Crimping pipe detection method based on magnetic flux leakage signal

Abstract

The invention discloses a crimping pipe detection method based on a magnetic flux leakage signal, and the method comprises the steps: firstly collecting the magnetic flux leakage signal of a crimping pipe to construct an initial sample data set, and carrying out the normalization processing of the constructed initial sample data set; taking the original threshold value and the original weight of the BP neural network as an initial population position of an improved information acquisition optimization algorithm for optimization, and acquiring an optimal threshold value and an optimal weight of the BP neural network; and training the BP neural network to obtain a crimping tube detection model, and carrying out detection by using the crimping tube detection model according to the magnetic flux leakage signal collected in real time. According to the invention, the BP neural network of the improved information acquisition optimization algorithm is adopted to detect the crimping pipe, and the state of the crimping pipe can be accurately detected in real time.

Description

Technical Field

[0001] The invention belongs to the technical field of crimping tube detection, and in particular relates to a crimping tube detection method based on magnetic flux leakage signals. Background Art

[0002] The crimping quality inspection of the crimping tube is an important part of ensuring the safe and stable operation of the power system. Most of the failures of the crimping tube are concentrated in the crimping groove of the steel anchor. Therefore, the crimping quality inspection of the crimping groove is the top priority of the entire crimping quality inspection of the crimping tube. At present, the appearance size inspection method is commonly used in the actual construction process. This method is widely used in the actual construction site. According to the experience of on-site construction and acceptance work, the length from the uncrimped crimping tube handle to the steel anchor crimping groove is measured using a steel tape measure, and then compared with the length of the crimping groove after crimping. According to the "Technical Specifications for Hydraulic Crimping of Overhead Conductors and Ground Wires in Transmission and Transformation Engineering", the crimping condition of the steel anchor groove of the crimping tube is analyzed. This method is simple to operate, and the equipment only needs a roll of steel tape measure, but the measurement process requires manual measurement by the operator, so it is easy to make misjudgments during the on-site acceptance process, resulting in unqualified tension-resistant crimping tubes being put into operation in the transmission line. In recent years, the X-ray detection method has also been used in this field, but the high cost of its equipment and the hidden dangers to human health are the shortcomings of this method. Therefore, there is an urgent need for a convenient, economical, highly accurate method that does not pose a health hazard to the human body to replace existing detection methods. Summary of the invention

[0003] In response to the above-mentioned problems, many researchers have proposed a variety of solutions, but the results are not significant. In order to actually solve the current technical difficulties, the present invention discloses a crimping tube detection method based on leakage magnetic signals. The method first collects the leakage magnetic signals of the crimping tube to construct an initial sample data set, and then normalizes the initial sample data set and uses a BP neural network optimized by an improved information acquisition optimization algorithm to realize the detection of the crimping tube.

[0004] A method for detecting a crimping tube based on a magnetic flux leakage signal comprises the following steps:

[0005] Step S1: Collect the leakage magnetic signal of the crimping tube to construct an initial sample data set, and mark the leakage magnetic signal according to the state type;

[0006] Step S2: construct a BP neural network model, import the feature data set in step S1 into the BP neural network model for training, and during the training process, optimize the number of nodes in the hidden layer of the BP neural network model by improving the information acquisition optimization algorithm to obtain the optimal number of nodes;

[0007] Step S3: collecting magnetic flux leakage signals in real time, importing the trained BP neural network model, and obtaining the risk type of the crimping tube in real time;

[0008] In step S2, the process of optimizing the number of nodes of the BP neural network model by improving the information acquisition optimization algorithm is as follows:

[0009] Step S21: Initialize the population;

[0010] Step S22: setting the initial position of the population;

[0011] Step S23: Information collection;

[0012] Step S24: filtering and evaluating information;

[0013] Step S25: analysis and organization of information;

[0014] Step S26: Determine whether the maximum number of iterations has been reached. If so, calculate the fitness of the individual at the current number of iterations and select the individual with the maximum fitness as the optimal number of nodes. If not, continue with steps S22 to S26.

[0015] Furthermore, step S21 is specifically as follows:

[0016] Set the initial population size and maximum number of iterations for the improved information acquisition optimization algorithm;

[0017] Step S22 specifically includes: taking the original threshold and original weight of the BP neural network as the initial position of the population; expressed as:

[0018]

[0019] a∈(0.9,1.08);

[0020] i∈1,2,…,n;

[0021] j∈1,2,…,d;

[0022] Among them, a is the control parameter, X represents the initialization population after Singer chaos mapping, represents the initial value of the i-th information individual in the j-th dimension, x i,j It represents the value of the i-th information individual in the j-th dimension after Singer chaotic mapping, n is the population size, d is the dimension of the problem; each information individual represents a set of parameter solutions of the original threshold and original weight of the BP neural network, ub and lb are the upper and lower bounds of the problem respectively, Rand represents a random number between 0 and 1, and r is a random number.

