A method for predicting faults in power supply cables for airport navigation aids

Abstract

The present invention discloses a method for predicting faults in power supply cables for airport navigation aids. An offline test method is adopted and insulation monitoring data of the cables is obtained online in real time; an adaptive grey Markov prediction model based on parameter self-optimization is used to predict the insulation resistance value; based on the dual CT (current transformer) method, the leakage current and dielectric loss factor of the cables are obtained; the data with different time and space scales collected are preprocessed, the high-dimensional matrices of different state variables are obtained, and a high-dimensional random matrix is combined; the empirical spectral distribution of the eigenvalues of the high-dimensional random matrix is calculated, and the corresponding loop is calculated; the loop distributions of the current data and historical state data are compared, and the region where the state evaluation value is located is calculated to judge the equipment state. The realization of fault prediction for power supply cables of navigation aids helps to establish a more perfect and accurate state evaluation and prediction model, make a comprehensive prediction of the cable state, and thus realize the safety monitoring and evaluation of cable insulation.

Description

Technical Field

[0001] The present invention belongs to the field of power equipment status prediction and fault diagnosis, and particularly relates to a method for predicting faults of airport navigation lighting power supply cables. Background Art

[0002] With the rapid development of China's civil aviation industry, more and more airports are currently under construction and expansion. By 2025, there will be about 250 civil aviation transport airports in China to meet the rapidly growing civil aviation transport demand. Airport navigation lighting is an important approach visual aid equipment for airports to ensure that aircraft can take off, land, and taxi smoothly at night and under complex weather conditions. With the increase in flight density, in busy transport airports, due to the limitation of the suspension time, the time available for equipment maintenance after problems occur in the equipment is getting shorter and shorter, the pressure on the maintenance work of navigation lighting is increasing, the number of undetected inspections is increasing, the maintenance force cannot keep up, and the potential safety hazards are on the rise. Once a fault occurs in the navigation lighting guidance system, it will cause major safety hazards and endanger the safety of personnel and property. According to investigations, at present, a large part of the faults in navigation lighting systems at home and abroad are caused by faults in the cables of the navigation lighting circuit. The insulation performance of the cables directly determines whether the airport navigation lighting system can operate normally.

[0003] Therefore, it is very important to carry out research on the state monitoring and prediction of the cables in the navigation lighting power supply circuit, which is of great significance for ensuring air transportation safety. Summary of the Invention

[0004] In order to solve the technical problems existing in the background art, the purpose of the present invention is to provide a method for predicting faults of airport navigation lighting power supply cables, which solves the problem that when predicting cable faults at present, the influence of equipment operating environment factors is not fully considered, and the comprehensive analysis and in-depth mining application of various historical data in state monitoring are lacking, resulting in deviations in fault diagnosis and prediction.

[0005] In order to solve the technical problems, the technical solution of the present invention is as follows:

[0006] A method for predicting faults of airport navigation lighting power supply cables, the method comprising:

[0007] Obtaining the elongation at break, insulation resistance value, leakage current, dielectric loss factor, and temperature data of the airport navigation lighting power supply cable through offline testing and online real-time;

[0008] Preprocessing the elongation at break, insulation resistance value, leakage current, dielectric loss factor, and temperature data, calculating the high-dimensional matrix of different state variables, and constructing a high-dimensional random matrix;

[0009] Calculating the empirical spectral distribution of the eigenvalues of the high-dimensional random matrix and the corresponding loop;

[0010] Based on the obtained loop, compare the loop distribution of the current data with that of the historical state data, calculate the region where the state evaluation value is located, and determine the device state.

[0011] Furthermore, the insulation resistance value is predicted by using an adaptive grey Markov prediction model based on parameter self-optimization. The updated and iterated insulation resistance prediction value is used in the model, which specifically includes:

[0012] For the randomness and non-repeatability of the time series, construct the following adaptive exponential cumulative generating operator;

[0013]

[0014] It is called an adaptive exponential cumulative generating operator (AEAGO), λ ∈ (0, 1] is called the adaptive parameter, and λ(1 - λ) t-j is called the contribution degree of the variable x (1) (j) at time j to the variable at time t, and is called the AEAGO time series constructed by all AEAGOs;

[0015] Adopt an adaptive grey background value, z (1) (t) = (x (1) (t)) 1-u *(x (1) (t - 1)) u , where μ ∈ [0, 1];

[0016] Establish an adaptive odd-period grey model, and based on the Markov chain MC, its residual correction model is given, that is, the adaptive grey Markov correction model;

