A temperature sensor fault diagnosis method, storage medium and traction system

Abstract

The present invention discloses a temperature sensor fault diagnosis method, storage medium, and traction system. The method collects multiple sets of temperature measurement data from multiple temperature sensors used to detect the temperature of a target device. For each set of temperature measurement data, the temperature measurement data collected by the target temperature sensor is reconstructed using the temperature measurement data collected by the remaining temperature sensors except the target temperature sensor to be diagnosed, thereby obtaining temperature reconstruction data corresponding to the target temperature sensor to be diagnosed. Based on the temperature reconstruction data, temperature estimation data is determined using a principal component analysis method. Finally, based on the temperature estimation data, whether the target temperature sensor to be diagnosed is faulty is determined. This method reconstructs the temperature measurement data of the target temperature sensor based on the correlated timing characteristics of the multiple temperature sensors and further determines whether the target temperature sensor to be diagnosed is faulty based on the reconstructed data. This improves the effectiveness and reliability of fault diagnosis and helps ensure the safety of the target device.

Description

Technical Field

[0001] The present invention relates to the technical field of fault diagnosis, and in particular to a temperature sensor fault diagnosis method, a storage medium, and a traction system. Background Art

[0002] Temperature sensors are widely used in train traction systems, for example, to measure the cooling oil temperature of traction transformers, the cooling water temperature of traction converters, and the stator and bearing temperatures of traction motors. Temperature is a key indicator of the operating environment and status of locomotive transmission equipment. Excessively high temperatures require protective actions such as reduced train power and tripping vacuum circuit breakers to ensure safe equipment operation. Due to the harsh operating environment of traction systems, temperature sensor failures are common, causing the measured value to deviate from the true value. If the measured value is higher than the true value, it can cause system malfunctions, impacting normal locomotive operation; if the measured value is lower than the true value, it can cause system failures, causing damage to locomotive equipment. Therefore, a failure can cause system malfunctions, severely impacting normal train operation. Therefore, an effective real-time diagnostic method for train temperature sensors is urgently needed to ensure safe and stable train operation. Summary of the Invention

[0003] The technical problem to be solved by the present invention is: how to improve the effectiveness of temperature sensor fault diagnosis.

[0004] In order to solve the above technical problems, the present invention provides a temperature sensor fault diagnosis method, a storage medium and a traction system.

[0005] A first aspect of the present invention provides a temperature sensor fault diagnosis method, comprising:

[0006] Collecting multiple groups of temperature measurement data from multiple temperature sensors for detecting the temperature of the target device, each group of temperature measurement data corresponds to the temperature measurement data of the multiple temperature sensors at a first moment;

[0007] For each set of the temperature measurement data, reconstructing the temperature measurement data collected by the temperature sensor to be diagnosed using the temperature measurement data collected by the remaining temperature sensors among the multiple temperature sensors except the temperature sensor to be diagnosed, so as to obtain temperature reconstructed data corresponding to the temperature sensor to be diagnosed;

[0008] Determining temperature estimation data based on the temperature reconstruction data using a principal component analysis method;

[0009] According to the temperature estimation data, it is determined whether the temperature sensor to be diagnosed is faulty; wherein the temperature reconstruction data is obtained by the following expression:

[0010]

[0011] zi represents the temperature reconstruction data of the i-th temperature sensor, i = 1, 2, ..., m; Represents the matrix transpose of the temperature measurement data collected by 0 to i-1 temperature sensors; c represents the matrix transpose of the temperature measurement data collected by temperature sensors i+1 to m; i represents the rank-deficient element corresponding to the i-th temperature sensor; c ii Represents the rank deficiency element corresponding to the i-th row and i-th column in the rank deficiency matrix.

[0012] In some embodiments, for each set of temperature measurement data, before reconstructing the temperature measurement data collected by the temperature sensor to be diagnosed using the temperature measurement data collected by the remaining temperature sensors among the multiple temperature sensors except the temperature sensor to be diagnosed, the method further includes:

[0013] Acquire historical temperature sample data when the multiple temperature sensors are in normal operating state;

[0014] Based on the historical temperature sample data, a rank-deficient matrix corresponding to the plurality of temperature sensors in the principal component analysis method is determined.