[0023] Furthermore, step S23 is specifically as follows: individuals use various methods to collect information from different sources, which is regarded as a differential evolution, expressed as:

[0024]

[0025] Where: t represents the current iteration number, represents the position of the i-th information at the t-th iteration, represents the position after the i-th information iteration, θ is a random number between [0,1], and represents the positions of any two different individuals in the tth iteration;

[0026] Step S23A: In the information collection phase, a triangle walk strategy is introduced, which is expressed as:

[0027]

[0028] L2 = rand() × L1;

[0029] λ = 2 × π × rand();

[0030] P=L1 2 +L2 2 -2×L1×L2×cos(λ);

[0031]

[0032] Where: L1 represents the distance between the ith individual and the best individual at the tth iteration, L2 represents the walking step length of the ith individual, r and rand() represent random numbers between 0 and 1, λ represents a random number between 0 and 2π, and P represents a constant.

[0033] Furthermore, step S24 is specifically as follows: filtering and evaluating the information is expressed as:

[0034]

[0035] In the formula, rand is a random number generated in [0,1]. represents the position of any individual in the tth iteration, Δ is the error caused by subjective factors when filtering and evaluating information, expressed as:

[0036]

[0037] In the formula, Ξ is the subjective influencing factor, which is a quantitative indicator of individual subjectivity. It reflects that the individual's preferences, experience, emotions and preconceived ideas will make overly optimistic or pessimistic judgments on information. Γ represents the reliability factor. The value of Ξ is calculated by the following formula:

[0038] Ξ=2×mod(3.468×υ×(1-β)×(acos(γ×10 4 ))),1)

[0039] Where υ, β, and γ are random numbers between [0,1], and mod(·) represents the remainder function;

[0040] Γ represents the ability of the algorithm to adjust itself to optimize its behavior according to the quality of information at different iteration stages; the calculation formula is as follows:

[0041]

[0042] Among them, T is the maximum number of iterations, and the mathematical model of Γ consists of three main parts: the sine function part, the logarithmic function part, and the information quality factor Φ;

[0043] The formula for information quality factor Φ is:

[0044]

[0045] Where δ is a random number generated between [0,1].

[0046] Further, step S24 is specifically: identifying existing useful information from the filtered information, and converting the convertible information identified in the previous stage into useful information, which is expressed as:

[0047]

[0048] in: represents the best information body generated in the previous iteration, represents the average value of the best information body generated in the previous iteration, ε, ζ, κ, ω represent random numbers generated between [0,1], and Λ represents the control factor for analyzing and organizing information, which is defined as:

[0049]

[0050] Furthermore, the number of input layer nodes of the BP neural network model is equal to the dimension of the input vector, and the number of output layer nodes is consistent with the number of prediction results; the number of hidden layer nodes is determined by the following formula:

[0051]

[0052] Among them, N h Represents the number of hidden layer nodes, N p Represents the number of input layer nodes, N o represents the number of output layer nodes, and α is a constant between [1,10].

[0053] Furthermore, the risk types in step S1 include: no risk, low risk, medium risk and high risk.

[0054] An electronic device includes a processor, a memory and a bus, wherein the processor and the memory are connected via the bus, wherein the memory is used to store a set of program codes, and the processor is used to call the program codes stored in the memory to execute the above-mentioned crimping tube detection method based on leakage magnetic signals.

[0055] A non-volatile computer storage medium stores computer executable instructions, and the computer executable instructions can execute the above-mentioned crimping tube detection method based on leakage magnetic signal.

[0056] The beneficial effects of the present invention are as follows: the BP neural network model optimized by the improved information acquisition optimization algorithm is used to detect the crimped tube, which solves the problem that the hidden layer nodes of the BP neural network model are difficult to accurately select. Through the leakage magnetic signal, the crimping condition of the crimped tube can be accurately detected in real time. After the threshold and weight are optimized by the improved information acquisition optimization algorithm, the crimped tube detection of the BP neural network will be more accurate. Moreover, the scheme can realize the detection of the crimped tube without destroying the crimped tube, which is of great significance to the safe dispatch and stable operation of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 The figure is a flow chart of a crimping tube detection method based on magnetic flux leakage signal.