[0017] The corresponding adaptive odd-period grey difference equation is:

[0018]

[0019] ρ k α j and β j represent the vectors constructed by the control variables, is the harmonic component, and γ is the coefficient of the adaptive grey background value ; the adaptive odd-period grey difference equation can be abbreviated as BP = Y;

[0020]

[0021] P = [γ, ρ0, ρ1, ρ2, α1, β1, …, α T-1 , β T-1 T ​(4)

[0022] Y = [x (0) (2),…,x (0) (n)] T (5)

[0023] The time response function of the adaptive odd-period grey prediction model is as follows:

[0024]

[0025] Among which, the expressions of each parameter are:

[0026]

[0027]

[0028]

[0029] Based on the optimization model of mean absolute percentage error, the ant lion optimization algorithm is used to search for the corresponding adaptive parameters, that is, the insulation resistance value is predicted.

[0030] Furthermore, the ant lion optimization algorithm is used to search for the corresponding adaptive parameters, specifically including:

[0031] S1: Set the initialization parameters of the ant lion algorithm;

[0032] S2: Generate the random positions of ants and ant lions;

[0033] S3: Determine the fitness function, that is, the optimization model;

[0034] S4: Evolve and obtain the offspring;

[0035] S5: Select the best ant lion, that is, the parameter estimation value.

[0036] Furthermore, based on the dual CT method, the leakage current and dielectric loss factor are obtained, specifically including:

[0037] Install a current transformer at both ends of the cable to measure the current flowing through the core wires at the head and end of the cable;

[0038] Subtract the current of the core wire at the end of the cable from the current of the core wire at the head of the cable measured to obtain the leakage current of the cable;

[0039] Use a voltage transformer to measure the voltage of the cable for navigational aids lighting, and combine the leakage current of the cable to calculate the complementary angle of the phase angle difference between the leakage current flowing through the main insulation of the cable and the cable voltage, which is the dielectric loss angle;

[0040] The dielectric loss factor of the main insulation of the cable can be expressed as:

[0041] where δ is the dielectric loss angle, is the phase angle of the leakage current, and θ is the phase angle of the voltage at the head and tail ends.

[0042] Furthermore, the temperature data is obtained through a wireless sensor network.

[0043] Furthermore, the construction of the high-dimensional random matrix is specifically as follows:

[0044] Select N measurable state parameters of the system for T samplings, then all the sampling data forms a matrix with N rows and T columns as:

[0045]

[0046] where the element x of the X matrix ij can be used to represent the value of the i-th measurable state parameter of the navigation aid lighting power supply cable at the j-th sampling moment. When N, T → ∞ and c = NT ∈ (0, 1], X is a high-dimensional random matrix;

[0047] Considering that the sampling time interval is relatively long, the row and column elements are adjusted by block and translation to obtain a better row-column ratio;

[0048] The number of rows of the X matrix remains unchanged, and the T column elements of the matrix are sequentially split into i blocks, then the matrix is divided into i sub-matrices, that is, X = (X 1 , X 2 ,..., X i ), where the number of rows of each sub-matrix is N and the number of columns is T / i. The sub-matrices X 2 ,..., X i are sequentially translated below X 1 to expand into a state data matrix X0, that is:

[0049]

[0050] Then the state data matrix X0 is a matrix with N × i rows and T / i columns, and its row-column ratio is c = N × i 2 / T.

[0051] Furthermore, the state evaluation and prediction utilize the random matrix single-loop theorem. According to the relationship between the average spectral radius MSR and the inner-loop radius, the fault state of the navigation aid lighting power supply cable is judged, and the key state evaluation and fault prediction are carried out according to the preset regional threshold.

[0052] Furthermore, according to the single-ring theorem of random matrices, when the elements of a random matrix are not interfered by external factors, its eigenvalues converge to a circular ring or close to the edge of the inner ring in the complex plane coordinate axes. When the elements of the random matrix are interfered by abnormal factors, its eigenvalues will no longer converge and their distribution tends to the center of the circle. The mean spectral radius MSR describes the distribution of the eigenvalues of the random matrix, reflects the trace of the random matrix, and is the distance of the matrix eigenvalues from the origin in the complex plane. According to the size relationship between MSR and the inner ring radius, the fault state of the navigation aid lighting power supply cable is judged. When MSR is less than the inner ring radius, the system is in an abnormal operating state. Finally, critical state assessment and fault prediction are carried out according to the set regional threshold.

