Multi-domain analysis and early warning method for deep-ultra-deep shaft optical fiber data

Abstract

The invention discloses a deep-ultra-deep shaft optical fiber data multi-domain analysis and early warning method, which comprises the following steps: acquiring real-time optical fiber monitoring data of a deep-ultra-deep shaft, and filtering the real-time optical fiber monitoring data by utilizing a self-adaptive filtering technology; performing normalization processing on the filtered optical fiber monitoring data to obtain normalized optical fiber data; respectively extracting a time domain feature, a frequency domain feature and a time-frequency domain feature of the normalized optical fiber data; and based on the time domain feature, the frequency domain feature and the time-frequency domain feature of the normalized optical fiber data, carrying out early warning on the abnormal condition of the deep-ultra-deep shaft. The optical fiber monitoring data under the complex oil and gas well working condition can be efficiently and accurately analyzed, more comprehensive and timely feedback and early warning are provided, the mining efficiency of the deep-ultra-deep oil and gas well is optimized, the mining safety is improved, and economic losses or safety accidents caused by wrong analysis are avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of oil and gas engineering, and in particular to a multi-domain analysis and early warning method for deep-ultra-deep wellbore optical fiber data. Background Art

[0002] With the development of sensing technology and data processing algorithms, fiber optic sensors have played an important role in downhole environmental monitoring in the field of oil and gas field development and production due to their high precision, anti-electromagnetic interference, long life and corrosion resistance. Fiber optic sensing systems can provide continuous and rich multi-parameter data streams, covering information on changes in multiple physical quantities such as temperature, pressure, sound waves and vibrations, which are essential for understanding and optimizing the production process of oil and gas wells. Therefore, it is necessary to conduct in-depth research on the changes in multiple parameters under complex working conditions inside oil and gas wells to improve the understanding and prediction of the dynamic behavior of the reservoir. Accurately capturing and analyzing these dynamic changes will not only help optimize oil and gas production efficiency, but also provide solid data support for safety management and fault warning.

[0003] The accuracy of downhole fiber optic monitoring data analysis is crucial to oil and gas well production. Incorrect data analysis may lead to significant economic losses or even irreversible consequences. At present, most deep-ultra-deep wellbore loss fiber optic data analysis still remains in the single-domain processing stage. This method cannot fully capture the multi-dimensional feature changes under complex working conditions, making it difficult to accurately evaluate and predict the operating status of oil and gas wells. Traditional fiber optic data analysis methods are difficult to efficiently process large-scale data sets, focus on single-dimensional feature extraction, and have poor real-time response capabilities, making it difficult to fully reflect complex downhole dynamic changes.

[0004] Therefore, it is necessary to propose a multi-domain analysis and early warning method for deep and ultra-deep wellbore leakage data, which can efficiently and accurately analyze the optical fiber data under complex oil and gas well conditions, provide more comprehensive and timely feedback and early warning, optimize the production efficiency of deep and ultra-deep oil and gas wells, improve production safety, and avoid economic losses or safety accidents caused by erroneous analysis. Summary of the invention

[0005] In view of this, the present invention provides a multi-domain analysis and early warning method for deep-ultra-deep wellbore optical fiber data, which is used to solve the technical problem that the existing technology for monitoring oil and gas wells often focuses on feature extraction of a single dimension, which is difficult to fully reflect the complex dynamic changes downhole, resulting in safety hazards in the exploitation process of oil and gas wells.

[0006] In order to achieve the above technical objectives, the present invention adopts the following technical solutions:

[0007] In a first aspect, the present invention provides a multi-domain analysis and early warning method for deep-ultra-deep wellbore optical fiber data, comprising:

[0008] Acquire real-time optical fiber monitoring data of deep-ultra-deep wellbore, and filter the real-time optical fiber monitoring data using adaptive filtering technology;

[0009] Normalizing the filtered optical fiber monitoring data to obtain normalized optical fiber data;

[0010] Extracting time domain features, frequency domain features and time-frequency domain features of normalized optical fiber data respectively;

[0011] Based on the time domain characteristics, frequency domain characteristics and time-frequency domain characteristics of normalized optical fiber data, early warning of abnormal conditions in deep and ultra-deep wellbores is provided.

[0012] Furthermore, the real-time optical fiber monitoring data is filtered using adaptive filtering technology, including:

[0013] The filter coefficients are iteratively updated based on the minimum mean square error method until the filter coefficients reach the optimal convergence state.

