CN107203746A - A kind of switch breakdown recognition methods - Google Patents

Abstract

The present invention provides a fault identification method for a turnout, wherein the method includes: collecting each action curve of a turnout; dividing the obtained action curves of a turnout into normal curves and fault curves; The representative curve is the representative curve of this type of curve; use the similarity algorithm to calculate the similarity between the curve to be identified and the normal representative curve 1, the similarity between the curve to be identified and the fault representative curve 2; compare the calculated similarity, if similar If the degree of similarity 1 is greater than the degree of similarity 2, the curve is a normal curve, and if the degree of similarity 1 is smaller than the degree of similarity 2, the curve is a fault curve. The present invention solves the problem in the prior art of judging the fault type of a turnout by manual experience, resulting in missed and false alarms, thereby realizing the automatic identification of the fault type of the turnout, improving maintenance efficiency and system reliability, and ensuring driving safety.

Description

A fault identification method for a turnout

technical field

The invention relates to the field of rail transit, in particular to a method for identifying a turnout failure.

Background technique

Turnout is an essential line equipment when a train transfers from one track to or crosses another track. It is an important part of the railway track and is also the equipment with the highest failure rate. Once the turnout breaks down and cannot complete the prescribed action, it will temporarily stop for several hours, which will delay the time of a large number of passengers;

At present, the microcomputer monitoring system is mainly used in my country to monitor the state of the turnout. From the actual application situation on the spot, it is mainly through manual observation of the current curve of the turnout collected by the microcomputer monitoring system to identify the fault type, which requires a large number of staff; the identification accuracy It mainly depends on whether the relevant personnel have rich work experience, which makes false positives and false alarms appear from time to time; it takes time for the staff to analyze the characteristics of the curve and judge the type of fault, which cannot be identified online in real time, and the efficiency is low. In addition, in order to further prevent the occurrence of accidents, the railway department arranges relevant personnel to regularly inspect and repair the turnouts, which requires a lot of manpower and material resources. This diagnosis method of turnout failure is no longer suitable for the rapid development of high-speed railways. How to quickly and accurately judge the type of turnout failure is an important measure to ensure the safety of traffic and the safety of passengers.

In the prior art, it is not possible to automatically identify the fault type of the turnout, and no quick and effective solution has been proposed.

Contents of the invention

The invention provides a method for identifying a switch fault, which at least solves the problem in the prior art of judging the fault type of a switch based on manual experience, resulting in missed and false alarms.

A kind of switch fault identification method that the present invention proposes, comprises the following steps:

(1): collect the action curve of each turnout;

(2): Divide the obtained action curves of each turnout into normal curves and fault curves;

(3): For the normal curve and each type of fault curve, select a curve with the most representative characteristics as the normal representative curve and each type of fault representative curve;

(4): Use the similarity algorithm to calculate the similarity between the curve to be identified and the normal representative curve 1, the similarity between the curve to be identified and the fault representative curve 2; the similarity algorithm is a dynamic time warping algorithm or an algorithm based on Fresche distance;

The similarity algorithm is a dynamic time warping algorithm, specifically:

(4a1): The curve to be identified can be expressed as T={T(1), T(2),...,T(n),...,T(N)}, n is the time series label of the time series, n= 1 is the starting point of the time series, n=N is the end point of the time series, and T(n) is the value of the time series;

(4b1): The normal representative curve and the fault representative curve can be expressed as R={R(1), R(2),...,R(m),...,R(M)}, where m is the time sequence of the time series label, m=1 is the starting point of the time series, m=M is the end point of the time series, and R(m) is the value of the time series;

(4c1): Mark each time series number n of the curve time series to be identified on the horizontal axis, mark each time series number m representing the time series of the curve on the vertical axis, and draw some vertical and horizontal lines through the integer coordinates of these time series numbers to form a Network, all grid points are (1, 1), ..., (n, m), ..., (N, M), search for the optimal path from (1, 1) to (N, M);

(4d1): After the path passes through (n, m), the next grid point to pass can only be (n, m+1), (n+1, m), (n+1, m+1), select ( The minimum distance from n, m) to the next grid point is the optimal path, and the accumulated minimum distance from (1, 1) to (N, M) is calculated;