[0053] Furthermore, the specific steps of the single-ring theorem are as follows:

[0054] Let Y = {y i,j} M×N be a random matrix with non-Hermitian characteristics, and each element is a random variable that conforms to independent and identically distributed. Its expectation and variance satisfy E(y i,j ) = 0, E(|y i,j | 2 ) = 1;

[0055] For multiple random matrices Y i (i = 1, 2,..., L) with non-Hermitian characteristics, the matrix product is defined as where Y u,i ∈C M×M is the singular value equivalent matrix of Y i , and C represents the set of complex numbers;

[0056] The row vectors of the matrix product Z satisfy a mean of 0 and a variance of 1. The matrix product Z is transformed to obtain the standard matrix product The empirical spectral density of the standardized matrix can converge to a given limit:

[0057]

[0058] In the formula: f(λ Z ) is the probability density function of the matrix , c is the ratio of the number of rows to the number of columns of , and c ∈ (0, 1]; λ Z is the eigenvalue of ;

[0059] According to the single-ring theorem, in the complex plane, the eigenvalues of L / 2 are roughly uniformly distributed within a circular ring with an outer circle radius r1 = 1 and an inner circle radius r2 = (1 - c)

[0060] Compared with the prior art, the advantages of the present invention are as follows:

[0061] The present invention provides a method for predicting faults in airport navigation lighting power supply cables, which solves the problems that the existing research on cable insulation aging mainly focuses on the aging characteristics under single influencing factors, the monitoring of navigation lighting power supply cables mainly collects the states of electrical parameters, the influence of equipment operating environment factors is not fully considered in the system state diagnosis information, the health assessment of equipment is still limited to the assessment based on the existing real-time monitoring information, the research on the assessment method for the current health status and future development trend of power supply loop equipment is insufficient, and the comprehensive analysis and in-depth mining application of various historical data in state monitoring are lacking, resulting in deviations in fault diagnosis.

[0062] A method for predicting faults in a navigation lighting power supply cable according to the present invention analyzes the elongation at break, insulation resistance value, leakage current data, dielectric loss factor value, and temperature data of the navigation lighting power supply cable obtained through offline testing and online real-time. Considering the problems of low measurement interval and low measurement accuracy of the insulation resistance value of the navigation lighting power supply cable, an adaptive grey Markov prediction model based on parameter self-optimization is used to predict the insulation resistance value. The double CT method is used to obtain the leakage current and dielectric loss factor of the cable. Based on physical models and data-driven modeling, a method for cable fault assessment and prediction based on the theory of the spectral characteristics distribution of random matrices is established, and various historical data in state monitoring are comprehensively analyzed and deeply mined and applied to establish a more perfect and accurate state assessment and prediction model to make a comprehensive prediction of the cable state, thereby realizing the safety monitoring and assessment of cable insulation. Description of the Drawings

[0063] Figure 1 is a schematic diagram of the whole process of predicting the state of a navigation lighting power supply cable in an embodiment of the present invention;

[0064] Figure 2 is a flowchart for predicting the insulation resistance value of a navigation lighting power supply cable in an embodiment of the present invention;

[0065] Figure 3 is a flowchart of the ant lion optimization algorithm in an embodiment of the present invention;

[0066] Figure 4 is a flowchart of the double CT method for a navigation lighting power supply cable in an embodiment of the present invention;

[0067] Figure 5 is a simulation diagram of a cable model for a navigation lighting power supply loop in an embodiment of the present invention;

[0068] Figure 6 is a curve graph of the conductivity of a cable varying with temperature in an embodiment of the present invention;

[0069] Figure 7It is the predicted broken line graph of the insulation resistance value of the power supply cable for the navigation aid lights in the embodiment of the present invention;

[0070] Figure 8 is the leakage current waveform graph of the cable in different states in the embodiment of the present invention;

[0071] Figure 9 is the single-loop graph of the cable in different states in the embodiment of the present invention;

[0072] Figure 10 It is the equipment state distribution determination graph based on the Gaussian random matrix theory in the embodiment of the present invention. Specific Embodiments

[0073] The following describes the specific embodiments of the present invention in conjunction with the embodiments:

[0074] It should be noted that the structures, ratios, sizes, etc. shown in this specification are only used to cooperate with the content disclosed in the specification for those familiar with this technology to understand and read, and are not used to limit the implementation conditions of the present invention. Any modification of the structure, change of the proportional relationship or adjustment of the size, without affecting the effects that the present invention can produce and the purposes that can be achieved, should still fall within the scope covered by the technical content disclosed in the present invention.