[0014] Furthermore, the filtered optical fiber monitoring data is normalized to obtain normalized optical fiber data, including:

[0015] The Z-Score method is used to normalize the filtered optical fiber monitoring data, which can be expressed as follows:

[0016]

[0017] Among them, X is the original data, μ is the mean of the data, σ is the standard deviation of the data, X norm is the normalized data.

[0018] Furthermore, the time domain features of the normalized optical fiber data are extracted, including:

[0019] Set the window length and window moving step parameters of the preset window function;

[0020] Using the preset window function to frame the normalized optical fiber data;

[0021] Extract the time domain features of the optical fiber data in each window length frame by frame;

[0022] Summarizing and obtaining the time domain characteristics of the normalized optical fiber data within a sampling time period;

[0023] The time domain features include short-time energy, short-time average amplitude, short-time zero-crossing rate, and short-time average amplitude difference.

[0024] Furthermore, the expression for extracting the short-time energy is:

[0025]

[0026] Where i is the frame number, x is i (n) represents the normalized fiber data, N is the frame length, which indicates the number of sampling points contained in each frame, n is the sampling point index in the window, ranging from 0 to N-1, and E(i) represents the short-time energy.

[0027] Furthermore, the frequency domain features of the normalized optical fiber data are extracted, including:

[0028] The normalized optical fiber data is converted from the time domain to the frequency domain by Fourier transform, and the frequency domain characteristics of the optical fiber data are extracted;

[0029] The frequency domain features include amplitude spectrum, amplitude spectrum density, power spectrum and power spectrum density.

[0030] Furthermore, the time-frequency domain features of the normalized optical fiber data are extracted, including:

[0031] The time-frequency characteristics of normalized optical fiber data are extracted at different time and frequency scales by using wavelet scattering transform, short-time Fourier transform and principal component analysis, respectively, combining time domain and frequency domain characteristics.

[0032] In a second aspect, the present invention provides a multi-domain analysis and early warning device for deep-ultra-deep wellbore optical fiber data, comprising:

[0033] A data acquisition module, used to acquire real-time optical fiber monitoring data of deep-ultra-deep wellbores, and filter the real-time optical fiber monitoring data using adaptive filtering technology;

[0034] A preprocessing module is used to normalize the filtered optical fiber monitoring data to obtain normalized optical fiber data;

[0035] A multi-dimensional feature extraction module, used to extract time domain features, frequency domain features and time-frequency domain features of normalized optical fiber data respectively;

[0036] The early warning module is used to warn of abnormal conditions in deep and ultra-deep wellbores based on the time domain characteristics, frequency domain characteristics and time-frequency domain characteristics of normalized optical fiber data.

[0037] In a third aspect, the present invention provides an electronic device comprising a processor and a memory, wherein the memory stores a computer program, and when the computer program is executed by the processor, a multi-domain analysis and early warning method for deep-ultra-deep wellbore optical fiber data as described in any of the above technical solutions is implemented.

[0038] In a fourth aspect, the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the multi-domain analysis and early warning method of deep-ultra-deep wellbore optical fiber data as described in any of the above technical solutions is implemented.

[0039] Compared with the prior art, the advantages provided by the present invention are: first, the present invention obtains real-time optical fiber monitoring data to monitor changes in temperature, pressure, vibration, etc. in deep-ultra-deep wellbores, thereby providing a data basis for early detection of potential anomalies in the wellbore; the deep-ultra-deep wellbore optical fiber data is filtered through adaptive filtering technology to remove noise interference, improve the quality of the optical fiber data set, and further improve the robustness and stability of the system under various complex conditions; secondly, through normalization processing, optical fiber data of different dimensions and ranges are converted to a unified scale, eliminating the interference of dimensional differences on the analysis results; finally, the time domain, frequency domain and time-frequency domain features of the data are extracted respectively, and based on the characteristic law analysis of each different dimension, the changes in the optical fiber data can be understood in an all-round and multi-level manner, forming a multi-domain analysis and early warning method for optical fiber data. The present invention can identify different types of anomalies by comprehensively analyzing the time domain, frequency domain and time-frequency domain features. Comprehensive analysis based on multiple features can reduce the false positives or omissions that may be caused by single feature analysis, improve the accuracy of the system in complex environments, ensure efficient early warning of wellbore abnormal conditions, and achieve high-precision real-time monitoring and abnormal early warning of downhole data. The method of the present invention can provide powerful technical support for oil and gas field development and open up new ways to improve the efficiency of oil and gas resource exploration and development. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 A schematic diagram of the steps of the multi-domain analysis and early warning method for deep-ultra-deep wellbore optical fiber data provided by the present invention;

