An optical cable fault detection method based on industrial data processing

Abstract

The application relates to the technical field of optical cable fault detection, and discloses an optical cable fault detection method based on industrial data processing, which comprises the following steps: collecting fiber splicing point data of optical fibers in an optical cable; judging whether the parameters of the fiber splicing points are normal; accurately verifying the parameters of the optical fibers in the optical cable; transmitting data or adjusting the transmitted data or prompting the abnormal parameters of the fiber splicing points; the optical cable fault detection method based on industrial data processing judges whether the parameters of the fiber splicing points are abnormal, generates a simulation model according to different interference conditions and different optical fiber conditions, judges whether the collected parameters of the fiber splicing points are accurate, automatically restores to the real data when the collected parameters are interfered, automatically adjusts the data to be transmitted when the parameters of the fiber splicing points are abnormal, reduces the error condition of data transmission, avoids causing the loss of users, improves the detection accuracy, and improves the detection efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of optical cable fault detection, and in particular to an optical cable fault detection method based on industrial data processing. Background Art

[0002] With the rapid development of communication networks, optical cables, as a key transmission medium, have become particularly important for maintenance and fault detection. Traditional optical cable fault detection relies heavily on on-site inspections and manual troubleshooting, which are costly, inefficient, and difficult to achieve real-time monitoring. With the continuous improvement of industrial information technology, the use of industrial data (such as data collected by sensors, network transmission status information, etc.) for optical cable fault detection has become a trend. The optical cable fault detection method based on industrial data processing uses optical cable-related data collected in industrial environments to achieve rapid and accurate detection and positioning of optical cable faults through data analysis, feature extraction, and intelligent algorithms. This aims to improve optical cable maintenance efficiency, reduce power outages, and ensure the reliability of communication networks. By collecting and analyzing multi-source data in industrial environments and using advanced data processing and intelligent algorithms, it can achieve early detection and precise positioning of optical cable faults, providing a safe and efficient maintenance method for the communications industry.

[0003] The existing optical cable fault detection method based on industrial data processing cannot determine whether there are abnormalities in the parameters of the optical fiber connection point based on the OTDR parameters of the optical fiber in the optical cable, cannot generate simulation models based on different interference conditions and different optical fiber conditions, and cannot determine whether the collected parameters of the optical fiber connection point are accurate based on the simulation model. When there is interference in the collected parameters, it cannot automatically recover to the real data. When there are abnormalities in the parameters of the optical fiber connection point, it cannot automatically adjust the data to be transmitted, which can easily lead to errors in data transmission, thereby causing certain losses and reducing the accuracy of detection. Its practicality has certain limitations. Summary of the Invention

[0004] The present invention provides an optical cable fault detection method based on industrial data processing, which is used to promote the solution of the problems mentioned in the background technology.

[0005] The present invention provides the following technical solution: an optical cable fault detection method based on industrial data processing, comprising:

[0006] Collect data on the fiber splicing points of the optical fibers in the optical cable;

[0007] For each fiber splice, select a smooth OTDR curve on the fiber segments before and after the splice.

[0008] Calculate the loss of optical fiber splicing points:

[0009] Among them, P beforeis the optical power before the connection point, P after is the optical power after the connection point;

[0010] For each fiber splice, measure the optical power of the reflection peak at the splice location and record it as P refl ;

[0011] Calculate the reflectivity of the optical fiber splice point:

[0012] Among them, P incident is the incident light power;

[0013] Set the loss threshold and reflectivity threshold, respectively denoted as Loss threshold and Reflectivity threshold ;

[0014] Set a parameter judgment function to determine whether the parameters of the optical fiber splicing point are normal:

[0015]

[0016] If F j (Loss co_p , Reflectivity co_p )=True, it is determined that the parameter detection of the optical fiber connection point is normal;

[0017] If F j (Loss co_p , Reflectivity co_p )=False, then it is determined that the parameter detection of the optical fiber connection point is abnormal;

[0018] Verify the accuracy of the parameters of the optical fiber in the cable through the detection model;

[0019] If the parameter detection of the optical fiber connection point is normal and the data collection is accurate, data transmission will proceed normally;

[0020] If the parameter detection of the optical fiber connection point is abnormal and the data collection is accurate, the transmitted data is adjusted according to the signal adjustment model so that the received data is consistent with the sent data before the adjustment;

[0021] If the data collection is wrong, it will prompt that the parameters of the optical fiber connection point are abnormal and data transmission will not be performed.

[0022] As an optional solution to the optical cable fault detection method based on industrial data processing of the present invention, wherein: the optical fiber connection point data of the optical fiber in the optical cable is collected, specifically:

[0023] Set OTDR parameters, including wavelength λ, pulse width τ, and average times N avgAnd the distance range L range_dis ;

[0024] Collect backscatter signal curve:

[0025] BS(x)=P(x)(x=1,2,...,N);

[0026] Where x is the distance from the OTDR starting point, and P(x) is the optical power at distance x.

[0027] Identify reflection peak positions:

[0028] PO peak_refl ={x i |P(x i )>P(x i -Δx)andP(x i )>P(x i +Δx)};

[0029] Among them, Δx is used to determine the window size of the local extreme value;

[0030] Identify the loss step location:

[0031] PO step_loss ={x i |P(x i )<P(x i -Δx)andP(x i )<P(x i +Δx)};

[0032] Among them, Δx is used to determine the window size of the local extreme value;

[0033] Record the connection point location:

[0034] PO point_con ={x joint，1 , x joint，2 ,...,x joint，m}.