[0075] At the same time, the terms such as "upper", "lower", "left", "right", "middle" and "one" cited in this specification are only for the convenience of clear narration, and are not used to limit the scope of implementation of the present invention. The change or adjustment of their relative relationship, without substantial change in the technical content, should also be regarded as the scope of implementation of the present invention.

[0076] Embodiment 1:

[0077] As Figure 1 shown, the prediction method of the present invention can be divided into three parts: data acquisition, data processing and analysis, and state assessment and prediction:

[0078] Data acquisition mainly obtains the elongation at break, insulation resistance value, leakage current, dielectric loss factor, and temperature data of the power supply cable for navigation aid lights through offline testing and online real-time.

[0079] Data processing and analysis utilize the high-dimensional random matrix spectrum analysis theory. Research shows that the random matrix spectrum analysis theory has obvious advantages in processing high-dimensional data. The random matrix spectrum theory mainly studies the asymptotic behavior of the sample eigenvalues, sample eigenvectors of random matrices, and their statistical functions under general asymptotic systems. RMT integrates various data into a high-dimensional matrix, studies the spatio-temporal correlation and volatility of data through data modeling and data analysis in high-dimensional space, and studies the characteristics of the matrix and the data distribution from the perspective of probability and statistics. The present invention conducts research on the state monitoring and fault prediction of navigation aid lights based on the random matrix spectrum analysis theory and in combination with historical operation data driven by data.

[0080] The state assessment prediction uses the single ring theorem of random matrices. When the elements of the random matrix are not disturbed by external factors, their eigenvalues ​​converge to a ring or close to the edge of the inner ring in the complex plane coordinate axis. When the elements of the random matrix are disturbed by abnormal factors, their eigenvalues ​​will no longer converge, and their distribution tends to the center of the circle. The mean spectral radius (MSR) describes the distribution of the eigenvalues ​​of the random matrix, which can reflect the trace of the random matrix. It is the distance of the matrix eigenvalue from the origin on the complex plane. According to the relationship between the MSR and the inner ring radius, the fault state of the navigation light power supply cable can be judged. When the MSR is smaller than the inner ring radius, the system is in an abnormal operating state. Finally, key state assessment and fault prediction are performed according to the set regional threshold.

[0081] The fault prediction of the power supply cable of the airport navigation lighting includes the following steps:

[0082] S1. Use offline testing method and online real-time to obtain cable insulation monitoring data.

[0083] S2. For step S1, the offline index data includes the elongation at break index of the in-service cable sample material obtained in the laboratory, the insulation resistance value obtained on site, and the leakage current, dielectric loss factor and temperature data of the navigation light power supply cable obtained online in real time.

[0084] S3, testing the elongation at break of the sampled cables is currently the most reliable method recognized by the industry to assess the degree of cable aging. Considering the low measurement interval and measurement accuracy of the insulation resistance value in step S2, an adaptive gray Markov prediction model based on parameter self-optimization is used to predict the insulation resistance value. The updated iterative insulation resistance prediction value is used in the model.

[0085] like Figure 2 As shown in the figure, the specific steps for predicting the insulation resistance value of the navigation light power supply cable are:

[0086] S301 constructs the following adaptive exponential accumulation generation operator in view of the randomness and non-repeatability of the time series;

[0087]

[0088] is called the adaptive exponential accumulation generation operator (AEAGO), λ∈(0,1] is called the adaptive parameter, and λ(1-λ) t-j is called the variable x at time j (1) (j) For the variable at time t The contribution of is the AEAGO time series constructed from all AEAGOs.

[0089] S302 adopts a new adaptive gray background value, z (1) (t) = (x (1) (t)) 1-u *(x (1) (t - 1)) u , where μ ∈ [0, 1];

[0090] S303 establishes an adaptive odd - period gray model. To further improve the prediction accuracy, a residual correction model is given based on the Markov chain (MC), that is, an adaptive gray - Markov correction model, to improve the prediction accuracy;

[0091] The corresponding adaptive odd - period gray difference equation is:

[0092]

[0093] ρ k , α j and β j represent vectors constructed by control variables, is the harmonic component, γ is the adaptive gray - level background value coefficient. The adaptive odd - period gray difference equation can be abbreviated as BP = Y;

[0094]

[0095] P = [γ, ρ0, ρ1, ρ2, α1, β1, …, α T-1 , β T-1 T (4)

[0096] Y = [x (0) (2), …, x (0) (n)] T (5)