(4e1): Calculate the Euclidean distance between the time series T of the curve to be identified and the time series R of the representative curve;

(4f1): The total accumulated distance from the start point (1, 1) to the end point (N, M) is the cumulative minimum distance from the start point (1, 1) to the end point (N, M), the time series T of the curve to be identified and the representative curve The sum of the Euclidean distances between the time series R;

(4g1): Negating the total accumulated distance, indicating the similarity between the curve to be identified and the normal representative curve or the curve to be identified and the fault representative curve;

Or: the similarity algorithm is an algorithm based on Fresche distance, specifically:

(4a2): The curve L ₁ to be identified can be expressed as P={P(1), P(2),..., P(n),..., P(N)}, P(n)=(x _n ,y _n ), n is the serial number of the sampling point on the curve L ₁ , n=1 is the starting sampling point, n=N is the end sampling point, x _n is the abscissa of the nth sampling point, x _n is the The vertical coordinates of n sampling points;

(4b2): Normal representative curve or fault representative curve L2 can be expressed as P'={P'(1), P'( ₂ ),...,P'(m),...,P'(M)} , P′(m)=(x′ ₀ , y′ _m ), m is the serial number of the sampling point on the curve L ₂ , m=1 is the starting sampling point, m=M is the ending sampling point, x’ _m is The abscissa of the mth sampling point, y' _m is the ordinate of the mth sampling point;

(4c2): Calculate the distance between each sampling point _on L2 and each sampling point _on L2, and obtain the distance matrix D as follows:

1≤m≤M, 1≤n≤N

in the above formula Indicates the distance from the mth sampling point _on the curve L2 to the nth sampling point on the curve L2 _;

(4d2): Select the maximum distance d _max =max(D) and the minimum distance d _min =min(D) in the distance matrix D, initialize the target distance f=d _min , and set the cycle interval

(4e2): Set the elements less than or equal to f in the distance matrix D to 1, and set the elements greater than f to 0, so as to obtain the binary matrix D′ as follows:

1≤m≤M, 1≤n≤N

(4f2): Search for a path that satisfies the following conditions in the binary matrix D′: the starting point of the path is d' ₁₁ , the end point of the path is d' _MN , and after the path passes through the point d' _mn , the next passing point is only It can be one of d' _m+1,n , d' _m,n+1 , d' _m+1,n+1 , and the value of all points in the path must be 1;

(4g2): If no path satisfying the condition is found in step (4f2), set the target distance f=f+res, then repeat steps (4e2) and (4f2), if a path satisfying the condition is found in step (4f2) Path or target distance f=d _max , enter the next step;

(4h2): Frechet distance Frechet=f between the curve to be identified and the normal representative curve or fault representative curve, Indicates the similarity between the curve to be identified and the normal representative curve or fault representative curve;

(5): Comparing the similarity obtained in the calculation step (4), if the similarity 1 is greater than the similarity 2, then the curve is a normal curve, and if the similarity 1 is smaller than the similarity 2, then the curve is a failure curve.

In the present invention, each action curve of the turnout collected in step (1) is the action curve data or image of the turnout generated in the microcomputer monitoring system, or the action curve data or image of the turnout in the paper file.

In the present invention, each action curve of the turnout collected in the step (1) is the action current curve data or image of the turnout; or the action power curve data or image of the turnout.

In the present invention, the obtained turnout action curve is divided into a normal curve and a fault curve as described in step (2), and the fault curve is specifically divided into: a disconnection curve of the starting circuit, a sudden stop rotation curve after the turnout is started, and a curve with Foreign matter curve, point machine stator-rotor mixed line curve, automatic switch action inflexible curve, point machine start-up delay curve, locking current exceeding the standard curve and turnout operating current in a zigzag curve; for each type of fault curve, Select a curve with the most representative characteristics as the representative curve of this type of fault curve; and calculate the similarity between the curve to be identified and the representative curve of each type of fault respectively; the type of curve with the highest similarity is the curve to be identified type of failure.