[0035] As an optional solution to the optical cable fault detection method based on industrial data processing of the present invention, wherein: for each optical fiber splicing point, a smooth OTDR curve is selected on the optical fiber segments before and after the splicing point position, specifically:

[0036] For each fiber splice, the location of the splice is denoted by x. joint ;

[0037] Set a length, record it as L ref ;

[0038] Take a distance range before and after the connection point as the reference interval, and the length of the distance range is L ref ;

[0039] The reference interval before the connection point is

[0040] The reference interval after the connection point is

[0041] Extracting optical power data of a reference interval from the backscattered signal curve BS(x)=P(x);

[0042] Among them, the reference curve data before the connection point is

[0043] The reference curve data after the connection point is

[0044] Smooth the extracted reference curve data:

[0045]

[0046] Where Δx is the distance between data points on the curve, and k is the half-window size of the moving average;

[0047] Compute the first derivative of the smoothed curve:

[0048]

[0049] Determine the smoothness of the curve:

[0050]

[0051] Among them, ε smooth is the smoothness threshold for verifying the smoothness of the curve;

[0052] If F smooth (P' smooth (x)) = 1, the curve is judged to be smooth;

[0053] If F smooth (P' smooth If (x))=0, the curve is judged to be not smooth.

[0054] As an optional solution of the optical cable fault detection method based on industrial data processing of the present invention, the detection model is specifically:

[0055] Get the length of the optical fiber, recorded as L current ;

[0056] Set the collection interval, denoted as d;

[0057] According to the collection interval d, several data collection points are set on the optical fiber;

[0058] Calculate the position of each data collection point:

[0059] x i =i·d(i=1,2,...,n);

[0060] Among them, x i represents the position of the i-th data collection point, n is the number of data collection points, and satisfies:

[0061]

[0062] in, Indicates a round-down operation;

[0063] like Then the position of the last data collection point is:

[0064] x n =L current ;

[0065] At this time, the corrected length of the last interval of the actual collection point interval is:

[0066] L current -x n-1 ;

[0067] Perform interference analysis.

[0068] As an optional solution of the optical cable fault detection method based on industrial data processing of the present invention, the interference analysis is specifically:

[0069] S1. For the i-th data collection point on the optical fiber, define its interference condition vector as:

[0070] C i =[C i1 , C i2 ,...,C ik ];

[0071] Where k is the number of different test conditions, C ij It represents the interference situation of the i-th data collection point under the j-th test condition, C ij =1 means there is interference, C ij =0 means no interference;

[0072] S2. For all n data collection points, define the interference condition vector set as:

[0073] C=[C1,C2,...,C n ];

[0074] S3. Calculate the intersection of the interference condition vectors of all data collection points:

[0075]

[0076] S4. Set a consistency judgment function to determine whether the interference conditions of all data collection points are consistent:

[0077]

[0078] Where j = 1, 2, ..., k, I = C j Indicates that all data collection points point to the jth interference condition, and I = / 0 indicates no interference condition;

[0079] S5, if F consistency (I) = 1, the data collection is judged to be accurate;

[0080] If F consistency (I) = 0, then take the fiber connection point as the origin and set an initial range, denoted as R0;

[0081] S6. Obtain the interference condition vector set of all data collection points within the range:

[0082]

[0083] Where m is the number of data collection points within the range;

[0084] S7. Calculate the intersection of the interference condition vectors of the data collection points within the range:

[0085]

[0086] S8. Use the range determination function to determine whether the interference conditions of the data collection points within the range are consistent:

[0087]

[0088] Among them, I range =C j Indicates that all data collection points within the range point to the jth interference condition, I range = / 0 means there is no interference condition within the range;

[0089] S9, if F range (I range )=1, then the data collection is determined to be accurate;

[0090] If F range (I range )=0, then obtain the range reduction model;

[0091] S10, substituting the initial range R0 into the range reduction model to generate a new range, and marking the new range as R0;

[0092] S11, repeat S6-S12, set the range threshold, record it as R threshold ;

[0093] Substitute the re-evaluated value into the loop evaluation function:

[0094]

[0095] Among them, STOP is the sign to stop the cycle, and CONTINUE is the sign to continue the cycle;

[0096] S12, if F cycle (I range )=STOP2, then stop the loop and determine that the data collection error;

[0097] If F cycle (I range )=STOP1, then stop the loop and determine that the data collection is accurate;

[0098] If F cycle (I range )=CONTINUE, then the loop continues and S6-S12 are repeatedly executed.

[0099] As an optional solution of the optical cable fault detection method based on industrial data processing of the present invention, the range reduction model is specifically:

[0100] Get the current range, recorded as R t ;

[0101] Taking the fiber splicing point as the origin, calculate the new range:

[0102]

[0103] Among them, R t+1 For the narrowed scope;

[0104] Get the current interval distance, recorded as d t ;

[0105] Calculate the new separation distance:

[0106]

[0107] Among them, d t+1 The shortened separation distance;

[0108] Then the number of data collection points in the new range is:

[0109]

[0110] The location of each new data collection point is:

[0111]

[0112] Among them, x center The center position of the current range.