[0097] The time - response function of the adaptive odd - period gray prediction model is:

[0098]

[0099] The expressions of each parameter therein:

[0100]

[0101]

[0102]

[0103] S304 is based on an optimization model of the mean absolute percentage error, and the ant - lion optimization algorithm is used to search for the corresponding adaptive parameters;

[0104] Such as​Figure 3 As shown in the figure, the overall process of the antlion optimization algorithm is as follows:

[0105] (1) Set the initialization parameters of the antlion algorithm;

[0106] (2) Generate the random positions of ants and antlions;

[0107] (3) Determine the fitness function, i.e., the optimization model;

[0108] (4) Evolve and obtain offspring;

[0109] (5) Select the best antlion, i.e., the parameter estimation value.

[0110] S4. Based on the dual CT (current transformer) method, obtain the leakage current and the dielectric loss factor in step S1;

[0111] As Figure 4 shown, the specific steps of the dual CT method are as follows:

[0112] S401. Install a current transformer (CT) at both ends of the cable to measure the current flowing through the core of the cable head and the end;

[0113] S402. Subtract the current flowing through the core of the cable end from the current flowing through the core of the cable head measured, and the leakage current of the cable can be obtained;

[0114] S403. Use a voltage transformer to measure the voltage of the cable for navigation light power supply. Combining with step S402, the complementary angle of the phase angle difference between the leakage current flowing through the main insulation of the cable and the cable voltage can be calculated, which is the dielectric loss angle;

[0115] S404. The dielectric loss factor of the main insulation of the cable can be expressed as:

[0116] where δ is the dielectric loss angle, is the phase angle of the leakage current, and θ is the phase angle of the head and end voltages;

[0117] Figure 5As shown in the figure, the dynamic simulation platform Simulink of Matlab software is used. Mainly, the SimPowerSystem toolbox is utilized to establish a simulation model of the navigational lighting circuit. The navigational lighting circuit is a floating circuit composed of a series of isolation transformers connected in series. The circuit is powered by a constant current source. The actual parameters of the power supply cable for navigational lights are selected for simulation to simulate the real situation. The navigational lights are embedded lamps with a rated power of 100 W and a rated current of 6.6 A. If regarded as a pure resistive load, its static resistance is calculated to be approximately 2.29 ohms. In this simulation system, a resistor with a resistance value of 2.29 ohms is used instead. An isolation transformer with a rated power of 100 VA and a rated current of 6.6 A is selected. Finally, the double CT method is adopted to obtain the leakage current value and the dielectric loss factor of the cable.

[0118] S5. Preprocess the data with different time and space scales collected in step S2, calculate the high-dimensional matrices of different state variables, and combine them into a high-dimensional random matrix;

[0119] As Figure 6 shown, the operating temperature of the cable will change continuously with the changes of the external environment and its own working state. Temperature has a significant impact on the cable insulation parameters, and the quality of insulation will also affect the change of the cable's own temperature. The navigational lighting cable is usually buried underground, so the cable temperature during normal operation is generally constant. When the cable insulation fails, it will cause the cable temperature to rise during operation. When the cable temperature exceeds the rated temperature, the conductivity will increase exponentially with the increase of temperature. The change of conductivity will affect the cable insulation resistance value, thus affecting the accuracy of cable insulation on-line monitoring. Therefore, it can be used as an important basis for judging whether the cable has failed.

[0120] S6. Calculate the empirical spectral distribution of the eigenvalues of the high-dimensional random matrix obtained in step S5, and calculate the corresponding ring;

[0121] S7. According to the single-ring theorem of random matrices, when the elements of the random matrix are not interfered by external factors, its eigenvalues converge to a ring or close to the inner ring edge in the complex plane coordinate axis. When the elements of the random matrix are interfered by abnormal factors, its eigenvalues will no longer converge, and its distribution tends to the center of the circle. The mean spectral radius (MSR) describes the distribution of the eigenvalues of the random matrix, can reflect the trace of the random matrix, and is the distance of the matrix eigenvalues from the origin in the complex plane. According to the size relationship between the MSR and the inner ring radius, the fault state of the navigational lighting power supply cable can be judged. When the MSR is less than the inner ring radius, the system is in an abnormal operating state. Finally, key state assessment and fault prediction are carried out according to the set regional threshold.

[0122] The specific steps of the single-ring theorem are as follows:

[0123] Let Y = {yi,j} M×N is a random matrix with non-Hermitian characteristics, and each element is an independent and identically distributed random variable. Its expectation and variance satisfy E(yi,j) = 0 and E(|y i,j | 2 ) = 1.