In the present invention, before the step (3), the obtained switch action curve is preprocessed, including the following steps:

(1): Get the mean value between the R, G, and B components of each pixel in the colored turnout action curve image as the gray value of the pixel, and convert the colored turnout action curve image into a grayscale image;

(2): A threshold is set so that the pixel point whose gray value is greater than the threshold value is 1, and the pixel point whose gray value is smaller than the threshold value is 0, and the gray image is converted into a binary image;

(3): Find the target area enclosed by the coordinate axes, remove the free and isolated pixels in the target area, and smooth the edge of the curve to remove noise;

(4): Each moment corresponds to a value, and its pixel value is 0, and the situation that there are multiple pixel points in a column is 0 is refined;

(5): Through function transformation, extract the coordinates of each point on the curve;

(6): Scale the coordinates of each point proportionally so that the horizontal and vertical coordinates of each point are within the same range.

Among the present invention, in step (3), for normal curve, select any one curve to be normal representative curve; For each type of fault curve, select any one curve in this type of fault curve to be the representative curve of this type of fault curve.

In summary, the beneficial effects of the present invention are:

(1) Each action curve of the turnout is collected in the microcomputer monitoring system, and the failure of the turnout can be identified without additional installation of other devices, which is economical, convenient, and practical.

(2) Preprocessing the obtained turnout action curves can not only eliminate interference such as grids and noise, and improve the accuracy of turnout fault identification; it can also analyze turnouts from different computer monitoring systems, different railway bureaus, and different weather The action curve is used for fault identification, so that the method of the present invention has a wide application range, is not limited to certain small areas, and has strong adaptability.

(3) Using the dynamic time warping algorithm and the Fresche distance algorithm, it does not require a large amount of historical data and expert knowledge base, and can identify the type of turnout fault by selecting the representative curve arbitrarily, reducing the difficulty of identification and reducing the need for correlation Professional needs.

(4) The automatic identification of switch faults is realized, which solves the low efficiency and unreliability caused by manual experience in judging the type of switch faults, saves a lot of manpower and material resources, and improves the accuracy of judgment.

Description of drawings

In order to more clearly illustrate the specific implementation of the present invention or the technical solutions in the prior art, the following will briefly introduce the accompanying drawings that need to be used in the specific implementation or description of the prior art. Obviously, the accompanying drawings in the following description The drawings show some implementations of the present invention, and those skilled in the art can obtain other drawings based on these drawings without any creative work.

Fig. 1 is a flow chart of a method for identifying a fault in a turnout according to an embodiment of the present invention;

Fig. 2 is a flow chart of a method for identifying a fault of a turnout according to Embodiment 1 of the present invention;

Fig. 3 is a similarity histogram between a curve to be identified and each representative curve according to Embodiment 1 of the present invention;

Fig. 4 is a flowchart of a method for identifying a fault of a turnout according to Embodiment 2 of the present invention;

Fig. 5 is according to embodiment 2 of the present invention, the normal representative curve and 8 kinds of fault representative curve images selected after preprocessing the acquired switch action current curve; (c) is the sudden stop rotation curve after the turnout is started, (d) is the foreign body curve in the turnout, (e) is the mixed line curve of the stator and rotor of the switch machine, and (f) is the inflexible curve of the automatic switch , (g) is the switch machine start-up delay curve, (h) is the locking current exceeding the standard curve, (i) is the zigzag curve of the switch action current;

Fig. 6 is a histogram of the similarity between a curve to be identified and representative curves according to Embodiment 2 of the present invention.

detailed description

The technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Apparently, the described embodiments are some of the embodiments of the present invention, but not all of them. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

In the description of the present invention, it should be noted that the terms "first", "second", and "third" are used for description purposes only, and should not be understood as indicating or implying relative importance.

In addition, the technical features involved in the different embodiments of the present invention described below may be combined with each other as long as there is no conflict with each other.

Example 1

In this embodiment, a method for identification of a turnout failure is provided. Fig. 1 is a flow chart of a method for identifying a failure of a turnout according to an embodiment of the present invention. As shown in Fig. 1, the flow chart includes the following steps:

Step S11: collect each action curve of the turnout;

Step S12: Divide the obtained switch action curves into normal curves and fault curves;

Step S13: For the normal curve and the fault curve, select a curve with the most representative characteristics as the representative curve of this type of curve;

Step S14: use the similarity algorithm to calculate the similarity 1 between the curve to be identified and the normal representative curve, and the similarity 2 between the curve to be identified and the fault representative curve;

Step S15: Compare the calculated similarities. If the similarity 1 is greater than the similarity 2, the curve is a normal curve. If the similarity 1 is smaller than the similarity 2, the curve is a faulty curve.