[0015] In some embodiments, for each set of temperature measurement data, before reconstructing the temperature measurement data collected by the temperature sensor to be diagnosed using the temperature measurement data collected by the remaining temperature sensors among the multiple temperature sensors except the temperature sensor to be diagnosed, the method further includes:

[0016] The plurality of sets of temperature measurement data are expanded according to a preset time lag factor to obtain an expansion matrix of temperature measurement data from the first moment to the second moment, wherein a difference between the first moment and the second moment is equal to the preset time lag factor.

[0017] In some embodiments, determining temperature estimation data based on the temperature reconstruction data using a principal component analysis method includes:

[0018] The temperature estimation data is determined based on the product of the temperature reconstruction data and the rank-deficient element corresponding to the temperature sensor to be diagnosed.

[0019] In some embodiments, determining whether the temperature sensor to be diagnosed is faulty based on the temperature estimation data includes:

[0020] Determine whether the temperature estimation data exceeds a preset temperature range. When the temperature estimation data exceeds the preset temperature range, determine that the temperature sensor to be diagnosed is faulty; when the temperature estimation data is within the preset temperature range, determine that the temperature sensor to be diagnosed is normal.

[0021] In some embodiments, determining whether the temperature sensor to be diagnosed is faulty based on the temperature estimation data includes:

[0022] Calculating a residual between the temperature measurement data corresponding to the temperature sensor to be diagnosed and the temperature estimation data;

[0023] The test statistic corresponding to the residual is determined by the following expression:

[0024]

[0025] represents the detection statistic corresponding to the i-th temperature sensor at time k, j = 1, 2, ..., n; q represents the length of the data window selected when obtaining the residual; p represents time, p = k-q+1, k-q+2, ..., k; is the variance of the residual calculated based on the historical temperature sample data of the i-th temperature sensor in a normal state;

[0026] Based on the detection statistic and the detection threshold, determining whether the temperature sensor to be diagnosed is faulty;

[0027] When the detection statistic is greater than the detection threshold, it is determined that the temperature sensor to be diagnosed is faulty; when the detection statistic is less than or equal to the detection threshold, it is determined that the temperature sensor to be diagnosed is normal.

[0028] In some embodiments, before determining whether the temperature sensor to be diagnosed is faulty based on the detection statistic and the detection threshold, the method further includes:

[0029] The corresponding detection threshold is determined according to the preset signal level and the data window length.

[0030] A second aspect of the present invention provides a storage medium storing a computer program, wherein the computer program, when executed by a processor, implements any one of the temperature sensor fault diagnosis methods described above.

[0031] According to a third aspect of the present invention, a traction system is provided, comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the temperature sensor fault diagnosis method as described above is implemented.

[0032] In some embodiments, the traction system further includes: a traction transformer, a traction converter, an oil temperature sensor for detecting the oil temperature in the traction transformer, and a water temperature sensor for detecting the cooling water in the traction converter.

[0033] Compared with the prior art, one or more embodiments of the above solutions may have the following advantages or beneficial effects:

[0034] The temperature sensor fault diagnosis method provided by the present invention collects multiple sets of temperature measurement data from multiple temperature sensors used to detect the temperature of a target device. For each set of temperature measurement data, the temperature measurement data collected by the temperature sensor to be diagnosed is reconstructed using the temperature measurement data collected by the remaining temperature sensors except the temperature sensor to be diagnosed, thereby obtaining temperature reconstruction data corresponding to the temperature sensor to be diagnosed. Based on the temperature reconstruction data, temperature estimation data is determined using a principal component analysis method. Finally, based on the temperature estimation data, whether the temperature sensor to be diagnosed is faulty is determined. This method reconstructs the temperature measurement data of the temperature sensor to be diagnosed based on the correlated timing characteristics of multiple temperature sensors, and further determines whether the temperature sensor to be diagnosed is faulty based on the obtained temperature reconstruction data. This improves the reliability of the temperature estimation data and the effectiveness and reliability of fault diagnosis, thereby facilitating the safety of the target device. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The scope of the present disclosure may be better understood by reading the following detailed description of exemplary embodiments in conjunction with the accompanying drawings, which include:

[0036] Figure 1 A schematic flow chart of a temperature sensor fault diagnosis method according to a first embodiment of the present invention is shown;

[0037] Figure 2 A schematic diagram of a temperature sensor fault diagnosis method according to a second embodiment of the present invention is shown;

[0038] Figure 3 A schematic structural diagram of a traction system provided by an embodiment of the present invention is shown;

[0039] Figure 4 A schematic diagram of the cooling structure of the traction system provided by an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0040] In order to make the objectives, technical solutions and advantages of the present invention clearer, the implementation method of the present invention will be described in detail below with reference to the accompanying drawings and embodiments, so that the implementation process of how the present invention applies technical means to solve technical problems and achieve technical effects can be fully understood and implemented accordingly.

[0041] Temperature sensors are widely used in train traction systems, for example, to measure the cooling oil temperature of traction transformers, the cooling water temperature of traction converters, and the stator and bearing temperatures of traction motors. Temperature is a key indicator of the operating environment and status of locomotive transmission equipment. Excessively high temperatures require protective actions such as reduced train power and tripping vacuum circuit breakers to ensure safe equipment operation. Due to the harsh operating environment of traction systems, sensor failures are common, causing sensor measurements to deviate from the true value. If the measured value is higher than the true value, it can cause system malfunctions, impacting normal locomotive operation; if the measured value is lower than the true value, it can cause system failures, potentially damaging locomotive equipment. Therefore, a fault can cause system malfunctions, severely impacting normal train operation. Therefore, an effective real-time diagnostic method for train temperature sensors is urgently needed to ensure safe and stable train operation.

[0042] In view of this, the present invention provides a temperature sensor fault diagnosis method. The method collects multiple sets of temperature measurement data from multiple temperature sensors used to detect the temperature of a target device. For each set of temperature measurement data, the temperature measurement data collected by the target temperature sensor is reconstructed using the temperature measurement data collected by the remaining temperature sensors except the target temperature sensor to be diagnosed, thereby obtaining temperature reconstruction data corresponding to the target temperature sensor to be diagnosed. Based on the temperature reconstruction data, temperature estimation data is determined using a principal component analysis method. Finally, based on the temperature estimation data, whether the target temperature sensor to be diagnosed is faulty is determined. This method reconstructs the temperature measurement data of the target temperature sensor to be diagnosed based on the correlated timing characteristics of the multiple temperature sensors, and further determines whether the target temperature sensor to be diagnosed is faulty based on the obtained temperature reconstruction data. This method improves the reliability of the temperature estimation data and the effectiveness and reliability of fault diagnosis, thereby facilitating the safety of the target device.

[0043] Example 1

[0044] See also Figure 1 As shown, Figure 1 A schematic flow chart of a temperature sensor fault diagnosis method provided in the first embodiment of the present invention is shown, which may include:

[0045] Step S101: collecting multiple sets of temperature measurement data from multiple temperature sensors for detecting the temperature of a target device, where each set of temperature measurement data corresponds to the temperature measurement data of the multiple temperature sensors at a first moment;

[0046] Step S102: For each set of temperature measurement data, reconstruct the temperature measurement data collected by the temperature sensor to be diagnosed using the temperature measurement data collected by the remaining temperature sensors among the multiple temperature sensors except the temperature sensor to be diagnosed, so as to obtain temperature reconstructed data corresponding to the temperature sensor to be diagnosed; wherein the temperature reconstructed data is obtained by the following expression:

[0047]

[0048] z i represents the temperature reconstruction data of the i-th temperature sensor, i = 1, 2, ..., m; Represents the matrix transpose of the temperature measurement data collected by 0 to i-1 temperature sensors; c represents the matrix transpose of the temperature measurement data collected by temperature sensors i+1 to m; i represents the rank-deficient element corresponding to the i-th temperature sensor; c ii Represents the rank deficiency element corresponding to the i-th row and i-th column in the rank deficiency matrix;

[0049] Step S103: Based on the temperature reconstruction data, temperature estimation data is determined using a principal component analysis method;

[0050] Step S104: Determine whether the temperature sensor to be diagnosed is faulty based on the temperature estimation data.