[0124] For multiple random matrices Yi (i = 1, 2,..., L) with non-Hermitian characteristics, the matrix product is defined as where Y u,i ∈C M×M is the singular value equivalent matrix of Y i , and C represents the set of complex numbers.

[0125] The row vectors of the matrix product Z have a mean of 0 and a variance of 1. The matrix product Z is transformed to obtain the standard matrix product The standardized matrix 's empirical spectral density can converge to a given limit:

[0126]

[0127] In the formula: f(λ Z ) is the probability density function of the matrix . c is the ratio of the number of rows to the number of columns of , and c ∈ (0, 1]; λ Z is the eigenvalue of .

[0128] According to the single-ring theorem, in the complex plane, 's eigenvalues are approximately uniformly distributed within the annulus with an outer circle radius r1 = 1 and an inner circle radius r2 = (1 - c) L / 2 .

[0129] Example 2:

[0130] Taking the insulation resistance value data of the actual navigation aid lighting power supply cable as an example, as Figure 7 is the predicted broken line graph of the insulation resistance value of the navigation aid lighting power supply cable. The actual parameters of the navigation aid lighting power supply cable are selected for simulation, and the leakage current waveforms of the cable under different states are shown in Figure 8. According to the standard regulations for buried cables in the airport navigation aid lighting circuit, the tanδ of the navigation aid lighting power supply cable ≤ 0.008, and the maximum rated temperature of the conductor of the airport navigation aid lighting power supply cable is 90°C. The elongation at break test is carried out on the sampled cable, which is the most reliable method currently recognized in the industry to evaluate the aging degree of the cable. The national standard for the airport navigation aid lighting power supply cable stipulates that the minimum elongation at break of the navigation aid lighting power supply cable is 250%.

[0131] The occurrence of cable faults is a degradation and gradual change process. When collecting on-site temperature through a wireless sensor network and combining it with monitored electrical parameters such as voltage and current for testing, a large number of sampled data constitute a high-dimensional data state matrix.

[0132] The obtained elongation at break, insulation resistance value, leakage current data, dielectric loss factor value, and temperature data are formed into an original data matrix as shown in Table 1:

[0133] Table 1 Cable state quantity matrix

[0134] State variable Sampling period Original matrix Insulation resistance 30 min 1×4000 Leakage current 10 min 1×4000 Dielectric loss factor 10 min 1×4000 Temperature 10 min 1×4000 Elongation at break 10 min 1×4000

[0135] If the 5 measurable state parameters of the system are sampled 4000 times, then all sampled data can form a matrix with 5 rows and 4000 columns as follows:

[0136]

[0137] The matrix X is divided into 20 blocks and then translated and extended to form a state data matrix X0 as follows:

[0138]

[0139] The adjusted row-column ratio is 0.5.

[0140] The power supply circuit of airport navigation lights has the characteristics of long cable lines, underground laying, and numerous joints. From on-site investigations and accident cases, it is found that cable insulation damage will lead to single-point grounding of the power supply circuit, and in severe cases, two or even multiple points of grounding will occur. The common faults of the main circuit of navigation light cables can be divided into cable open-circuit faults, cable grounding faults, and composite faults containing both of these faults.

[0141] From the perspective of the scope of influence of faults, open-circuit faults directly affect the lighting of navigation lights and may cause the airport to close in severe cases. When there is only one grounding point for grounding faults, it will not affect the operation of the lights. If there are more than two fault points, it will cause a section of the lights to be unlit or dim. Composite faults can cause the closure of navigation lights at any time and are the greatest hidden danger to flight safety.

[0142] The ring distribution under different cable fault types is obtained by applying the single-ring theorem as shown in Figure 9.

[0143] The single ring in the normal state without faults of the navigation light power supply cable is shown in Figure 9(a). The singular value points of the high-dimensional random matrix are evenly distributed between the outer ring and the inner ring, and the radius of the inner ring is 0.7077. The average spectral radius of the singular values is 0.8384, which is greater than the radius of the inner ring, and the distribution of the singular values conforms to the rules of the single-ring theorem.

[0144] When a single loop is needed for fault warning of the navigation aid lighting power supply cable, as shown in Figure 9(b), the inner loop radius is 0.7077, and the average spectral radius of singular values is 0.7583, which is close to the inner loop radius. At this time, it means that the navigation aid lighting cable has a fault, but it will not affect the normal operation of the entire navigation aid lighting system. Therefore, in this case, fault warning is required.