Through the above steps, the fault category of the collected turnout action curve will be automatically identified. Compared with the prior art, the inefficiency and unreliability caused by manual experience to judge the fault type of the turnout are compared. The above steps solve the problem in the prior art. Judging the fault type of the turnout by manual experience leads to the problem of missing and false alarms, thereby realizing the automatic identification of the fault type of the turnout, improving the maintenance efficiency and system reliability, and ensuring driving safety.

Fig. 2 is the flow chart of the fault identification method of a turnout according to Embodiment 1 of the present invention. As shown in Fig. 2, the method comprises the following steps: collecting each action power curve of the turnout in the microcomputer monitoring system; Classification can be divided into normal curves and fault curves. The fault curves can be further divided into: start circuit disconnection curve, turnout sudden stop rotation curve after start-up, turnout with foreign body curve, automatic switch action inflexible curve and lock Closed current exceeding the standard curve; for each type of turnout action power curve, randomly select one such turnout action power curve as the representative curve of this type of power curve; use the dynamic time warping algorithm to calculate the power curve to be identified and the normal representative curve and various The distance of the fault representative power curve; invert the calculated distance to indicate the similarity between the curve to be identified and the normal representative curve and various fault representative curves; compare the similarity between the curve to be identified and the normal representative curve and various fault representative curves, The type of curve with the highest similarity is the type of the curve to be identified.

The following description will be made in conjunction with a specific optional embodiment.

(1): Collect the power curve of each action of the turnout in the microcomputer monitoring system;

(2): Classify the obtained turnout action power curves, which can be divided into normal curves and fault curves. The fault curves can be further divided into: starting circuit disconnection curves, turnout sudden stop rotation curves after starting, and foreign objects in the turnout Curve, inflexible curve of automatic switch and lock current exceeding the standard curve;

(3): For each type of turnout action power curve, arbitrarily select one such turnout action power curve as the representative curve of this type of power curve;

(4): Use the dynamic time warping algorithm to calculate the distance between the power curve to be identified and the normal representative curve and various fault representative power curves. The steps are as follows:

(1) Take a curve to be identified, which can be expressed as T={T(1), T(2),..., T(219)}, T(1)=0, T(2)=1.84684,... , T(219)=2.29189;

(2) The normal representative curve can be expressed as R={R(1), R(2),...,R(144)}, R(1)=0, R(2)=5.435897,...,R( 144) = 0.090598;

(3) Mark each time series number 219 of the curve time series to be identified on the horizontal axis, mark each time series number 144 representing the time series of the curve on the vertical axis, and draw some vertical and horizontal lines through the integer coordinates of these time series numbers to form a network , all grid points are (1, 1), ..., (219, 144), search for the optimal path from (1, 1) to (219, 144);

(4): After the path passes through (1, 1), the next passing grid point can only be (1, 2), (2, 1), (2, 2), and the calculation can be obtained from (1, 1) to ( 219, 144) the accumulated minimum distance is 85.78232;

(5): The Euclidean distance between the time series T of the curve to be identified and the time series R of the normal representative curve is 0.28344;

(6): The total accumulated distance from the starting point (1, 1) to the ending point (219, 144) is 86.06576;

The method of calculating the distance between the power curve to be identified and the representative power curves of various faults is the same as above. The power curve to be identified and the disconnection curve of the starting circuit, the curve of the sudden stop of the turnout after starting, the curve of the foreign object caught in the turnout, and the action of the automatic switch The distances of the inflexible curve and the lockout current exceeding the standard curve are: 886.71484, 87.18578, 1.00232, 103.44763, 140.06902.

(5): The distances between the power curve to be identified and the normal representative power curve and various fault representative power curves are respectively reversed to 0.01162, 0.00112, 0.01147, 0.99769, 0.00967, 0.00714, indicating that the curve to be identified is disconnected from the normal representative curve and the starting circuit The similarities of the line curve, turnout sudden stop rotation curve after start, turnout with foreign matter curve, automatic switch inflexible curve and locking current exceeding the limit curve are 0.01162, 0.00112, 0.01147, 0.99769, 0.00967, 0.00714 respectively;

(6): Comparing the similarity between the curve to be identified and the normal representative curve and various fault representative curves, it can be obtained that the similarity between the curve to be identified and the curve with foreign matter in the turnout is the highest, and the type of the curve to be identified is the curve with foreign matter in the turnout.