[0058] Figure 2 Flowchart of optimization algorithm for information acquisition.

[0059] Figure 3 Comparison of the convergence curves before and after algorithm improvement. DETAILED DESCRIPTION

[0060] The following describes the embodiments of the present invention through specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention.

[0061] A method for detecting a crimping tube based on a magnetic leakage signal, in one example, comprises the following steps:

[0062] Step S1: Collect the leakage magnetic signal of the crimping tube to construct an initial sample data set, and mark the leakage magnetic signal according to the state type;

[0063] Step S2: construct a BP neural network model, import the feature data set in step S1 into the BP neural network model for training, and during the training process, optimize the number of nodes in the hidden layer of the BP neural network model by improving the information acquisition optimization algorithm to obtain the optimal number of nodes;

[0064] Step S3: Using the real-time collected leakage magnetic signal to detect the crimping tube using the crimping tube detection model, and obtaining the real-time state type of the crimping tube;

[0065] In step S2, the process of optimizing the number of nodes of the BP neural network model by improving the information acquisition optimization algorithm is as follows:

[0066] Step S21: Initialize the population;

[0067] Step S22: setting the initial position of the population;

[0068] Step S23: Information collection;

[0069] Step S24: filtering and evaluating information;

[0070] Step S25: analysis and organization of information;

[0071] Step S26: Determine whether the maximum number of iterations has been reached. If so, calculate the fitness of the individual at the current number of iterations and select the individual with the maximum fitness as the optimal number of nodes. If not, continue with steps S22 to S26.

[0072] Furthermore, the initial sample data set constructed in step S1 is expressed as:

[0073] S=[S1、S2、......S n ];

[0074] Among them, S represents the constructed initial sample data set, S n Data representing the magnetic flux leakage signal at time n.

[0075] Furthermore, the constructed initial sample data set is normalized; in order to make the initial sample data set meet the requirements of the BP neural network model, the normalization is performed as follows:

[0076]

[0077] Among them, G represents the value of the data in the initial sample data set after normalization in the interval [0,1], S max is the maximum value in the initial sample data set, S min is the minimum value in the initial sample data set, and S represents the value before normalization in the initial sample data set.

[0078] Furthermore, in one example, step S21 is specifically as follows:

[0079] Set the initial population size and maximum number of iterations for the improved information acquisition optimization algorithm;

[0080] Step S22 specifically includes: taking the original threshold and original weight of the BP neural network as the initial position of the population; expressed as:

[0081]

[0082] a∈(0.9,1.08);

[0083] i∈1,2,…,n;

[0084] j∈1,2,…,d;

[0085] Among them, a is the control parameter, X represents the initialization population after Singer chaos mapping, represents the initial value of the i-th information individual in the j-th dimension, x i,j It represents the value of the i-th information individual in the j-th dimension after Singer chaotic mapping, n is the population size, d is the dimension of the problem; each information individual represents a set of parameter solutions of the original threshold and original weight of the BP neural network, ub and lb are the upper and lower bounds of the problem respectively, Rand represents a random number between 0 and 1, and r is a random number.

[0086] Furthermore, step S23 is specifically as follows: individuals use various methods to collect information from different sources, which is regarded as a differential evolution, expressed as:

[0087]

[0088] Where: t represents the current iteration number, represents the position of the i-th information at the t-th iteration, represents the position after the i-th information iteration, θ is a random number between [0,1], and represents the positions of any two different individuals in the tth iteration;

[0089] Step S23A: In the information collection phase, a triangle walk strategy is introduced, which is expressed as:

[0090]

[0091] L2 = rand() × L1;

[0092] λ = 2 × π × rand();

[0093] P=L1 2 +L22 -2×L1×L2×cos(λ);

[0094]

[0095] Where: L1 represents the distance between the ith individual and the best individual at the tth iteration, L2 represents the walking step length of the ith individual, r and rand() represent random numbers between 0 and 1, λ represents a random number between 0 and 2π, and P represents a constant.