[0113] As an optional solution to the optical cable fault detection method based on industrial data processing of the present invention, the detection model further includes troubleshooting the interference cause and determining and updating the final optical fiber connection point data, specifically:

[0114] like or This means that the collected data is free of interference and can be used directly;

[0115] If I range =C j or I=C j , it means that the collected data is interfered with, collect the deviation data set D deviation ={(C i , ΔP i ) | i=1, 2, ..., m};

[0116] Among them, C i Represents the interference condition vector of the i-th sample, specifically:

[0117] C i =[C i1 , C i2 ,...,C ik ];

[0118] ΔP i is the parameter deviation vector of the optical fiber splicing point of the i-th sample, which represents the difference between the parameters under interference conditions and the actual parameters. Specifically:

[0119]

[0120] Train a perturbation model:

[0121] f(C)=β0+β1C i1 +β2C i2 +...+β k C ik ;

[0122] Among them, β0, β1, ..., β k are model parameters, estimated by minimizing the loss function:

[0123]

[0124] Get the current fiber connection point parameters, recorded as P measured ;

[0125] Extract the predicted deviation, denoted as ΔP predicted =f(C current );

[0126] Among them, C current is the current interference condition vector;

[0127] Calculate the actual parameters of the fiber splice point:

[0128] P true =P measured -ΔP predicted ;

[0129] Recalculate the loss and reflectivity of optical fiber splices;

[0130] Use parameter determination function to judge whether the parameters of the optical fiber splicing point are normal;

[0131] If F j (Loss co_p , Reflectivity co_p )=True, it is determined that the parameter detection of the optical fiber connection point is normal;

[0132] If F j (Loss co_p , Reflectivity co_p )=False, it is determined that the parameter detection of the optical fiber connection point is abnormal.

[0133] As an optional solution of the optical cable fault detection method based on industrial data processing of the present invention, the signal adjustment model is specifically:

[0134] Get all the input data to form an input data set, denoted as D in ={D in，1 , D in，2 ,...,D in，m};

[0135] Among them, each D in，i Represents a specific set of input data;

[0136] Get different fiber line lengths to form a line data set, denoted as L line ={L line，1 , L line，2 ,...,L line，l};

[0137] Among them, each Lline，h Indicates a specific line length;

[0138] Obtain different numbers of fiber splicing points to form a splicing point data set, denoted as N = {N1, N2, ..., N k};

[0139] Among them, each N j Indicates the number of optical fiber splicing points;

[0140] For each fiber connection point number, obtain different fiber connection point parameters for each fiber connection point, denoted as P n ={P n，1 , P n，2 ,...,P n，p};

[0141] Among them, each P n，q Indicates the qth parameter;

[0142] For each fiber connection point parameter, obtain its different parameter values, recorded as V n，p ={V n，p，1 , V n，p，2 ,...,V n，p，s};

[0143] Among them, each V n，p，r Indicates a possible value for the parameter;

[0144] For each combination (D in，i , N j , P n，q =V n，p，r , L line，h ), use the optical fiber transmission model to simulate and get the actual output data, denoted as D out , and generate a training data set, denoted as D:

[0145]

[0146] Define an adjustment amount, denoted as ΔD in ;

[0147] For each input data D in，i , adjustment amount ΔD in，i satisfy:

[0148] D out (D in，i +ΔD in，i , N j , P n，q =V n，p，r , L line，h )=D in，i ;

[0149] Use the gradient descent method to optimize and calculate the updated adjustment amount ΔD in，i :

[0150]

[0151] Where t is the number of iterations, is the adjustment amount at the tth iteration, η is the learning rate, Calculate the difference function J versus ΔD in，i The gradient at iteration t, Represents the gradient operator, calculating the difference function J for the adjustment amount ΔD in，i Gradient:

[0152]

[0153] Where d is the adjustment amount ΔD in，i The dimension of the difference function J is the actual output data D out And the expected output data, that is, the input data D in，i The function of the difference between them is:

[0154] J(ΔD in，i )=||D out (D in，i +ΔD in，i , N j , P n，q =V n，p，r , L line，h )-D in，i || 2 ;

[0155] Set the maximum number of iterations, denoted as T;

[0156] Set a given threshold, denoted as ε;

[0157] Set an optimization judgment function to judge the updated adjustment amount ΔD in，i Is it suitable?

[0158]

[0159] If F optimize (ΔD in，i )=1, then the iterative optimization is stopped and the adjustment amount ΔD obtained after iterative optimization is in，i Integrate into the training dataset D;

[0160] If F optimize (ΔD in，i )=0, then continue iterative optimization.

[0161] As an optional solution of the optical cable fault detection method based on industrial data processing of the present invention, wherein: the transmitted data is adjusted according to the signal adjustment model so that the received data is consistent with the sent data before the adjustment, specifically:

[0162] Get the fiber optic line length of the current line, recorded as L line_current ;

[0163] Get the input data that needs to be transmitted, denoted as D in_current ;

[0164] Get the number of fiber connection points on the current line, recorded as N current ;

[0165] Obtain the optical fiber connection point parameters of each optical fiber connection point and the parameter value of each optical fiber connection point parameter, which is recorded as {(P n，c , V n，c，c_c )|P n，c =V n，c，c_c};

[0166] Integration generates the transmission data set, denoted as (D in_current , N current , P n，c =V n，c，c_c , L line_current );

[0167] Extract the corresponding adjustment amount of the transmission data set in the signal adjustment model, recorded as ΔD in_current ;

[0168] Calculate the final input data D in_final :

[0169] D in_final =D in_current +ΔD in_current ;

[0170] Input data D in_current Adjust to the final input data D in_final Then data transmission is performed.