[0145] When a single loop occurs when the navigation aid lighting power supply cable fails, as shown in Figure 9(c), the singular value points are completely contracted within the range of the center of the complex plane. The inner loop radius is 0.7077, and the average spectral radius of singular values is 0.1527, which is much smaller than the inner loop radius. The distribution of the singular value points does not conform to the single loop theorem, so it can be determined that the navigation aid lighting power supply cable has a fault.

[0146] Based on the single loop law, key state evaluation and fault prediction are carried out according to the set regional threshold as Figure 10 shown. The cable state determination is shown in Table 2:

[0147] Table 2 Navigation aid lighting power supply cable state determination:

[0148]

[0149] The preferred embodiments of the present invention have been described in detail above, but the present invention is not limited to the above embodiments. Within the scope of knowledge possessed by those of ordinary skill in the art, various changes can be made without departing from the purpose of the present invention.

[0150] Many other changes and modifications can be made without departing from the concept and scope of the present invention. It should be understood that the present invention is not limited to specific embodiments, and the scope of the present invention is defined by the appended claims.

Claims

1. A method for predicting the faults of airport navigation lighting power supply cables, characterized in that, The method includes: Obtaining the elongation at break, insulation resistance value, leakage current, dielectric loss factor and temperature data of the navigation aid lighting power supply cable through off-line testing and on-line real-time; The insulation resistance value is predicted by using an adaptive grey Markov prediction model based on parameter self-optimization. The updated iterative insulation resistance prediction value is used in the model. The specific steps are as follows: In view of the randomness and non-repeatability of the time series, an adaptive exponential cumulative generating operator is constructed, an adaptive grey background value is adopted, an adaptive odd-period grey model is established, and its residual correction model, namely the adaptive grey Markov correction model, is given based on the Markov chain; an optimization model based on the mean absolute percentage error is used, and the ant lion optimization algorithm is adopted to search for the corresponding adaptive parameters, that is, the insulation resistance value is predicted; Preprocess the elongation at break, insulation resistance value, leakage current, dielectric loss factor and temperature data, calculate the high-dimensional matrix of different state variables, and construct a high-dimensional random matrix; Calculate the empirical spectral distribution of the eigenvalues of the high-dimensional random matrix and the corresponding ring; Based on the obtained ring, compare the ring distribution of the current data with the ring distribution of the historical state data, calculate the area where its state evaluation value is located, and judge the equipment state; According to the single-ring theorem of the random matrix, when the elements of the random matrix are not interfered by external factors, its eigenvalues converge to a circular ring or close to the inner ring edge in the complex plane coordinate axis. When the elements of the random matrix are interfered by abnormal factors, its eigenvalues will no longer converge and its distribution tends to the center of the circle; the mean spectral radius MSR describes the distribution of the eigenvalues of the random matrix, reflects the trace of the random matrix, and is the distance of the matrix eigenvalues from the origin in the complex plane. According to the size relationship between the MSR and the inner ring radius, judge the fault state of the navigation aid lighting power supply cable. When the MSR is less than the inner ring radius, the system is in an abnormal operating state. Finally, perform key state evaluation and fault prediction according to the set regional threshold.

2. The method for predicting the failure of an airport navigation lighting power supply cable according to claim 1, characterized in that The insulation resistance value is predicted by using an adaptive grey Markov prediction model based on parameter self-optimization. The updated iterative insulation resistance prediction value is used in the model, and it specifically includes: In view of the randomness and non-repeatability of the time series, construct the following adaptive exponential cumulative generating operator; It is called the Adaptive Exponential Accumulation Generation Operator (AEAGO), and λ ∈ (0, 1] is called the adaptive parameter, and λ(1 - λ) t-j is called the variable x at time j (1) (j)'s contribution to the variable at time t , and is called the AEAGO time series constructed by all AEAGOs; Adopt an adaptive gray background value, z (1) (t) = (x (1) (t)) 1-u *(x (1) (t - 1)) u , where μ ∈ [0, 1]; Establish an adaptive odd-period grey model, and based on the Markov chain MC, its residual correction model, namely the adaptive grey Markov correction model, is given; The corresponding adaptive odd-period grey difference equation is: ρ k , α j and β j represent vectors constructed from control variables, is the harmonic component, and γ is the coefficient of the adaptive gray background value ; the adaptive odd-period gray difference equation can be abbreviated as BP = Y; P = [γ, ρ0, ρ1, ρ2, α1, β1,..., α T-1 , β T-1 T (4)​ Y = [x (0) (2), …, x (0) (n)] T (5) The time response function of the adaptive odd-period grey prediction model is: The expressions of each parameter therein: Based on the optimization model of the mean absolute percentage error, and adopt the ant lion optimization algorithm to search for the corresponding adaptive parameters, that is, the insulation resistance value is predicted.