[0041] Figure 2 A schematic diagram of the processing flow of the multi-domain analysis and early warning method for deep-ultra-deep wellbore optical fiber data provided by the present invention;

[0042] Figure 3 A comparison diagram of the original signal provided by the present invention and the signal after adaptive filtering and noise reduction;

[0043] Figure 4 A result diagram of the short-time zero-crossing rate analysis of optical fiber data provided by the present invention;

[0044] FIG5( a) is a diagram showing the analysis results of the optical fiber data amplitude spectrum and the amplitude spectrum density provided by the present invention;

[0045] FIG5( b ) is a diagram showing the power spectrum and power spectrum density analysis results of optical fiber data provided by the present invention;

[0046] Figure 6 A schematic diagram of a three-dimensional surface formed by the characteristic values ​​of the principal components and their perception weights provided by the present invention;

[0047] Figure 7 A schematic diagram of the structure of a multi-domain analysis and early warning device for deep-ultra-deep wellbore optical fiber data provided by the present invention;

[0048] Figure 8 The present invention provides a schematic structural diagram of an electronic device. DETAILED DESCRIPTION

[0049] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not used to limit the scope of the present invention.

[0050] See also Figure 1 This embodiment provides a multi-domain analysis and early warning method for deep-ultra-deep wellbore optical fiber data, including:

[0051] Step S101: acquiring real-time optical fiber monitoring data of deep-ultra-deep wellbore, and filtering the real-time optical fiber monitoring data using adaptive filtering technology;

[0052] Step S102: normalizing the filtered optical fiber monitoring data to obtain normalized optical fiber data;

[0053] Step S103: respectively extracting the time domain features, frequency domain features and time-frequency domain features of the normalized optical fiber data;

[0054] Step S104: Based on the time domain characteristics, frequency domain characteristics and time-frequency domain characteristics of the normalized optical fiber data, an early warning is issued for abnormal conditions in deep and ultra-deep wellbores.

[0055] The method of this embodiment uses adaptive filtering technology to process deep-ultra-deep wellbore optical fiber data, removes noise interference, and improves the quality of optical fiber data sets. Based on the optical fiber data after adaptive filtering noise reduction processing, it is normalized to eliminate the interference of dimensional differences on the analysis results, and the time domain, frequency domain and time-frequency domain characteristics of the normalized optical fiber data are extracted respectively. Based on the characteristic law analysis of each different dimension, a multi-domain analysis and early warning method for optical fiber data is formed, which can accurately use the multi-analysis domain characteristic parameters of downhole optical fiber data to evaluate and predict the operating status of oil and gas wells, realize high-precision real-time monitoring and abnormal early warning of downhole data, provide strong technical support for oil and gas field development, and open up new ways to improve the efficiency of oil and gas resource exploration and development.

[0056] like Figure 2 As shown, Figure 2The processing flow diagram of this method is shown, and the specific time domain features, frequency domain features and time-frequency domain features are shown. The Z-Score normalization method is used for normalization; the time domain features include short-time energy, short-time average amplitude, short-time zero-crossing rate, and short-time average amplitude difference; the frequency domain features include amplitude spectrum, amplitude spectrum density, power spectrum and power spectrum density; the analysis of time-frequency domain features uses wavelet scattering transform and short-time Fourier transform principal component analysis method.

[0057] As a preferred embodiment, in step S101, based on the deep-ultra-deep wellbore optical fiber data collected by real-time monitoring, the filter parameters are automatically adjusted using adaptive filtering technology to process it, so as to remove noise interference and improve the quality of deep-ultra-deep wellbore optical fiber monitoring data, and provide a data basis for the subsequent analysis of the multi-domain characteristics of the optical fiber data within the required frequency range.

[0058] The adaptive filter can automatically adjust its coefficients according to the changes of the input signal to minimize a certain performance indicator (such as mean square error). The core idea is to continuously optimize the filter parameters through an iterative algorithm so that the filtered output signal is as close to the expected signal as possible.

[0059] This embodiment uses the LMS algorithm (Least Mean Squares, minimum mean square error algorithm) to automatically adjust the filter parameters for filtering. The specific steps are as follows:

[0060] Step 1: Initialize the filter parameters (the initial value is usually set to zero or a small random value) and select the appropriate order according to the requirements.

[0061] Step 2: Calculate the filter output. The output y(n) of the adaptive filter can be expressed as:

[0062] y(n)=w T (n) x(n),

[0063] Wherein, w(n) is the filter coefficient vector, T represents transpose, and x(n) is the input optical fiber monitoring data.