Fig. 3 is a histogram of the similarity between a curve to be recognized and each representative curve according to Embodiment 1 of the present invention. As can be seen from Fig. 3, the curve to be recognized and the normal curve and five kinds of fault curves are calculated using the dynamic time warping algorithm The similarity between the curve to be identified and the fault with foreign matter in the turnout is the highest, so it is judged that the fault type of the curve to be identified is foreign matter in the turnout. After verification, the judgment result is correct.

Example 2

In this embodiment, a method for identifying a fault of a switch is also provided.

Fig. 4 is the flow chart of the fault identification method of a turnout according to Embodiment 2 of the present invention. As shown in Fig. 4, the method comprises the following steps: collecting each action current curve of the turnout in the microcomputer monitoring system; Classification can be divided into normal curves and fault curves. The fault curves can be further divided into: start circuit disconnection curve, turnout sudden stop rotation curve after start, turnout with foreign matter curve, switch machine stator rotor mixed line curve, automatic The action inflexibility curve of the switch, the start-up delay curve of the switch machine, the locking current exceeding the standard curve, and the switch action current are zigzag curves; the difference between the R, G, and B components of each pixel in the color switch action current curve image is taken The average value of the pixel is used as the gray value of the pixel, and the color switch action current curve image is converted into a gray image; a threshold is set so that the pixel with a gray value greater than the threshold takes a value of 1, and the pixel with a gray value smaller than the threshold The value of the point is 0, transform the grayscale image into a binary image; find out the target area surrounded by the coordinate axes, remove the free and isolated pixels in the target area, and smooth the edge of the curve to remove noise; Each moment corresponds to a value, and its pixel value is 0, and the case where there are multiple pixel points in a column is 0 is refined; through function transformation, the coordinates of each point on the curve are extracted; the coordinates of each point are scaled proportionally , so that the horizontal and vertical coordinates of each point are within the same range; for each type of turnout action current curve, arbitrarily select one such turnout action current curve as the representative curve of this type of current curve; use the Frescher distance algorithm to calculate The distance between the current curve and the normal representative curve and various fault representative current curves; the calculated distance is reversed to indicate the similarity between the curve to be identified and the normal representative curve and various fault representative curves; compare the curve to be identified with the normal representative curve and Various types of faults represent the similarity of the curves, and the type of curve with the highest similarity is the type of the curve to be identified.

The following will describe in conjunction with another specific optional embodiment.

(1): In the microcomputer monitoring system, the current curve of each action of the turnout is collected;

(2): Classify the acquired action current curves of turnouts, which can be divided into normal curves and fault curves. The fault curves can be further divided into: start circuit disconnection curves, turnout sudden stop rotation curves after start-up, and foreign objects caught in turnouts Curve, point machine stator rotor mixed line curve, automatic switch action inflexible curve, switch machine start-up delay curve, locking current exceeding the standard curve and switch operating current zigzag curve;

(3): Get the average value between the R, G, and B components of each pixel in the color switch action current curve image as the gray value of the pixel point, and convert the color switch action current curve image into a grayscale image;

(4): a threshold is set so that the pixel point whose grayscale value is greater than the threshold value is 1, and the pixel point whose grayscale value is less than the threshold value is 0, and the grayscale image is transformed into a binary image;

(5): Find the target area surrounded by the coordinate axes, remove the free and isolated pixels in the target area, and smooth the edge of the curve to remove noise;

(6): Make each moment correspond to a value, and its pixel point value is 0, and the situation that there are a plurality of pixel points that are 0 in a column is refined;

(7): Through function transformation, extract the coordinates of each point on the curve;

(8): Scale the coordinates of each point proportionally, so that the horizontal and vertical coordinates of each point are within the same range

(9): For each type of turnout action power curve, arbitrarily select one such turnout action current curve as the representative curve of this type of current curve;