[0171] The present invention has the following beneficial effects:

[0172] 1. This optical cable fault detection method based on industrial data processing uses the OTDR method to collect OTDR parameters of the optical fiber in the optical cable, determine the loss and reflectivity of the optical fiber connection point, judge whether there are abnormalities in the parameters of the optical fiber connection point, and perform preliminary detection and analysis on the optical fiber in the optical cable, thereby reducing errors in data transmission, avoiding losses to users, improving detection accuracy, and improving detection efficiency.

[0173] 2. The optical cable fault detection method based on industrial data processing generates a detection model by simulating different interference conditions and different optical fiber conditions, and sets a number of data collection points on the light. According to the data collected by the data collection points within a certain range of the optical fiber connection point, it is judged whether the data all point to the same interference condition or all point to the non-interference condition. When the data collected by the data collection points point to different interference conditions, the judgment range is continuously narrowed, and the interval distance between the data collection points is continuously shortened, and the judgment is continued to determine whether the data all point to the same interference condition or all point to the non-interference condition, that is, whether the parameters of the collected optical fiber connection point are accurate. If the data all point to the same interference condition or all point to the non-interference condition before the range is narrowed to the threshold, it means that the data is accurate. If the data collected by the data collection points still point to different interference conditions until the range is narrowed to the threshold, it means that the data is inaccurate. At this time, a data collection error is prompted, which reduces the error in data transmission, avoids user losses, improves detection accuracy, and improves detection efficiency.

[0174] 3. This optical cable fault detection method based on industrial data processing, when the data collection is determined to be accurate, if there is no interference in the collected parameters, the data transmission is carried out normally; if there is interference in the collected parameters, it automatically recovers to the real data; when the data collection is determined to be accurate and there is an abnormality in the parameters of the optical fiber connection point, the sending data to be transmitted is automatically adjusted so that the final received data is consistent with the sending data before adjustment, reducing the situation where there are errors in data transmission, avoiding losses to users, improving the accuracy of detection, and improving the efficiency of detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0175] Figure 1 This is a flow chart of the optical cable fault detection method based on industrial data processing of the present invention. DETAILED DESCRIPTION

[0176] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0177] Example 1: A method for detecting optical cable faults based on industrial data processing, see Figure 1 ,include:

[0178] Collect data on the fiber splicing points of the optical fibers in the optical cable;

[0179] For each fiber splice, select a smooth OTDR curve on the fiber segments before and after the splice.

[0180] Calculate the loss of optical fiber splicing points:

[0181] Among them, P before is the optical power before the connection point, P after is the optical power after the connection point;

[0182] For each fiber splice, measure the optical power of the reflection peak at the splice location and record it as P refl ;

[0183] Calculate the reflectivity of the optical fiber splice point:

[0184] Among them, P incident is the incident light power;

[0185] Set the loss threshold and reflectivity threshold, respectively denoted as Loss threshold and Reflectivity threshold , the loss threshold is a threshold set according to actual needs, such as the loss threshold is 0.1dB, and the reflectivity threshold is a threshold set according to actual needs, such as the reflectivity threshold is -50dB;

[0186] Set a parameter judgment function to determine whether the parameters of the optical fiber splicing point are normal:

[0187]

[0188] If F j (Loss co_p , Reflectivity co_p )=True, it is determined that the parameter detection of the optical fiber connection point is normal;

[0189] If F j (Loss co_p , Reflectivity co_p )=False, then it is determined that the parameter detection of the optical fiber connection point is abnormal;

[0190] Verify the accuracy of the parameters of the optical fiber in the cable through the detection model;

[0191] If the parameter detection of the optical fiber connection point is normal and the data collection is accurate, data transmission will proceed normally;

[0192] If the parameter detection of the optical fiber connection point is abnormal and the data collection is accurate, the transmitted data is adjusted according to the signal adjustment model so that the received data is consistent with the sent data before the adjustment;

[0193] If the data collection is wrong, it will prompt that the parameters of the optical fiber connection point are abnormal and data transmission will not be performed.

[0194] The acquisition of optical fiber connection point data of optical fibers in optical cables is specifically as follows:

[0195] Set OTDR parameters, including wavelength λ, pulse width τ, and average times N avg And the distance range L range_dis ;

[0196] Collect the backscattered signal curve. The backscattered signal curve can be represented by a two-dimensional array, where each element contains the distance and the corresponding optical power value:

[0197] BS(x)=P(x)(x=1,2,...,N);

[0198] Where x is the distance from the OTDR starting point, and P(x) is the optical power at distance x.