[0096] Furthermore, step S24 is specifically as follows: filtering and evaluating the information is expressed as:

[0097]

[0098] In the formula, rand is a random number generated in [0,1]. represents the position of any individual in the tth iteration, Δ is the error caused by subjective factors when filtering and evaluating information, expressed as:

[0099]

[0100] In the formula, Ξ is the subjective influencing factor, which is a quantitative indicator of individual subjectivity. It reflects that the individual's preferences, experience, emotions and preconceived ideas will make overly optimistic or pessimistic judgments on information. Γ represents the reliability factor. The value of Ξ is calculated by the following formula:

[0101] Ξ=2×mod(3.468×υ×(1-β)×(acos(γ×10 4 ))),1)

[0102] Where υ, β, and γ are random numbers between [0,1], and mod(·) represents the remainder function;

[0103] Γ represents the ability of the algorithm to adjust itself to optimize its behavior according to the quality of information at different iteration stages; the calculation formula is as follows:

[0104]

[0105] Among them, T is the maximum number of iterations, and the mathematical model of Γ consists of three main parts: the sine function part, the logarithmic function part, and the information quality factor Φ;

[0106] The formula for information quality factor Φ is:

[0107]

[0108] Where δ is a random number generated between [0,1].

[0109] Further, step S24 is specifically: identifying existing useful information from the filtered information, and converting the convertible information identified in the previous stage into useful information, which is expressed as:

[0110]

[0111] in: represents the best information body generated in the previous iteration, represents the average value of the best information body generated in the previous iteration, ε, ζ, κ, ω represent random numbers generated between [0,1], and Λ represents the control factor for analyzing and organizing information, which is defined as:

[0112]

[0113] Furthermore, the number of input layer nodes of the BP neural network model is equal to the dimension of the input vector, and the number of output layer nodes is consistent with the number of prediction results; the number of hidden layer nodes is determined by the following formula:

[0114]

[0115] Among them, N h Represents the number of hidden layer nodes, N p Represents the number of input layer nodes, N o represents the number of output layer nodes, and α is a constant between [1,10].

[0116] Furthermore, the risk types in step S1 include: no risk, low risk, medium risk and high risk.

[0117] It is worth noting that in the process of constructing the BP neural network model, the number of input layer nodes is equal to the dimension of the input vector. In the present invention, the dimension of the input vector is the dimension of the selected initial sample data set. Therefore, the number of input layer nodes of the BP neural network is 10; and the number of output layer nodes is consistent with the number of prediction results.

[0118] Furthermore, the original threshold and weight of the BP neural network are used as the initial population position of the improved information acquisition optimization algorithm to obtain the optimal threshold and optimal weight of the BP neural network model; the improved information acquisition optimization algorithm accelerates the algorithm convergence speed and optimization accuracy while increasing its ability to jump out of the local optimal solution. The convergence curve of the improved information acquisition optimization algorithm is shown in Figure 3 As shown in Figure 2, the convergence speed and accuracy have been improved. The optimization process of improving the information acquisition optimization algorithm is as follows: Figure 2 shown.

[0119] An electronic device includes a processor, a memory and a bus, wherein the processor and the memory are connected via the bus, wherein the memory is used to store a set of program codes, and the processor is used to call the program codes stored in the memory to execute the above-mentioned crimping tube detection method based on leakage magnetic signals.

[0120] A non-volatile computer storage medium stores computer executable instructions, and the computer executable instructions can execute the above-mentioned crimping tube detection method based on leakage magnetic signal.

[0121] The present invention is described in detail above in conjunction with the embodiments of the accompanying drawings. A person skilled in the art can make various variations of the present invention according to the above description. Therefore, some details in the embodiments should not be construed as limiting the present invention, and the present invention shall be protected by the scope defined by the attached claims.

Claims

1. A method for detecting a crimped tube based on a magnetic flux leakage signal, characterized in that: The following steps are involved: Step S1: Collect the leakage magnetic signals of the crimping tube to construct an initial sample data set, and mark the leakage magnetic signals according to the risk type; Step S2: construct a BP neural network model, import the feature data set in step S1 into the BP neural network model for training, and during the training process, optimize the number of nodes in the hidden layer of the BP neural network model by improving the information acquisition optimization algorithm to obtain the optimal number of nodes; Step S3: The magnetic leakage signal collected in real time is imported into the trained BP neural network model to obtain the risk type of the crimping tube in real time; In step S2, the process of optimizing the number of nodes of the BP neural network model by improving the information acquisition optimization algorithm is as follows: Step S21: Initialize the population; Step S22: setting the initial position of the population; Step S23: Information collection; Step S24: filtering and evaluating information; Step S25: analysis and organization of information; Step S26: Determine whether the maximum number of iterations has been reached. If so, calculate the fitness of the individual at the current number of iterations and select the individual with the maximum fitness as the optimal number of nodes. If not, continue with steps S22 to S26.