(10): Using the Fresche-based distance algorithm, calculate the distance between the current curve to be identified and the normal representative curve and various fault representative current curves;

(1) Take a curve whose fault type is starting and stopping suddenly as the curve to be identified L ₁ can be expressed as P={P(1), P(2),...,P(97)}, P(1) =(0,0), P(1)=(0.0104,1), ..., P(1)=(1,0.0061), P(1) is the starting sampling point, P(97) is the ending sampling point ;

(2) Taking the normal curve as the representative curve L ₂ can be expressed as P'={P'(1), P'(2),...,P'(M)}, P'(1)=(0,0 ), P'(2)=(0.0015,0.1372), ..., P'(654)=(1,0.0020), P'(1) is the starting sampling point, P'(654) is the ending sampling point, ;

( ₃ ) Calculate the distance between each sampling point _on L1 and each sampling point on L2, and obtain the distance matrix D as follows:

1≤m≤654, 1≤n≤97

in the above formula Indicates the distance from the mth sampling point _on the curve L2 to the nth sampling point _on the curve L1.

(4) Find the maximum distance d _max =max(D)=1.4054 and the minimum distance d _min =min(D)=0 in the distance matrix D, initialize the target distance f=d _min =0, and set the cycle interval

(5) Set the elements less than or equal to f in the distance matrix D to 1, and set the elements greater than f to 0, so as to obtain the binary matrix D′ as follows:

1≤m≤654, 1≤n≤97

(6) Search for a path that satisfies the following conditions in the binary matrix D′: the starting point of the path is d′ ₁₁ , and the end point of the path is d′ _MN ; after the path passes through the point d′ _mn , the next passing point can only be is one of d′ _{m+1, n} , d′ _{m, n+1} , d′ _{m+1, n+1} ; all points in the path must have a value of 1.

(7) If in step 8f, do not find a path satisfying the condition, then set the target distance f=f+res=f+0.0141, then repeat steps 8e and 8f; if find a path satisfying the condition or target distance f in step 8f ＝d _m ^x away, then go to the next step.

Through (5), (6), (7) three-step cycle calculation, under the condition of f=0.5059, we find a path satisfying the condition at step 8f, and then jump out of this cycle and enter step (8).

(8) Through the above calculation, we can get the discrete Frechet distance Frechet=f=0.5059 between the curve to be identified and the normal curve, then the similarity between the curve to be identified and the normal curve

Through the same method, we can obtain the curve to be identified and the sudden stop rotation curve after the turnout is started, the stator and rotor confusion curve of the switch machine, the foreign body curve added to the switch, the disconnection curve of the starting circuit, the locking current exceeding the standard curve, and the turnout of the switch machine. The Frescher distances of the zigzag curve of the operating current, the start-up delay curve of the switch machine, and the inflexible curve of the automatic switch are 0.1266, 0.8278, 0.9777, 0.8406, 0.4908, 0.4907, 0.5030, and 0.8416, respectively. The curve to be identified and the switch suddenly stop turning after starting, the stator and rotor chaos curve of the switch machine, the foreign body curve added to the switch, the disconnection curve of the starting circuit, the lock-up current exceeding the limit curve, the zigzag curve of the switch operation current of the switch machine, and the turnout curve. The similarities of the start-up delay curve of the track machine and the inflexibility curve of the automatic switch are 7.8963, 1.2079, 1.02277, 1.1896, 2.0372, 2.0378, 1.9878, 1.1881, respectively. Therefore, the similarity between the curve to be identified and the curve that suddenly stops turning after the turnout starts is the highest, which is 7.8963, that is, the fault type of the curve to be identified is that the turnout suddenly stops turning after starting.

(11): Reverse the calculated distance to indicate the similarity between the curve to be identified and the normal representative curve and various fault representative curves;

(12): Compare the similarity between the curve to be identified and the normal representative curve and various fault representative curves, the type of curve with the highest similarity is the type of the curve to be identified.

Fig. 5 is an image of normal representative curves and 8 types of fault representative curves selected after preprocessing the acquired switch operating current curves in Embodiment 2 of the present invention.