[0199] Identify the reflection peak position, that is, find the local maximum point on the curve:

[0200] PO peak_refl ={x i |P(x i )>P(x i -Δx)and P(x i )>P(x i +Δx)};

[0201] Among them, Δx is used to determine the window size of the local extreme value;

[0202] Identify the loss step location, that is, find the local minimum point on the curve:

[0203] PO step_loss ={x i |P(x i )<P(x i -Δx)and P(x i )<P(x i +Δx)};

[0204] Among them, Δx is used to determine the window size of the local extreme value;

[0205] Record the location of the connection point, that is, the distance from the OTDR starting point:

[0206] PO point_con ={x joint，1 , x joint，2 ,...,x joint，m}.

[0207] This embodiment further provides that, for each optical fiber splicing point, a smooth OTDR curve is selected on the optical fiber segments before and after the splicing point, specifically:

[0208] For each fiber splice, the location of the splice is denoted by x. joint ;

[0209] Set a length, record it as L ref ;

[0210] Take a distance range before and after the connection point as the reference interval, and the length of the distance range is L ref ;

[0211] The reference interval before the connection point is

[0212] The reference interval after the connection point is

[0213] Extracting optical power data of a reference interval from the backscattered signal curve BS(x)=P(x);

[0214] Among them, the reference curve data before the connection point is

[0215] The reference curve data after the connection point is

[0216] Smooth the extracted reference curve data to remove noise and small fluctuations:

[0217]

[0218] Where Δx is the distance between data points on the curve, and k is the half-window size of the moving average;

[0219] Compute the first derivative of the smoothed curve:

[0220]

[0221] Determine the smoothness of the curve:

[0222]

[0223] Among them, ε smooth The smoothness threshold for verifying the smoothness of the curve is set according to actual needs, for example, the smoothness threshold is 0.1;

[0224] If F smooth (P' smooth (x)) = 1, the curve is judged to be smooth;

[0225] If F smooth (P' smooth If (x))=0, the curve is judged to be not smooth.

[0226] Through the above method, based on the OTDR parameters of the optical fiber in the optical cable, it is judged whether the parameters of the optical fiber connection point are abnormal, and a simulation model is generated according to different interference conditions and different optical fiber conditions. Based on the simulation model, it is judged whether the collected parameters of the optical fiber connection point are accurate. When there is interference in the collected parameters, the real data is automatically restored. When there is an abnormality in the parameters of the optical fiber connection point, the data to be transmitted is automatically adjusted to reduce errors in data transmission, avoid losses to users, improve the accuracy of detection, and improve the efficiency of detection.

[0227] Embodiment 2: This embodiment is an improvement made on the basis of embodiment 1. The optical cable fault detection method based on industrial data processing, the detection model is specifically:

[0228] Get the length of the optical fiber, recorded as L current ;

[0229] Set the collection interval, denoted as d;

[0230] According to the collection interval d, several data collection points are set on the optical fiber;

[0231] Calculate the position of each data collection point:

[0232] x i =i·d(i=1,2,...,n);

[0233] Among them, x i represents the position of the i-th data collection point, n is the number of data collection points, and satisfies:

[0234]

[0235] in, Indicates a round-down operation;

[0236] like Then the position of the last data collection point is:

[0237] x n =L current ;

[0238] That is, the last acquisition point is set at the end of the optical fiber;

[0239] At this time, the corrected length of the last interval of the actual collection point interval is:

[0240] L current -x n-1 ;

[0241] Perform interference analysis.

[0242] The interference analysis is specifically performed as follows:

[0243] S1. For the i-th data collection point on the optical fiber, define its interference condition vector as:

[0244] C i =[C i1 , C i2 ,...,C ik ];

[0245] Among them, k is the number of different test conditions, such as light, temperature, etc., C ij It represents the interference situation of the i-th data collection point under the j-th test condition, C ij =1 means there is interference, C ij =0 means no interference;

[0246] S2. For all n data collection points, define the interference condition vector set as:

[0247] C=[C1,C2,...,C n ];

[0248] S3. Calculate the intersection of the interference condition vectors of all data collection points:

[0249]

[0250] S4. Set a consistency judgment function to determine whether the interference conditions of all data collection points are consistent:

[0251]

[0252] Where j = 1, 2, ..., k, I = C j Indicates that all data collection points point to the jth interference condition, and I = / 0 indicates no interference condition;

[0253] S5, if F consistency (I) = 1, the data collection is judged to be accurate;

[0254] If F consistency (I) = 0, then an initial range is set with the fiber connection point as the origin, denoted as R0. The initial range is the range formed by taking the value set according to actual needs as the radius and the fiber connection point as the origin;

[0255] S6. Obtain the interference condition vector set of all data collection points within the range:

[0256]

[0257] Where m is the number of data collection points within the range;

[0258] S7. Calculate the intersection of the interference condition vectors of the data collection points within the range:

[0259]

[0260] S8. Use the range determination function to determine whether the interference conditions of the data collection points within the range are consistent:

[0261]

[0262] Among them, I range =C j Indicates that all data collection points within the range point to the jth interference condition, Indicates that there are no interference conditions within the range;

[0263] S9, if F range (I range )=1, then the data collection is determined to be accurate;

[0264] If F range (I range )=0, then obtain the range reduction model;

[0265] S10, substituting the initial range R0 into the range reduction model to generate a new range, and marking the new range as R0;

[0266] S11, repeat S6-S12, set the range threshold, record it as R threshold ;

[0267] Substitute the re-evaluated value into the loop evaluation function:

[0268]

[0269] Among them, STOP is the sign to stop the cycle, and CONTINUE is the sign to continue the cycle;

[0270] S12, if F cycle (I range )=STOP2, then stop the loop and determine that the data collection error;

[0271] If F cycle (I range )=STOP1, then stop the loop and determine that the data collection is accurate;

[0272] If F cycle (I range )=CONTINUE, then the loop continues and S6-S12 are repeatedly executed.