[0064] Step 3: Calculate the error signal, which is e(n)=d(n)-y(n).

[0065] Wherein, d(n) is the expected signal. In this embodiment, the expected signal refers to a pure signal that is not interfered by noise; y(n) is the filter output signal.

[0066] Step 4: Apply the LMS algorithm to iteratively update the filter coefficients according to the error signal until the filter coefficients converge to the optimal state;

[0067] At each iteration, the updated filter coefficient vector is expressed as:

[0068] w(n+1)=w(n)+μe(n)x(n),

[0069] Among them, μ is the step size parameter, which is used to control the convergence speed and stability.

[0070] As a specific embodiment, in the actual processing process, the step size parameter is set to 0.01 to control the convergence speed of the filter, the LMS adaptive filter order is set to 4, the weight is initialized to a zero vector with a length of 4, and the fiber data is looped through to calculate the error and update the weight, and the filter parameters are automatically adjusted to remove noise interference.

[0071] like Figure 3 As shown, Figure 3 This is a comparison diagram of the original signal provided by the present invention and the signal after adaptive filtering and denoising. After adaptive filtering and denoising, the quality of the deep-ultra-deep well fiber data set can be improved. This embodiment obtains deep and ultra-deep wellbore fiber data in the 0-800Hz frequency band for subsequent analysis.

[0072] As a preferred embodiment, in step S102, the filtered optical fiber monitoring data is normalized to obtain normalized optical fiber data, which specifically includes the following steps:

[0073] The Z-Score method is used to normalize the filtered optical fiber monitoring data, which can be expressed as follows:

[0074]

[0075] Among them, X is the original data, μ is the mean of the data, σ is the standard deviation of the data, X norm is the normalized data.

[0076] As a preferred embodiment, in step S103, extracting the time domain features of the normalized optical fiber data specifically includes:

[0077] Set the window length and window moving step parameters of the preset window function;

[0078] Using the preset window function to frame the normalized optical fiber data;

[0079] Extract the time domain features of the optical fiber data in each window length frame by frame;

[0080] Summarizing and obtaining the time domain characteristics of the normalized optical fiber data within a sampling time period;

[0081] The time domain features include short-time energy, short-time average amplitude, short-time zero-crossing rate, and short-time average amplitude difference.

[0082] It should be noted here that the window functions that can be selected in this step include rectangular windows, Hamming windows, Hening windows and other types. Each window function has different spectral characteristics. Selecting a suitable window function helps to reduce spectrum leakage and improve analysis accuracy. In practical applications, a suitable window function is selected according to specific needs to improve the quality and reliability of the analysis results. The window length is usually set according to the sampling frequency of the signal and the accuracy of the analysis. A longer window length will have better frequency resolution, but will reduce the time domain resolution. The step size usually refers to the sliding amount of the window in the time domain, which affects the accuracy of the frame. The data is divided into multiple frames according to the set step size. Each frame contains the optical fiber data within a specific time period, which is convenient for time domain feature extraction of each frame of data.

[0083] As a specific embodiment, a Hamming window is selected to frame the deep and ultra-deep wellbore optical fiber data after Z-Score normalization, the window length is set to 512 sampling points, and the window moving step is set to 50% of the window size.

[0084] As a preferred embodiment, short-time energy can measure the energy of a signal in a short period of time, and is used to reflect the strength or fluctuation amplitude of the signal. In wellbore monitoring, sudden changes in energy may indicate abnormal events in the equipment or wellbore (such as equipment failure, abnormal vibration, etc.). By monitoring the changes in short-time energy, energy anomalies can be identified in a timely manner, thereby predicting possible failures or abnormal conditions. The short-time energy expression in the time domain feature is:

[0085]

[0086] Among them, i is the frame number, x is i (n) represents the normalized fiber data, N is the frame length, which indicates the number of sampling points contained in each frame, n is the sampling point index in the window, ranging from 0 to N-1, and E(i) represents the short-time energy.

[0087] The short-time average amplitude is used to measure the amplitude of the signal and describe the amplitude of the signal change. If the signal amplitude suddenly increases, it may indicate abnormal phenomena such as pressure fluctuations and liquid level changes in the wellbore; monitoring this time domain feature helps to discover abnormal fluctuations in the short term and identify potential abnormal events. The short-time average amplitude expression in the time domain feature is:

[0088]

[0089] Among them, i is the frame number, x is i (n) represents the sound wave signal, N is the frame length, which indicates the number of sampling points contained in each frame, n is the sampling point index in the window, ranging from 0 to N-1, and M(i) represents the average amplitude.