Fig. 6 is a histogram of the similarity between a curve to be recognized and each representative curve according to Embodiment 2 of the present invention. As can be seen from Fig. 6, the curve to be recognized and the normal curve and 8 kinds of normal curves are calculated using the Fresche distance algorithm For the similarity of the fault curve, the similarity between the curve to be identified and the sudden stop of the turnout after starting is the highest, so it is judged that the fault type of the curve to be identified is the sudden stop of the turnout after starting. After verification, the judgment result is correct.

Those skilled in the art should understand that the embodiments of the present invention may be provided as methods, systems, or computer program products. Accordingly, the present invention can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.

The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It should be understood that each procedure and/or block in the flowchart and/or block diagram, and a combination of procedures and/or blocks in the flowchart and/or block diagram can be realized by computer program instructions. These computer program instructions may be provided to a general purpose computer, special purpose computer, embedded processor, or processor of other programmable data processing equipment to produce a machine such that the instructions executed by the processor of the computer or other programmable data processing equipment produce a An apparatus for realizing the functions specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.

These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing apparatus to operate in a specific manner, such that the instructions stored in the computer-readable memory produce an article of manufacture comprising instruction means, the instructions The device realizes the function specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.

These computer program instructions may also be loaded onto a computer or other programmable data processing equipment,

Causes a series of operational steps to be performed on a computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide a process for implementing one or more processes and/or block diagrams in the flowchart A step of a function specified in a box or boxes.

Apparently, the above-mentioned embodiments are only examples for clear description, rather than limiting the implementation. For those of ordinary skill in the art, on the basis of the above description, other changes or changes in different forms can also be made. It is not necessary and impossible to exhaustively list all the implementation manners here. And the obvious changes or changes derived therefrom are still within the scope of protection of the present invention.

Claims (6)

