With the accumulation of data, the system can become more intelligent and continuously improve the accuracy and efficiency of resource scheduling.

In this embodiment, the real-time monitoring system can be combined to calculate the resource scheduling priority to achieve real-time monitoring and early warning functions. Once a high-priority task is found, the system can immediately notify relevant personnel and take corresponding measures.

Corresponding to the method described in FIG. 1 , the embodiment of the present application further provides an optical fiber vibration detection system, which is applied to electronic equipment and is used to specifically implement the method described in FIG. 4 . The structural schematic diagram of the system is shown in FIG. 4 , and includes:

A first acquisition unit 401 is used to acquire a vibration signal from the optical fiber sensor;

A feature extraction unit 402 is used to extract features from the vibration signal to obtain vibration features, where the vibration features include vibration amplitude, vibration frequency and phase;

A first execution unit 403 is used to input the vibration feature into a pre-trained regression model to obtain the vibration intensity of the vibration signal;

The second execution unit 404 is used to input the vibration feature into a pre-trained neural network model to obtain a category label corresponding to the vibration signal;

A third execution unit 405 is used to obtain a vibration score of the vibration signal according to the vibration intensity, the category label, the vibration frequency and the position information of the optical fiber sensor;

The first determining unit 406 is configured to determine that the vibration signal is an abnormal vibration signal if the vibration score is greater than a preset score threshold.

In an embodiment provided in the present application, based on the above solution, optionally, the first execution unit 403 includes:

A first acquisition subunit is used to acquire a first training sample set and an initial regression model to be trained; the first training sample set includes a plurality of first training samples and a vibration intensity corresponding to each first training sample; the first training sample includes a vibration amplitude, a vibration frequency and a phase of a historical vibration signal;

A first determining subunit, configured to determine a first target training sample currently used for training in the first training sample set;

A first execution subunit is used to input a first target training sample into an initial regression model to obtain a first recognition result output by the initial regression model;

A first calculation subunit, configured to calculate a first loss function value of the initial regression model based on a preset mean square error loss function, the vibration intensity corresponding to the first target training sample, and the first recognition result;

A first updating subunit, configured to update the model parameters of the initial regression model using the first loss function value;

a second execution subunit, configured to return to trigger the first determination subunit to determine a first target training sample currently used for training from each of the first training samples in the first training sample set when the initial regression model after updating the model parameters does not meet the preset first training completion condition;

The third execution subunit is used to use the initial regression model that meets the first training completion condition as the trained regression model when the initial regression model after updating the model parameters meets the preset first training completion condition.

In an embodiment provided in the present application, based on the above solution, optionally, the second execution unit 404 includes:

a second acquisition subunit, configured to acquire a second training sample set and an initial neural network model to be trained; the second training sample set includes a plurality of second training samples and a classification label of each second training sample; the second training sample includes a vibration amplitude, a vibration frequency and a phase of a historical vibration signal, and the classification label includes one of a normal operation vibration type, an equipment failure vibration type, a man-made activity vibration type and a natural activity vibration type;

A second determining subunit, configured to determine a second target training sample currently used for training in the second training sample set;

A fourth execution subunit, used for inputting a second target training sample into the initial neural network model to obtain a second recognition result output by the initial neural network model;

A second calculation subunit, configured to calculate a second loss function value of the initial neural network model based on a preset cross entropy loss function, the classification label of the second target training sample, and the second recognition result;

A second updating subunit, used to update the model parameters of the initial neural network model using the second loss function value;

a fifth execution subunit, configured to return to trigger the second determination subunit to determine a second target training sample currently used for training from each of the second training samples in the second training sample set when the initial neural network model after updating the model parameters does not meet the preset second training completion condition;

The sixth execution subunit is used to use the initial neural network model that meets the second training completion condition as the trained neural network model when the initial neural network model after updating the model parameters meets the preset second training completion condition.

In an embodiment provided in the present application, based on the above solution, optionally, it further includes:

A second acquisition unit is used to acquire an average value of the historical scoring thresholds corresponding to each historical vibration signal;

A calculation unit, configured to calculate a target threshold value according to the scoring threshold value and the average value;

A replacing unit is used to replace the current scoring threshold with a target threshold so as to use the target threshold as a new scoring threshold.

In an embodiment provided in the present application, based on the above solution, optionally, it further includes:

A third acquisition unit is used to acquire the facility importance parameter corresponding to the optical fiber sensor and the cumulative number of occurrences of the vibration signal corresponding to the category label;

A second determining unit is used to determine the resource scheduling priority corresponding to the vibration signal according to the vibration intensity, the facility importance parameter corresponding to the optical fiber sensor, and the cumulative number of occurrences of the vibration signal corresponding to the category label;

An output unit is used to generate a maintenance work order according to the resource scheduling priority corresponding to the vibration signal, and output the maintenance work order.