3. A method for predicting faults in an airport navigation lighting power supply cable according to claim 1, characterized in that, Adopt the ant lion optimization algorithm to search for the corresponding adaptive parameters, and it specifically includes: S1: Set the initialization parameters of the ant lion algorithm; S2: Generate the random positions of the ants and ant lions; S3: Determine the fitness function, that is, the optimization model; S4: Evolve and obtain the offspring; S5: Select the best ant lion, that is, the parameter estimation value.

4. A method for predicting faults in airport navigation lighting power supply cables according to claim 1, characterized in that Based on the double CT method, obtain the leakage current and dielectric loss factor, and it specifically includes: Install a current transformer at both ends of the cable to measure the current flowing through the core of the cable at the head end and the end end; Subtract the current of the cable end core measured from the current of the cable head end core to obtain the leakage current of the cable; Use a voltage transformer to measure the voltage of the cable for the navigation aid lights, and calculate the complementary angle of the phase angle difference between the leakage current flowing through the main insulation of the cable and the cable voltage in combination with the leakage current of the cable, which is the dielectric loss angle; The dielectric loss factor of the main insulation of the cable can be expressed as: where δ is the dielectric loss angle, is the phase angle of the leakage current, and θ is the phase angle of the voltage at the head and tail ends.

5. A method for predicting faults in an airport navigation lighting power supply cable according to claim 1, characterized in that, The temperature data is obtained through a wireless sensor network.

6. A method for predicting faults in an airport navigation lighting power supply cable according to claim 1, characterized in that, The construction of the high-dimensional random matrix is specifically as follows: Select N measurable state parameters of the system for T times of sampling, then the matrix composed of all sampling data with N rows and T columns is: Among them, the element x of the X matrix ij can be used to represent the value of the i-th measurable state parameter of the navigation aid lighting power supply cable at the j-th sampling moment. When N, T → ∞ and c = N / T ∈ (0, 1], X is a high-dimensional random matrix; Considering that the sampling time interval is relatively long, the row-column ratio is optimized by block and translation adjustment of the row-column elements; The number of rows of the X matrix remains unchanged. The T columns of elements of the matrix are split into i blocks in sequence, and the matrix is divided into i submatrices, that is, X = (X 1 , X 2 ,..., X i ), where the number of rows of each submatrix is N and the number of columns is T / i. The submatrices X 2 ,..., X i are translated to the bottom of X 1 in sequence and expanded into the state data matrix X0, that is: Then the state data matrix X0 is a matrix with N×i rows and T / i columns, and its row-column ratio is c = N×i 2 / T.

7. The method for predicting the fault of the airport navigation lighting power supply cable according to claim 1, characterized in that, The state evaluation and prediction uses the single-ring theorem of the random matrix. According to the size relationship between the average spectral radius MSR and the inner ring radius, the fault state of the cable for the navigation aid lights is judged, and the key state evaluation and fault prediction are carried out according to the preset regional threshold.

8. A method for predicting faults in an airport navigation lighting power supply cable according to claim 7, characterized in that The specific steps of the single-ring theorem are as follows: Let \(Y = \{y i,j \}\ M×N be a random matrix with non-Hermitian eigenvalues, where each element is an independent and identically distributed random variable, and its expectation and variance satisfy \(E(y i,j ) = 0\), \(E(|y i,j | 2 ) = 1\); For a random matrix Y with multiple non-Hermitian eigenvalues i (i = 1, 2,..., L), define the matrix product as where Y u,i ∈ C M×M is the singular value equivalent matrix of Y i , and C represents the set of complex numbers; The row vectors of the matrix product Z satisfy a mean of 0 and a variance of 1, and the matrix product Z is transformed to obtain a standard matrix product The standardized matrix The empirical spectral density of can converge to a given limit: where: f(λ Z ) is the probability density function of the matrix , c is the ratio of the number of rows to the number of columns of , and c ∈ (0, 1]; λ Z is the eigenvalue of; According to the single-ring theorem, in the complex plane, the eigenvalues are approximately uniformly distributed within the annulus with outer circle radius r1 = 1 and inner circle radius r2 = (1 - c). L / 2 ​

Images