1. A kind of switch fault identification method is characterized in that, comprises the following steps: (1): collect each action curve of the turnout; (2): Divide the obtained action curves of each turnout into normal curves and fault curves; (3): For each type of normal curve and fault curve, select a curve with the most representative characteristics as the normal representative curve and fault representative curve of this type of curve; (4): Use the similarity algorithm to calculate the similarity between the curve to be identified and the normal representative curve 1, the similarity between the curve to be identified and the fault representative curve 2; the similarity algorithm is a dynamic time warping algorithm or an algorithm based on Fresche distance; The similarity algorithm is a dynamic time warping algorithm, specifically: (4a1): The curve to be identified can be expressed as T={T(1), T(2),...,T(n),...,T(N)}, n is the time series label of the time series, n= 1 is the starting point of the time series, n=N is the end point of the time series, and T(n) is the value of the time series; (4b1): The normal representative curve and the fault representative curve can be expressed as R={R(1), R(2),...,R(m),...,R(M)}, where m is the time sequence of the time series label, m=1 is the starting point of the time series, m=M is the end point of the time series, and R(m) is the value of the time series; (4c1): Mark each time series number n of the curve time series to be identified on the horizontal axis, mark each time series number m representing the time series of the curve on the vertical axis, and draw some vertical and horizontal lines through the integer coordinates of these time series numbers to form a Network, all grid points are (1, 1), ..., (n, m), ..., (N, M), search for the optimal path from (1, 1) to (N, M); (4d1): After the path passes through (n, m), the next grid point to pass can only be (n, m+1), (n+1, m), (n+1, m+1), select ( The minimum distance from n, m) to the next grid point is the optimal path, and the accumulated minimum distance from (1, 1) to (N, M) is calculated; (4e1): Calculate the Euclidean distance between the time series T of the curve to be identified and the time series R of the representative curve; (4f1): The total accumulated distance from the start point (1, 1) to the end point (N, M) is the cumulative minimum distance from the start point (1, 1) to the end point (N, M), the time series T of the curve to be identified and the representative curve The sum of the Euclidean distances between the time series R; (4g1): Negating the total accumulated distance, indicating the similarity between the curve to be identified and the normal representative curve or the curve to be identified and the fault representative curve; Or: the similarity algorithm is an algorithm based on Fresche distance, specifically: (4a2): The curve L ₁ to be identified can be expressed as P={P(1), P(2),..., P(n),..., P(N)}, P(n)=(x _n ,y _n ), n is the serial number of the sampling point on the curve L ₁ , n=1 is the starting sampling point, n=N is the end sampling point, x _n is the abscissa of the nth sampling point, x _n is the The vertical coordinates of n sampling points; (4b2): Normal representative curve or fault representative curve L2 can be expressed as P'={P'(1), P'( ₂ ),...,P'(m),...,P'(M)} , P′(m)=(x′ _m , y′ _m ), m is the serial number of the sampling point on the curve L ₂ , m=1 is the starting sampling point, m=M is the ending sampling point, x’ _m is The abscissa of the mth sampling point, y' _m is the ordinate of the mth sampling point; (4c2): Calculate the distance between each sampling point _on L2 and each sampling point _on L2, and obtain the distance matrix D as follows: 1≤m≤M, 1≤n≤N in the above formula Indicates the distance from the mth sampling point _on the curve L2 to the nth sampling point on the curve L2 _; (4d2): Select the maximum distance d _max =max(D) and the minimum distance d _min =min(D) in the distance matrix D, initialize the target distance f=d _min , and set the cycle interval (4e2): Set the elements less than or equal to f in the distance matrix D to 1, and set the elements greater than f to 0, so as to obtain the binary matrix D′ as follows: 1≤m≤M, 1≤n≤N (4f2): Search for a path that satisfies the following conditions in the binary matrix D′: the starting point of the path is d' ₁₁ , the end point of the path is d' _MN , and after the path passes through the point d' _mn , the next passing point is only It can be one of d' _m+1,n , d' _m,n+1 , d' _m+1,n+1 , and the value of all points in the path must be 1; (4g2): If no path satisfying the condition is found in step (4f2), set the target distance f=f+res, then repeat steps (4e2) and (4f2), if a path satisfying the condition is found in step (4f2) Path or target distance f=d _max , enter the next step; (4h2): Frechet distance Frechet=f between the curve to be identified and the normal representative curve or fault representative curve, Indicates the similarity between the curve to be identified and the normal representative curve or fault representative curve; (5): Comparing the similarity obtained in the calculation step (4), if the similarity 1 is greater than the similarity 2, then the curve is a normal curve, and if the similarity 1 is smaller than the similarity 2, then the curve is a failure curve. 2. the turnout fault identification method according to claim 1, is characterized in that, described in the step (1) collects the turnout action curve data or the image of turnout action curve that generates in the microcomputer monitoring system every time, or is a paper document Turnout action curve data or images in . 3. The fault identification method of a turnout according to claim 1, wherein the step (1) of collecting each action curve of the turnout is the action current curve data or image of the turnout; or the action power curve data or image of the turnout. 4. The switch failure identification method according to claim 1, characterized in that, the acquired switch action curves in step (2) are divided into normal curves and fault curves, and the fault curves are specifically divided into: Line curve, sudden stop rotation curve after turnout starts, foreign body curve in turnout, mixed stator and rotor curve of switch machine, inflexible action curve of automatic switch, start-up delay curve of switch machine, lock current exceeding standard curve and turnout The operating current is a zigzag curve; for each type of fault curve, select a curve with the most representative characteristics as the representative curve of this type of fault curve; and calculate the similarity between the curve to be identified and the representative curve of each type of fault; The type of curve with the highest similarity is the fault type of the curve to be identified. 5. the switch fault identification method according to claim 1, is characterized in that, before step (3), the obtained switch action curve is preprocessed, comprising the following steps: (1): Get the mean value between the R, G, and B components of each pixel in the colored turnout action curve image as the gray value of the pixel, and convert the colored turnout action curve image into a grayscale image; (2): A threshold is set so that the pixel point whose gray value is greater than the threshold value is 1, and the pixel point whose gray value is smaller than the threshold value is 0, and the gray image is converted into a binary image; (3): Find the target area enclosed by the coordinate axes, remove the free and isolated pixels in the target area, and smooth the edge of the curve to remove noise; (4): Each moment corresponds to a value, and its pixel value is 0, and the situation that there are multiple pixel points in a column is 0 is refined; (5): Through function transformation, extract the coordinates of each point on the curve; (6): Scale the coordinates of each point proportionally so that the horizontal and vertical coordinates of each point are within the same range. 6. The switch fault identification method according to claim 1, characterized in that, in step (3), for normal curves, selecting any one curve is a normal representative curve; One curve is representative of such failure curves.