[0051] In some embodiments, the first moments corresponding to different groups of temperature measurement data may be different; in other embodiments, the first moments corresponding to different groups of temperature measurement data may also be the same.

[0052] In some embodiments, step S101 may specifically include collecting n sets of temperature measurement data from a plurality of temperature sensors used to detect the temperature of the target device. As an example, temperature measurement data from m temperature sensors may be collected, thereby obtaining a temperature measurement data matrix X formed by the n sets of temperature measurement data from the m temperature sensors:

[0053]

[0054] in, Represents the temperature measurement data measured by the mth temperature sensor in the nth set of temperature measurement data.

[0055] In some embodiments, step S102 may specifically include: in response to an instruction to perform fault diagnosis on the i-th temperature sensor among the multiple temperature sensors, the i-th temperature sensor is used as the temperature sensor to be diagnosed, and the following expression (1) can be used for each set of temperature measurement data to determine the temperature reconstruction data of the temperature sensor to be diagnosed:

[0056]

[0057] In some embodiments, before step S102, the following steps may also be included:

[0058] Acquire historical temperature sample data of multiple temperature sensors in normal operating state;

[0059] Based on historical temperature sample data, the rank-deficient matrix corresponding to multiple temperature sensors in the principal component analysis method is determined.

[0060] As an example, n groups of historical temperature sample data from m temperature sensors can be obtained to obtain a historical temperature sample data matrix X' with n rows and m columns:

[0061]

[0062] The covariance matrix of the historical temperature sample data matrix X' is calculated based on the following expression:

[0063]

[0064] Among them, ∑ is a real symmetric matrix,

[0065] Take the number of principal elements as l, diagonalize ∑, and arrange the eigenvalues from large to small, that is, λ1≥λ2≥…≥λ m ≥0, we get:

[0066]

[0067] Among them, P is the principal component load matrix composed of the first l eigenvectors, P e is the matrix composed of the remaining ml eigenvectors.

[0068]

[0069] Among them, PP T The matrix is represented by C, which is the rank-deficient matrix. From the above, the rank-deficient matrix can be determined:

[0070]

[0071] C is a matrix with m rows and m columns, c mm Represents the rank-deficient element in the mth row and mth column of the rank-deficient matrix.

[0072] In the embodiment of the present invention, step S102 may be specifically as follows: obtaining the rank defect element c corresponding to the i-th temperature sensor, i.e., the temperature sensor to be diagnosed, from the predetermined rank defect matrix C. i And the rank deficiency element c corresponding to the i-th row and i-th column in the rank deficiency matrix ii; Obtain the temperature reconstruction data to be diagnosed according to the following expression:

[0073]

[0074] Formula (6) is constructed based on the correlation timing characteristics of multiple temperature sensors and the sum of the weight values of the temperature measurement data corresponding to the multiple temperature sensors is 1, so as to avoid the influence of the temperature estimation data when the temperature sensor failure to be diagnosed is damaged, and improve the validity of the temperature estimation data.

[0075] As an example, a set of temperature measurement data X = [x1, x2, x3, x4, x5] is collected, and the fourth temperature sensor is set as the temperature sensor to be diagnosed. Substituting into formula (6), we can obtain the temperature reconstruction data of the temperature sensor to be diagnosed.

[0076] In an embodiment of the present invention, step S103 may specifically be determining temperature estimation data based on the product of the temperature reconstruction data and the rank-deficient element corresponding to the temperature sensor to be diagnosed.