[0090] The zero-crossing rate refers to the number of times a signal passes through the zero point per unit time, reflecting the frequency change or volatility of the signal. In wellbore vibration monitoring, the change in the zero-crossing rate indicates that the vibration mode has changed. If the zero-crossing rate of a certain signal suddenly increases or decreases, it may indicate a mechanical failure or structural problem. The expression of the short-term zero-crossing rate in the time domain feature is:

[0091]

[0092] Among them, i is the frame number, x is i (n) represents the acoustic signal, N is the frame length, which indicates the number of sampling points contained in each frame, n is the sampling point index in the window, ranging from 0 to N-1, Z(i) represents the short-time zero-crossing rate, and sgn[·] is the sign function.

[0093] The short-time average amplitude difference describes the amplitude difference between two adjacent signal segments, reflecting the severity of the signal fluctuation. If the amplitude difference increases abnormally, it indicates that the signal has changed significantly, which may be caused by a sudden abnormal event. For example, changes in pressure or liquid level in the wellbore may cause the amplitude difference to increase. The calculation formula for the short-time average amplitude difference in the time domain feature is:

[0094]

[0095] Among them, i is the frame number, x is i (n) represents the sound wave signal, N is the frame length, which indicates the number of sampling points contained in each frame, n is the sampling point index in the window, ranging from 0 to N-1, k is the maximum delay point number, and Z(i) represents the short-time average amplitude difference.

[0096] By analyzing and extracting the time domain characteristics of optical fiber data, we can obtain the time domain laws reflected by the downhole optical fiber data and identify abnormal situations from multiple angles such as signal amplitude, energy and volatility. By monitoring these characteristics, we can timely warn of potential faults in the wellbore from the time domain dimension and avoid safety accidents caused by failure to discover them in time.

[0097] The specific analysis method of time domain characteristics is described below.

[0098] See also Figure 4 , as a specific example, Figure 4 This is the result of the short-time zero-crossing rate analysis of optical fiber data. Under the condition of a wellbore inclination of 15°, the dispersion characteristics of bubbles in the liquid medium and their weak influence by gravity work together to promote the stability of the flow state, which is manifested as a reduction in the fluctuation amplitude of the bubbly flow. Figure 4Analysis shows that the low instantaneous zero-crossing rate in a specific area can be attributed to the local sparseness or uneven distribution of bubble density, which limits the intensity of signal changes and is insufficient to frequently trigger the change of signal signs. It is worth noting that although a single bubble has little effect on the signal, the cumulative effect of the bubble group in the bubbly flow is significant, which overall causes the signal to frequently cross the zero point, causing the short-term zero-crossing rate of the bubbly flow to exceed that of other flow types.

[0099] As a preferred embodiment, in step S103, extracting the frequency domain features of the normalized optical fiber data includes:

[0100] The normalized optical fiber data is converted from the time domain to the frequency domain by Fourier transform, and the frequency domain characteristics of the optical fiber data are extracted;

[0101] The frequency domain features include amplitude spectrum, amplitude spectrum density, power spectrum and power spectrum density.

[0102] Among them, the amplitude spectrum is mainly used to identify the most significant frequency components in the signal, reflecting the periodic changes and vibration patterns of the equipment. For example, the normal vibration frequency of downhole equipment may appear as some prominent frequencies in the amplitude spectrum, which can be used to confirm the normal operation of the equipment. Amplitude spectral density (ASD) helps to identify noise levels, abnormal vibrations or high-frequency components by evaluating the amplitude energy distribution of each frequency. In the monitoring of deep and ultra-deep wells, by analyzing the amplitude spectrum density, it can be determined whether the noise is too high, and then whether there is external interference or equipment abnormality. The power spectrum can help quantify the energy distribution of the signal at each frequency and is used to identify high-energy frequencies and low-energy frequencies. If the power spectrum shows a sudden increase or change in the power of certain frequencies, it may mean that the equipment is abnormal or damaged. For example, an abnormal increase in certain frequency components may represent excessive wear of the equipment or loose parts. The power spectrum density can reflect the trend of signal power changes over time. If the power spectrum density changes abnormally, such as a sudden increase in the power of certain frequency bandwidths, it may be an early sign of equipment failure.