[0273] The scope reduction model is specifically:

[0274] Get the current range, recorded as R t ;

[0275] Taking the fiber splicing point as the origin, calculate the new range:

[0276]

[0277] Among them, R t+1 For the narrowed scope;

[0278] Get the current interval distance, recorded as d t ;

[0279] Calculate the new separation distance:

[0280]

[0281] Among them, d t+1 The shortened separation distance;

[0282] Then the number of data collection points in the new range is:

[0283]

[0284] The location of each new data collection point is:

[0285]

[0286] Among them, x center The center position of the current range, that is, the location of the optical fiber connection point.

[0287] This embodiment also provides that the detection model further includes troubleshooting interference causes and determining and updating final fiber connection point data, specifically:

[0288] like or This means that the collected data is free of interference and can be used directly;

[0289] If I range =C j or I=C j , it means that the collected data is interfered with, collect the deviation data set D deviation ={(C i , ΔP i )|i=1, 2, ..., m};

[0290] Among them, C i Represents the interference condition vector of the i-th sample, specifically:

[0291] C i =[C i1 , C i2 ,...,Cik ];

[0292] ΔP i is the parameter deviation vector of the optical fiber splicing point of the i-th sample, which represents the difference between the parameters under interference conditions and the actual parameters. Specifically:

[0293]

[0294] Train a perturbation model:

[0295] f(C)=β0+β1C i1 +β2C i2 +...+β k C ik ;

[0296] Among them, β0, β1, ..., β k are model parameters, estimated by minimizing a loss function such as mean square error:

[0297]

[0298] Get the current fiber connection point parameters, recorded as P measured ;

[0299] Extract the predicted deviation, denoted as ΔP predicted =f(C current );

[0300] Among them, C current is the current interference condition vector;

[0301] Calculate the actual parameters of the fiber splice point:

[0302] P true =P measured -ΔP predicted ;

[0303] Recalculate the loss and reflectivity of optical fiber splices;

[0304] Use parameter determination function to judge whether the parameters of the optical fiber splicing point are normal;

[0305] If F j (Loss co_p , Reflectivity co_p )=True, it is determined that the parameter detection of the optical fiber connection point is normal;

[0306] If F j (Loss co_p , Reflectivity co_p )=False, it is determined that the parameter detection of the optical fiber connection point is abnormal.

[0307] Example 3: This example is an improvement made on the basis of Example 2. In this example, the signal adjustment model is specifically as follows:

[0308] Get all the input data to form an input data set, denoted as D in ={D in，1 , D in，2 ,...,D in，m};

[0309] Among them, each D in，i Represents a specific set of input data;

[0310] Get different fiber line lengths to form a line data set, denoted as L line ={L line，1 , L line，2 ,...,L line，l};

[0311] Among them, each L line，h Indicates a specific line length;

[0312] Obtain different numbers of fiber splicing points to form a splicing point data set, denoted as N = {N1, N2, ..., N k};

[0313] Among them, each N j Indicates the number of optical fiber splicing points;

[0314] For each fiber connection point number, obtain different fiber connection point parameters for each fiber connection point, denoted as P n ={P n，1 , P n，2 ,...,P n，p};

[0315] Among them, each P n，q Indicates the qth parameter;

[0316] For each fiber connection point parameter, obtain its different parameter values, recorded as V n，p ={V n，p，1 , V n，p，2 ,...,V n，p，s};

[0317] Among them, each V n，p，r Indicates a possible value for the parameter;

[0318] For each combination (D in，i , N j , P n，q =V n，p，r , L line，h), use the optical fiber transmission model to simulate and get the actual output data, denoted as D out , and generate a training data set, denoted as D:

[0319]

[0320] Define an adjustment amount, denoted as ΔD in ;

[0321] For each input data D in，i , adjustment amount ΔD in，i satisfy:

[0322] D out (D in，i +ΔD in，i , N j , P n，q =V n，p，r , L line，h )=D in，i ;

[0323] Use the gradient descent method to optimize and calculate the updated adjustment amount ΔD in，i :

[0324]

[0325] Where t is the number of iterations, is the adjustment amount at the tth iteration, η is the learning rate, which is used to control the step size of each update, Calculate the difference function J versus ΔD in，i The gradient at iteration t, Represents the gradient operator, calculating the difference function J for the adjustment amount ΔD in，i Gradient:

[0326]

[0327] Where d is the adjustment amount ΔD in，i The dimension of the difference function J is the actual output data D out And the expected output data, that is, the input data D in，i The function of the difference between them is:

[0328] J(ΔD in，i )=||D out (D in，i +ΔD in，i , N j , P n，q =V n，p，r , L line，h )-D in，i || 2 ;

[0329] Set the maximum number of iterations, denoted as T. The maximum number of iterations is a value set according to specific requirements, such as 3;

[0330] Set a given threshold, denoted as ε, which is a value set according to specific needs. For example, the given threshold is 10 -6 ;

[0331] Set an optimization judgment function to judge the updated adjustment amount ΔD in，i Is it suitable?