2. A method for detecting a crimped tube based on a magnetic flux leakage signal according to claim 1, characterized in that: Step S21 is specifically as follows: Set the initial population size and maximum number of iterations for the improved information acquisition optimization algorithm; Step S22 specifically includes: taking the original threshold and original weight of the BP neural network as the initial position of the population; expressed as: a∈(0.9,1.08)； i∈1,2,…,n; j∈1,2,…,d; Among them, a is the control parameter, X represents the initialization population after Singer chaos mapping, represents the initial value of the i-th information individual in the j-th dimension, x i,j It represents the value of the i-th information individual in the j-th dimension after Singer chaotic mapping, n is the population size, d is the dimension of the problem; each information individual represents a set of parameter solutions of the original threshold and original weight of the BP neural network, ub and lb are the upper and lower bounds of the problem respectively, Rand represents a random number between 0 and 1, and r is a random number.

3. A method for detecting a crimped tube based on a magnetic flux leakage signal according to claim 2, characterized in that: Step S23 is specifically as follows: individuals use various methods to collect information from different sources, which is regarded as a differential evolution, expressed as: Where: t represents the current iteration number, represents the position of the i-th information at the t-th iteration, represents the position after the i-th information iteration, θ is a random number between [0,1], and represents the positions of any two different individuals in the tth iteration; Step S23A: In the information collection phase, a triangle walk strategy is introduced, which is expressed as: L2 = rand() × L1; λ = 2 × π × rand(); P=L1 2 +L2 2 -2×L1×L2×cos(λ); Where: L1 represents the distance between the ith individual and the best individual at the tth iteration, L2 represents the walking step length of the ith individual, r and rand() represent random numbers between 0 and 1, λ represents a random number between 0 and 2π, and P represents a constant.

4. A method for detecting a crimped tube based on a magnetic flux leakage signal according to claim 3, characterized in that: Step S24 is specifically: filtering and evaluating the information is expressed as: In the formula, rand is a random number generated in [0,1]. represents the position of any individual in the tth iteration, Δ is the error caused by subjective factors when filtering and evaluating information, expressed as: In the formula, Ξ is the subjective influencing factor, which is a quantitative indicator of individual subjectivity. It reflects that the individual's preferences, experience, emotions and preconceived ideas will make overly optimistic or pessimistic judgments on information. Γ represents the reliability factor. The value of Ξ is calculated by the following formula: Ξ=2×mod(3.468×υ×(1-β)×(acos(γ×10 4 ))),1) Where υ, β, and γ are random numbers between [0,1], and mod(·) represents the remainder function; Γ represents the ability of the algorithm to adjust itself to optimize its behavior according to the quality of information at different iteration stages; the calculation formula is as follows: Among them, T is the maximum number of iterations, and the mathematical model of Γ consists of three main parts: the sine function part, the logarithmic function part, and the information quality factor Φ; The formula for information quality factor Φ is: Where δ is a random number generated between [0,1].

5. A method for detecting a crimped tube based on a magnetic flux leakage signal according to claim 4, characterized in that: Step S24 is specifically: identifying existing useful information from the filtered information, and converting the convertible information identified in the previous stage into useful information, which is expressed as: in: represents the best information body generated in the previous iteration, represents the average value of the best information body generated in the previous iteration, ε, ζ, κ, ω represent random numbers generated between [0,1], and Λ represents the control factor for analyzing and organizing information, which is defined as:

6. A method for detecting a crimped tube based on a magnetic flux leakage signal according to claim 5, characterized in that: The number of nodes in the input layer of the BP neural network model is equal to the dimension of the input vector, and the number of nodes in the output layer is consistent with the number of prediction results; the initial number of nodes in the hidden layer is determined by the following formula: Among them, N h Represents the initial number of nodes in the hidden layer, N p Represents the number of input layer nodes, N o represents the number of output layer nodes, and α is a constant between [1,10].

7. The method for detecting a crimped tube based on a magnetic flux leakage signal according to claim 5, characterized in that: The risk types in step S1 include: no risk, low risk, medium risk and high risk.

8. An electronic device, characterized in that: It includes a processor, a memory and a bus, wherein the processor and the memory are connected via the bus, wherein the memory is used to store a set of program codes, and the processor is used to call the program codes stored in the memory to execute a crimping tube detection method based on leakage magnetic signals as described in any one of claims 1 to 7.

9. A non-volatile computer storage medium storing computer executable instructions, characterized in that: The computer executable instructions execute the crimping tube detection method based on magnetic flux leakage signals as described in any one of claims 1 to 7.

Images