[0103] As a specific embodiment, the data of the best channel position of the signal response on the optical fiber is selected for subsequent processing. Combined with Figure 5(a) and Figure 5(b), Figure 5(a) and Figure 5(b) respectively show the amplitude spectrum, amplitude spectrum density, power spectrum and power spectrum density analysis results of the optical fiber data. Under the condition of a wellbore inclination of 15°, there are obvious amplitude difference characteristics between slug flow, bubbly flow and annular flow, which is specifically manifested as the slug flow has the largest amplitude value, followed by the bubbly flow, and the smallest is the annular flow; the power relationship shows that the slug flow occupies a dominant position, and its power value decreases successively to the bubbly flow. However, as the wellbore inclination increases to 30 degrees, this amplitude distribution pattern changes. At this time, the slug flow still maintains its maximum amplitude position unchanged, but the secondary position changes from the bubbly flow to the annular flow. The power distribution characteristics have not changed, and the order of slug flow is still maintained, followed by annular flow, and bubbly flow is the smallest.

[0104] As a preferred embodiment, in step S103, extracting the time-frequency domain features of the normalized optical fiber data includes:

[0105] The time-frequency characteristics of normalized optical fiber data are extracted at different time and frequency scales by using wavelet scattering transform, short-time Fourier transform and principal component analysis, respectively, combining time domain and frequency domain characteristics.

[0106] As a specific embodiment, two extraction methods are used in the time-frequency domain analysis, namely, wavelet scattering transform and short-time Fourier transform method fused with principal component analysis, wherein principal component analysis is used to reduce the dimension of the data after short-time Fourier transform. Based on historical data and expert knowledge, one or more thresholds are set to distinguish between normal status and potential leakage. When certain wavelet coefficients exceed the preset threshold, an alarm may be triggered. At the same time, when certain principal component eigenvalues ​​or eigenvectors deviate significantly from the normal range within a certain time period, an alarm may be triggered, indicating that there may be wellbore leakage or other abnormal conditions. Since the DAS system provides spatial resolution along the optical fiber, the problem area can be further accurately located according to the specific location where the anomaly occurs.

[0107] For example, wellbore equipment may cause periodic or non-periodic fault vibrations during certain operations, and these sudden frequency changes can be captured in real time through STFT. Wavelet transform decomposes the signal through scaling functions of different frequencies, which can accurately capture the local changes of the signal, and is particularly suitable for analyzing sudden, nonlinear wellbore anomalies. Wavelet transform can efficiently capture sudden pressure fluctuations, changes in liquid flow, or equipment failure signals in the well. When wavelet transform captures high-frequency short-term burst signals in the wellbore signal, the system can promptly identify these changes and issue warnings, indicating possible equipment failures or environmental abnormalities.

[0108] Specifically, db1 (the first member of the Daubechies wavelet family) is used to perform wavelet decomposition on the data, with a decomposition level of 3, to obtain wavelet coefficients at different levels. Short-time Fourier transform (STFT) is applied to the acoustic data to divide the acoustic signal into small parts so that they can overlap and be windowed. Then, DFT calculations are performed on each part separately, and principal component analysis (PCA) realizes the linear mapping from the original high-dimensional space to the low-dimensional space by calculating the main eigenvectors of the covariance matrix of the signal set, while maximizing the retention of the variance information of the data, effectively removing irrelevant noise, and improving data processing efficiency.

[0109] See also Figure 6 , Figure 6 The three-dimensional surface diagram composed of the eigenvalues ​​and perceptual weights of the first 50 principal components provided in this embodiment. The first principal component has a significantly higher eigenvalue than other principal components, and its corresponding eigenvector carries the richest information. Therefore, we select the eigenvector of the first principal component as the key feature.

[0110] As a specific embodiment, in step S104, based on the time domain characteristics, frequency domain characteristics and time-frequency domain characteristics of the normalized optical fiber data, an early warning is given for abnormal conditions in deep-ultra-deep wellbores, including:

[0111] According to the changing trend of the time domain characteristics, when the changing trend of the time domain characteristics exceeds a first preset threshold, an initial warning information is issued;

[0112] Analyze the frequency component changes in the frequency domain characteristics, and when the frequency offset exceeds a second preset threshold, issue an alarm to preliminarily confirm the warning information;

[0113] The time period in which the time domain feature exceeds the first preset threshold is used as the analysis window, anomaly detection is performed on the time-frequency feature at each moment in the analysis window, and final abnormal condition warning information is issued according to the detection result.