[0332]

[0333] If F optimize (ΔD in，i )=1, then the iterative optimization is stopped and the adjustment amount ΔD obtained after iterative optimization is in，i Integrate into the training dataset D;

[0334] If F optimize (ΔD in，i )=0, then continue iterative optimization.

[0335] This embodiment also provides that the transmitted data is adjusted according to the signal adjustment model so that the received data is consistent with the sent data before the adjustment, specifically:

[0336] Get the fiber optic line length of the current line, recorded as L line_current ;

[0337] Get the input data that needs to be transmitted, denoted as D in_current ;

[0338] Get the number of fiber connection points on the current line, recorded as N current ;

[0339] Obtain the optical fiber connection point parameters of each optical fiber connection point and the parameter value of each optical fiber connection point parameter, which is recorded as {(P n，c , V n，c，c_c )|P n，c =V n，c，c_c};

[0340] Integration generates the transmission data set, denoted as (D in_current , N current , P n，c =V n，c，c_c , L line_current );

[0341] Extract the corresponding adjustment amount of the transmission data set in the signal adjustment model, recorded as ΔD in_current ;

[0342] Calculate the final input data D in_final :

[0343] D in_final =D in_current +ΔD in_current ;

[0344] Input data D in_current Adjust to the final input data D in_final Then data transmission is performed.

[0345] In this embodiment, based on the OTDR parameters of the optical fiber in the optical cable, it is determined whether the parameters of the optical fiber connection point are abnormal, and a simulation model is generated according to different interference conditions and different optical fiber conditions. Based on the simulation model, it is determined whether the collected parameters of the optical fiber connection point are accurate. When there is interference in the collected parameters, the data are automatically restored to the real data. When there is an abnormality in the parameters of the optical fiber connection point, the data to be transmitted is automatically adjusted to reduce errors in data transmission, avoid losses to users, improve the accuracy of detection, and improve the efficiency of detection.

Claims

1. A method for detecting optical cable faults based on industrial data processing, characterized in that: include: Collect data on the fiber splicing points of the optical fibers in the optical cable; For each fiber splice, select a smooth OTDR curve on the fiber segments before and after the splice. Calculate the loss of optical fiber splicing points: ; in, is the optical power before the connection point, is the optical power after the connection point; For each optical fiber splice, measure the optical power of the reflection peak at the splice location and record it as ; Calculate the reflectivity of the optical fiber splice point: ; in, is the incident light power; Set the loss threshold and reflectivity threshold, respectively, as and ; Set a parameter judgment function to determine whether the parameters of the optical fiber splicing point are normal: ； like , it is determined that the parameter detection of the optical fiber connection point is normal; like , then it is determined that the parameter detection of the optical fiber connection point is abnormal; Verify the accuracy of the parameters of the optical fiber in the cable through the detection model; If the parameter detection of the optical fiber connection point is normal and the data collection is accurate, data transmission will proceed normally; If the parameter detection of the optical fiber connection point is abnormal and the data collection is accurate, the transmitted data is adjusted according to the signal adjustment model so that the received data is consistent with the sent data before the adjustment; If the data collection is wrong, it will prompt that the parameters of the optical fiber connection point are abnormal and data transmission will not be carried out; The detection model is specifically: Get the length of the optical fiber, recorded as ; Set the collection interval, recorded as ; According to the collection interval , set up several data collection points on the optical fiber; Calculate the position of each data collection point: ； in, Indicates the The location of the data collection points, is the number of data collection points, satisfying: ； in, Indicates a round-down operation; like , then the position of the last data collection point is: ； At this time, the corrected length of the last interval of the actual collection point interval is: ； Conduct interference analysis; The interference analysis is specifically as follows: S1, for the first data collection points, and define the interference condition vector as: ； in, is the number of different test conditions, Indicates the The data collection point is Interference conditions under various test conditions, Indicates that there is interference. Indicates no interference; S2. For all data collection points, and define the interference condition vector set as: ； S3. Calculate the intersection of the interference condition vectors of all data collection points: ； S4. Set a consistency judgment function to determine whether the interference conditions of all data collection points are consistent: ； in, , Indicates that all data collection points point to the interference conditions, represents the no-interference condition; S5, if , then the data collection is determined to be accurate; like , then take the fiber connection point as the origin and set an initial range, which is recorded as ; S6. Obtain the interference condition vector set of all data collection points within the range: ； in, is the number of data collection points within the range; S7. Calculate the intersection of the interference condition vectors of the data collection points within the range: ； S8. Use the range determination function to determine whether the interference conditions of the data collection points within the range are consistent: ； in, Indicates that all data collection points within the range point to the interference conditions, Indicates that there are no interference conditions within the range; S9, if , then the data collection is determined to be accurate; like , then obtain the scope reduction model; S10, the initial range Substitute the scope reduction model to generate a new scope and mark the new scope as ; S11, repeat S6-S12, set the range threshold, record it as ; Substitute the re-evaluated value into the loop evaluation function: ； in, A sign to stop the loop. A sign to continue the cycle; S12, if , then stop the loop and determine that the data collection is wrong; like , then stop the loop and judge that the data collection is accurate; like , then continue the loop and repeat S6-S12.