The specific principles and execution processes of each unit and module in the optical fiber vibration detection system disclosed in the above-mentioned embodiment of the present application are the same as those of the optical fiber vibration detection method disclosed in the above-mentioned embodiment of the present application. Please refer to the corresponding parts of the optical fiber vibration detection method provided in the above-mentioned embodiment of the present application, and will not be repeated here.

It should be noted that each embodiment in this specification is described in a progressive manner, and each embodiment focuses on the differences from other embodiments, and the same or similar parts between the embodiments can be referred to each other. For system embodiments, since they are basically similar to method embodiments, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

Finally, it should be noted that, in this article, relational terms such as first and second, etc. are merely used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations.

For the convenience of description, the above device is described in various units according to their functions. Of course, when implementing the present application, the functions of each unit can be implemented in the same or multiple software and/or hardware.

It can be known from the description of the above implementation methods that those skilled in the art can clearly understand that the present application can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present application can be essentially or partly contributed to the prior art in the form of a software product, which can be stored in a storage medium such as ROM/RAM, a magnetic disk, an optical disk, etc., and includes several instructions for enabling a computer device (which can be a personal computer, a server, etc.) to execute the methods described in the various embodiments of the present application or certain parts of the embodiments.

The above is a detailed introduction to the optical fiber vibration detection method provided by the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea; at the same time, for general technical personnel in this field, according to the idea of the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims (10)