[0114] By integrating the advantages of different analysis techniques, it provides a more comprehensive and accurate fault warning capability than a single analysis method. Time domain feature analysis mainly starts from the change of the signal in time, focusing on the amplitude, waveform, energy change of the signal, etc., which can reflect the overall trend of the wellbore state. These features are usually used to capture time-sensitive emergencies. Frequency domain feature analysis can reveal the changes in periodic components, characteristic frequencies and frequency amplitudes in the signal by converting the time domain signal to the frequency domain. It is mainly used to analyze the operating status and vibration mode of equipment in the wellbore. The time-frequency domain feature combines the advantages of the time domain and the frequency domain, and can analyze the changes of the signal in the dual dimensions of time and frequency, which is particularly suitable for capturing non-stationary signals or emergencies. Time-frequency domain analysis can capture the occurrence of anomalies faster by being able to perform data analysis at a fine time and frequency level. Set thresholds based on historical data or equipment operating parameters, and then set different alarm levels based on the intensity and duration of abnormal signals. When certain features in the time domain, frequency domain or time-frequency domain exceed the set range, the system will immediately trigger an alarm. By combining the time domain, frequency domain and time-frequency domain characteristics, the early warning mechanism can reflect the operating status of the wellbore more comprehensively and accurately, detect potential problems in a timely manner, and respond to them, thereby greatly improving the safety and stability of the wellbore.

[0115] The embodiment of the present invention also provides a multi-domain analysis and early warning device for deep-ultra-deep wellbore optical fiber data, such as Figure 7 As shown, the multi-domain analysis and early warning device 700 for deep-ultra-deep wellbore optical fiber data includes:

[0116] The data acquisition module 701 is used to acquire the real-time optical fiber monitoring data of the deep-ultra-deep wellbore and filter the real-time optical fiber monitoring data using adaptive filtering technology;

[0117] A preprocessing module 702 is used to normalize the filtered optical fiber monitoring data to obtain normalized optical fiber data;

[0118] A multi-dimensional feature extraction module 703, used to extract time domain features, frequency domain features and time-frequency domain features of normalized optical fiber data respectively;

[0119] The early warning module 704 is used to warn of abnormal conditions in deep-ultra-deep wellbores based on the time domain characteristics, frequency domain characteristics and time-frequency domain characteristics of the normalized optical fiber data.

[0120] like Figure 8 As shown, the multi-domain analysis and early warning method of deep-ultra-deep wellbore optical fiber data described above, the present invention also provides an electronic device 800 accordingly, the electronic device can be a computing device such as a mobile terminal, a desktop computer, a notebook, a palm computer and a server. The electronic device includes a processor 801, a memory 802 and a display 803.

[0121] In some embodiments, the memory 802 may be an internal storage unit of a computer device, such as a hard disk or memory of a computer device. In other embodiments, the memory 802 may also be an external storage device of a computer device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), etc. equipped on the computer device. Further, the memory 802 may also include both an internal storage unit of a computer device and an external storage device. The memory 802 is used to store application software and various types of data installed on the computer device, such as program codes for installing the computer device. The memory 802 may also be used to temporarily store data that has been output or is to be output. In one embodiment, a multi-domain analysis and early warning method program 804 of deep-ultra-deep wellbore optical fiber data is stored on the memory 802, and the multi-domain analysis and early warning method program 804 of deep-ultra-deep wellbore optical fiber data can be executed by the processor 801, thereby realizing a multi-domain analysis and early warning method of deep-ultra-deep wellbore optical fiber data in each embodiment of the present invention.

[0122] In some embodiments, the processor 801 can be a central processing unit (CPU), a microprocessor or other data processing chip, which is used to run the program code or process data stored in the memory 802, such as executing a multi-domain analysis and early warning method program for deep-ultra-deep wellbore optical fiber data.

[0123] In some embodiments, the display 803 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch device, etc. The display 803 is used to display information on the computer device and to display a visual user interface. The components 801-803 of the computer device communicate with each other through a system bus.

[0124] This embodiment also provides a computer-readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, the multi-domain analysis and early warning method of deep-ultra-deep wellbore optical fiber data described in any of the above technical solutions is implemented.

[0125] The computer-readable storage medium and computing device provided according to the above-mentioned embodiments of the present invention can be implemented with reference to the contents specifically described in the method for multi-domain analysis and early warning of deep-ultra-deep wellbore optical fiber data according to the present invention, and have similar beneficial effects as the method for multi-domain analysis and early warning of deep-ultra-deep wellbore optical fiber data as described above, which will not be described in detail here.