2. The optical cable fault detection method based on industrial data processing according to claim 1, characterized in that: The collecting of optical fiber connection point data of optical fibers in optical cables is specifically as follows: Set OTDR parameters, including wavelength , pulse width , average number of times and distance range ; Collect backscatter signal curve: ； in, is the distance from the starting point of the OTDR, For the distance The optical power at Identify reflection peak positions: ； in, The window size used to determine local extreme values; Identify the loss step location: ； in, The window size used to determine local extreme values; Record the connection point location: 。 3. The optical cable fault detection method based on industrial data processing according to claim 1, characterized in that: For each optical fiber splicing point, a smooth OTDR curve is selected on the optical fiber segments before and after the splicing point, specifically: For each fiber splice, the splice location is recorded as ; Set a length, recorded as ; Take a distance range before and after the connection point as the reference interval, and the length of the distance range is ; The reference interval before the connection point is ; The reference interval after the connection point is ; From the backscattered signal curve Extract the optical power data of the reference interval; Among them, the reference curve data before the connection point is ; The reference curve data after the connection point is ; Smooth the extracted reference curve data: ； in, is the distance between data points on the curve, is the half window size of the moving average; Compute the first derivative of the smoothed curve: ； Determine the smoothness of the curve: ； in, is the smoothness threshold for verifying the smoothness of the curve; like , then the curve is judged to be smooth; like , then the curve is judged to be not smooth.

4. The optical cable fault detection method based on industrial data processing according to claim 1, characterized in that: The scope reduction model is specifically: Get the current scope, recorded as ; Taking the fiber splicing point as the origin, calculate the new range: ； in, For the narrowed scope; Get the current interval distance, recorded as ; Calculate the new separation distance: ； in, The shortened separation distance; Then the number of data collection points in the new range is: ； The location of each new data collection point is: ； in, The center position of the current range.

5. The optical cable fault detection method based on industrial data processing according to claim 4, characterized in that: The detection model also includes troubleshooting interference causes and determining and updating final fiber splicing point data, specifically: like or , it means that the collected data has no interference and can be used directly; like or , it means that the collected data is interfered with, collect the deviation data set ; in, Indicates the The interference condition vector of samples is: ； For the The fiber splicing point parameter deviation vector of the sample represents the difference between the parameters under interference conditions and the actual parameters, specifically: ； Train a perturbation model: ； in, are model parameters, estimated by minimizing the loss function: ； Get the current fiber connection point parameters, recorded as ; Extract the predicted deviation, denoted as ; in, is the current interference condition vector; Calculate the actual parameters of the fiber splice point: ； Recalculate the loss and reflectivity of optical fiber splices; Use parameter determination function to judge whether the parameters of the optical fiber splicing point are normal; like , it is determined that the parameter detection of the optical fiber connection point is normal; like , it is determined that the parameter detection of the optical fiber connection point is abnormal.

6. The optical cable fault detection method based on industrial data processing according to claim 1, characterized in that: The signal adjustment model is specifically: Get all the input data to form an input data set, denoted as ; Among them, each Represents a specific set of input data; Get different fiber line lengths to form a line data set, recorded as ; Among them, each Indicates a specific line length; Obtain different numbers of fiber splicing points to form a splicing point dataset, recorded as ; Among them, each Indicates the number of optical fiber splicing points; For each fiber connection point number, obtain different fiber connection point parameters for each fiber connection point, which are recorded as ; Among them, each Indicates the parameters; For each fiber connection point parameter, obtain its different parameter values, which are recorded as ; Among them, each Indicates a possible value for the parameter; For each combination , using the optical fiber transmission model to simulate, the actual output data is obtained, which is recorded as , and generate a training data set, denoted as : ； Define an adjustment amount, denoted as ; For each input data , adjustment amount satisfy: ； Use gradient descent to optimize and calculate the updated adjustment amount : ； in, is the number of iterations, For the first The adjustment amount at the iteration, is the learning rate, Calculate the difference function right In the The gradient at iteration , Represents the gradient operator, calculating the difference function Adjustment amount Gradient: ； in, Adjustment amount Dimension, difference function For actual output data and the expected output data, i.e., the input data The function of the difference between them is: ； Set the maximum number of iterations, denoted as ; Set a given threshold, denoted as ; Set an optimization judgment function to determine the adjustment amount after the update Is it suitable? ； like , then stop the iterative optimization and set the adjustment amount obtained after iterative optimization Integrate into the training dataset ; like , then continue iterative optimization.

7. The optical cable fault detection method based on industrial data processing according to claim 1, characterized in that: The signal adjustment model is used to adjust the transmitted data so that the received data is consistent with the transmitted data before the adjustment, specifically: Get the fiber line length of the current line, recorded as ; Get the input data that needs to be transmitted, recorded as ; Get the number of fiber connection points of the current line, recorded as ; Get the fiber connection point parameters of each fiber connection point and the parameter value of each fiber connection point parameter, recorded as ; Integration generates the transmission data set, denoted as ; Extract the corresponding adjustment amount of the transmission data set in the signal adjustment model, denoted as ; Calculate the final input data : ； Input data Adjust to the final input data Then data transmission is performed.

Images