1. A method for detecting optical fiber vibration, comprising: Acquire vibration signals from optical fiber sensors; Extracting features from the vibration signal to obtain vibration features, wherein the vibration features include vibration amplitude, vibration frequency and phase; Inputting the vibration feature into a pre-trained regression model to obtain the vibration intensity of the vibration signal; Inputting the vibration feature into a pre-trained neural network model to obtain a category label corresponding to the vibration signal; Obtaining a vibration score of the vibration signal according to the vibration intensity, the category label, the vibration frequency, and the position information of the optical fiber sensor; When the vibration score is greater than a preset score threshold, the vibration signal is determined to be an abnormal vibration signal. 2. The method according to claim 1, characterized in that the training process of the regression model comprises: Acquire a first training sample set and an initial regression model to be trained; the first training sample set includes a plurality of first training samples and a vibration intensity corresponding to each first training sample; the first training sample includes a vibration amplitude, a vibration frequency and a phase of a historical vibration signal; Determining a first target training sample currently used for training from each of the first training samples in the first training sample set; Inputting the first target training sample into the initial regression model to obtain a first recognition result output by the initial regression model; Calculate a first loss function value of the initial regression model based on a preset mean square error loss function, the vibration intensity corresponding to the first target training sample, and the first recognition result; Using the first loss function value to update the model parameters of the initial regression model; If the initial regression model after updating the model parameters does not meet the preset first training completion condition, returning to the step of determining a first target training sample currently used for training from each of the first training samples in the first training sample set; When the initial regression model after updating the model parameters satisfies the preset first training completion condition, the initial regression model that satisfies the first training completion condition is used as the trained regression model. 3. The method according to claim 1, characterized in that the training process of the neural network model comprises: Acquire a second training sample set and an initial neural network model to be trained; the second training sample set includes a plurality of second training samples and a classification label of each second training sample; the second training sample includes the vibration amplitude, vibration frequency and phase of the historical vibration signal, and the classification label includes one of a normal operation vibration type, an equipment failure vibration type, a man-made activity vibration type and a natural activity vibration type; Determining a second target training sample currently used for training from each of the second training samples in the second training sample set; Inputting a second target training sample into the initial neural network model to obtain a second recognition result output by the initial neural network model; Calculate a second loss function value of the initial neural network model based on a preset cross entropy loss function, the classification label of the second target training sample, and the second recognition result; Updating the model parameters of the initial neural network model using the second loss function value; When the initial neural network model after updating the model parameters does not meet the preset second training completion condition, returning to the step of determining the second target training sample currently used for training from each of the second training samples in the second training sample set; When the initial neural network model after updating the model parameters satisfies the preset second training completion condition, the initial neural network model that satisfies the second training completion condition is used as the trained neural network model. 4. The method according to claim 1, characterized in that the step of determining that the vibration signal is an abnormal vibration signal further comprises: Obtaining an average value of the historical scoring thresholds corresponding to each historical vibration signal; Calculate a target threshold value according to the scoring threshold value and the average value; The current scoring threshold is replaced by the target threshold so as to use the target threshold as a new scoring threshold. 5. The method according to claim 1, characterized in that after determining that the vibration signal is an abnormal vibration signal, it also includes: Obtaining the facility importance parameter corresponding to the optical fiber sensor and the cumulative number of occurrences of the vibration signal corresponding to the category label; Determine the resource scheduling priority corresponding to the vibration signal according to the vibration intensity, the facility importance parameter corresponding to the optical fiber sensor, and the cumulative number of occurrences of the vibration signal corresponding to the category label; A maintenance work order is generated according to the resource scheduling priority corresponding to the vibration signal, and the maintenance work order is output. 6. An optical fiber vibration detection system, comprising: A first acquisition unit, used for acquiring a vibration signal from the optical fiber sensor; A feature extraction unit, used to extract features from the vibration signal to obtain vibration features, wherein the vibration features include vibration amplitude, vibration frequency and phase; A first execution unit, configured to input the vibration feature into a pre-trained regression model to obtain the vibration intensity of the vibration signal; A second execution unit is used to input the vibration feature into a pre-trained neural network model to obtain a category label corresponding to the vibration signal; A third execution unit, configured to obtain a vibration score of the vibration signal according to the vibration intensity, the category label, the vibration frequency, and the position information of the optical fiber sensor; The first determining unit is configured to determine that the vibration signal is an abnormal vibration signal when the vibration score is greater than a preset score threshold. 7. The system according to claim 6, wherein the first execution unit comprises: A first acquisition subunit is used to acquire a first training sample set and an initial regression model to be trained; the first training sample set includes a plurality of first training samples and a vibration intensity corresponding to each first training sample; the first training sample includes a vibration amplitude, a vibration frequency and a phase of a historical vibration signal; A first determining subunit, configured to determine a first target training sample currently used for training in the first training sample set; A first execution subunit is used to input a first target training sample into an initial regression model to obtain a first recognition result output by the initial regression model; A first calculation subunit, configured to calculate a first loss function value of the initial regression model based on a preset mean square error loss function, the vibration intensity corresponding to the first target training sample, and the first recognition result; A first updating subunit, configured to update the model parameters of the initial regression model using the first loss function value; a second execution subunit, configured to return to trigger the first determination subunit to determine a first target training sample currently used for training from each of the first training samples in the first training sample set when the initial regression model after updating the model parameters does not meet the preset first training completion condition; The third execution subunit is used to use the initial regression model that meets the first training completion condition as the trained regression model when the initial regression model after updating the model parameters meets the preset first training completion condition. 8. The system according to claim 6, wherein the second execution unit comprises: a second acquisition subunit, configured to acquire a second training sample set and an initial neural network model to be trained; the second training sample set includes a plurality of second training samples and a classification label of each second training sample; the second training sample includes a vibration amplitude, a vibration frequency and a phase of a historical vibration signal, and the classification label includes one of a normal operation vibration type, an equipment failure vibration type, a man-made activity vibration type and a natural activity vibration type; A second determining subunit, configured to determine a second target training sample currently used for training in the second training sample set; A fourth execution subunit, used for inputting a second target training sample into the initial neural network model to obtain a second recognition result output by the initial neural network model; A second calculation subunit, configured to calculate a second loss function value of the initial neural network model based on a preset cross entropy loss function, the classification label of the second target training sample, and the second recognition result; A second updating subunit, used to update the model parameters of the initial neural network model using the second loss function value; a fifth execution subunit, configured to return to trigger the second determination subunit to determine a second target training sample currently used for training from each of the second training samples in the second training sample set when the initial neural network model after updating the model parameters does not meet the preset second training completion condition; The sixth execution subunit is used to use the initial neural network model that meets the second training completion condition as the trained neural network model when the initial neural network model after updating the model parameters meets the preset second training completion condition. 9. The system according to claim 6, further comprising: A second acquisition unit is used to acquire an average value of the historical scoring thresholds corresponding to each historical vibration signal; A calculation unit, configured to calculate a target threshold value according to the scoring threshold value and the average value; A replacing unit is used to replace the current scoring threshold with a target threshold so as to use the target threshold as a new scoring threshold. 10. The system according to claim 6, further comprising: A third acquisition unit is used to acquire the facility importance parameter corresponding to the optical fiber sensor and the cumulative number of occurrences of the vibration signal corresponding to the category label; A second determining unit is used to determine the resource scheduling priority corresponding to the vibration signal according to the vibration intensity, the facility importance parameter corresponding to the optical fiber sensor, and the cumulative number of occurrences of the vibration signal corresponding to the category label; An output unit is used to generate a maintenance work order according to the resource scheduling priority corresponding to the vibration signal, and output the maintenance work order.