[0126] In summary, the present invention provides a multi-domain analysis and early warning method for deep-ultra-deep wellbore optical fiber data, which extracts the time domain, frequency domain, and time-frequency domain characteristics of the data respectively, analyzes the laws of the characteristics of each domain, and forms a multi-domain analysis and early warning method for deep and ultra-deep wellbore leakage optical fiber data. The method can accurately use the multi-analysis domain characteristic parameters of downhole optical fiber data to evaluate and predict the operating status of oil and gas wells, realize high-precision real-time monitoring and abnormal early warning of downhole data, provide strong technical support for oil and gas field development, and open up new ways to improve the efficiency of oil and gas resource exploration and development.

[0127] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A multi-domain analysis and early warning method for deep-ultra-deep wellbore optical fiber data, characterized in that: include: Acquire real-time optical fiber monitoring data of deep-ultra-deep wellbore, and filter the real-time optical fiber monitoring data using adaptive filtering technology; Normalizing the filtered optical fiber monitoring data to obtain normalized optical fiber data; Extracting time domain features, frequency domain features and time-frequency domain features of normalized optical fiber data respectively; Based on the time domain characteristics, frequency domain characteristics and time-frequency domain characteristics of normalized optical fiber data, early warning of abnormal conditions in deep and ultra-deep wellbores is provided.

2. The multi-domain analysis and early warning method for deep-ultra-deep wellbore optical fiber data according to claim 1, characterized in that: Adaptive filtering technology is used to filter real-time fiber monitoring data, including: The filter coefficients are iteratively updated based on the minimum mean square error method until the filter coefficients reach the optimal convergence state.

3. The multi-domain analysis and early warning method for deep-ultra-deep wellbore optical fiber data according to claim 1, characterized in that: The filtered optical fiber monitoring data is normalized to obtain normalized optical fiber data, including: The Z-Score method is used to normalize the filtered optical fiber monitoring data, which can be expressed as follows: Among them, X is the original data, μ is the mean of the data, σ is the standard deviation of the data, X norm is the normalized data.

4. The multi-domain analysis and early warning method for deep-ultra-deep wellbore optical fiber data according to claim 1, characterized in that: Extract time domain features of normalized fiber data, including: Set the window length and window moving step parameters of the preset window function; Using the preset window function to frame the normalized optical fiber data; Extract the time domain features of the optical fiber data in each window length frame by frame; Summarizing and obtaining the time domain characteristics of the normalized optical fiber data within a sampling time period; The time domain features include short-time energy, short-time average amplitude, short-time zero-crossing rate, and short-time average amplitude difference.

5. The multi-domain analysis and early warning method for deep-ultra-deep wellbore optical fiber data according to claim 4, characterized in that: The expression for extracting the short-time energy is: Where i is the frame number, x is i (n) represents the normalized fiber data, N is the frame length, which indicates the number of sampling points contained in each frame, n is the sampling point index in the window, ranging from 0 to N-1, and E(i) represents the short-time energy.

6. The multi-domain analysis and early warning method for deep-ultra-deep wellbore optical fiber data according to claim 1, characterized in that: Extract the frequency domain features of the normalized fiber data, including: The normalized optical fiber data is converted from the time domain to the frequency domain by Fourier transform, and the frequency domain characteristics of the optical fiber data are extracted; The frequency domain features include amplitude spectrum, amplitude spectrum density, power spectrum and power spectrum density.

7. The multi-domain analysis and early warning method for deep-ultra-deep wellbore optical fiber data according to claim 1, characterized in that: Extract time-frequency domain features of normalized fiber data, including: Wavelet scattering transform, short-time Fourier transform and principal component analysis are used respectively, and the time domain characteristics and frequency domain characteristics of the normalized optical fiber data are combined to obtain the time-frequency characteristics of the normalized optical fiber data on different time and frequency scales.

8. A multi-domain analysis and early warning device for deep-ultra-deep wellbore optical fiber data, characterized in that: include: A data acquisition module, used to acquire real-time optical fiber monitoring data of deep-ultra-deep wellbores, and filter the real-time optical fiber monitoring data using adaptive filtering technology; A preprocessing module is used to normalize the filtered optical fiber monitoring data to obtain normalized optical fiber data; A multi-dimensional feature extraction module, used to extract time domain features, frequency domain features and time-frequency domain features of normalized optical fiber data respectively; The early warning module is used to warn of abnormal conditions in deep and ultra-deep wellbores based on the time domain characteristics, frequency domain characteristics and time-frequency domain characteristics of normalized optical fiber data.

9. An electronic device, characterized in that: The method comprises a processor and a memory, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the multi-domain analysis and early warning method of deep-ultra-deep wellbore optical fiber data as described in any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by the processor, it implements a multi-domain analysis and early warning method for deep-ultra-deep wellbore optical fiber data as described in any one of claims 1-7.

Images