[0077] Among them, the principal component analysis method can be used to obtain the temperature estimation data and the temperature measurement data satisfying the above expression (5). After replacing the temperature measurement data with the temperature reconstruction data, the expression of the temperature estimation data can be obtained:

[0078]

[0079] Among them, Z (i) =[x1,x2,……,z i ,……x m ], Represents the temperature estimation data corresponding to the i-th temperature sensor, i.e., the temperature sensor to be diagnosed.

[0080] In some embodiments, step S104 may specifically include:

[0081] Determine whether the temperature estimation data exceeds a preset temperature range. When the temperature estimation data exceeds the preset temperature range, determine that the temperature sensor to be diagnosed is faulty; when the temperature estimation data is within the preset temperature range, determine that the temperature sensor to be diagnosed is normal.

[0082] In some other embodiments, step S104 may specifically include:

[0083] Step 1: Calculate the residual between the temperature measurement data and the temperature estimation data corresponding to the temperature sensor to be diagnosed;

[0084] Step 2: Determine the test statistic corresponding to the residual using the following expression:

[0085]

[0086] represents the detection statistic corresponding to the i-th temperature sensor at time k, j = 1, 2, ..., n; q represents the length of the data window selected when obtaining the residual; p represents time, p = k-q+1, k-q+2, ..., k; is the variance of the residual calculated based on the historical temperature sample data of the i-th temperature sensor in a normal state;

[0087] Step 3: Based on the detection statistics and the detection threshold, determine whether the temperature sensor to be diagnosed is faulty;

[0088] Step 4: When the detection statistic is greater than the detection threshold, it is determined that the temperature sensor to be diagnosed is faulty;

[0089] Step 5: When the detection statistic is less than or equal to the detection threshold, it is determined that the temperature sensor to be diagnosed is normal.

[0090] In the embodiment of the present invention, the residual of the temperature measurement data and the temperature estimation data corresponding to the temperature sensor to be diagnosed in step 1 can be expressed as:

[0091]

[0092] in, represents the residual of the jth temperature measurement data when the i-th temperature sensor, i.e., the temperature sensor to be diagnosed, is performing fault detection; represents the jth temperature measurement data of the temperature sensor to be diagnosed, express The j-th temperature estimation data of the temperature sensor to be diagnosed.

[0093] In the embodiment of the present invention, the data window length is selected as q, that is, q The data samples are taken as a group, and the overall distribution obeys the chi-square distribution with q degrees of freedom. Combined with the residuals determined by formula (8), the test statistic can be further determined by the following formula (9):

[0094]

[0095] In some embodiments, steps 3 to 5 may specifically be to construct a fault diagnosis logic of the temperature sensor based on formula (10):

[0096]

[0097] in, represents χ with q degrees of freedom 2 distribution; α represents the preset confidence level, which is usually understood as the probability of allowing false detection. This means that the temperature sensor to be diagnosed is faulty; This means that the temperature sensor to be diagnosed is operating normally.

[0098] In some embodiments, before step three, a corresponding detection threshold may be determined according to a preset signal level and a data window length.

[0099] Among them, the preset signal level α can be set according to actual conditions and needs, and the data window length can also be set according to needs. Based on the preset signal level, data window length and the correspondence table between the preset signal level, data window length and detection threshold, the corresponding detection threshold is determined.

[0100] The above is a temperature sensor fault diagnosis method provided by the present invention. It collects multiple sets of temperature measurement data from multiple temperature sensors used to detect the temperature of a target device; for each set of temperature measurement data, reconstructs the temperature measurement data collected by the temperature sensor to be diagnosed using the temperature measurement data collected by the remaining temperature sensors except the temperature sensor to be diagnosed, so as to obtain temperature reconstruction data corresponding to the temperature sensor to be diagnosed; determines temperature estimation data based on the temperature reconstruction data using the principal component analysis method; and finally, determines whether the temperature sensor to be diagnosed is faulty based on the temperature estimation data. This method reconstructs the temperature measurement data of the temperature sensor to be diagnosed based on the correlated timing characteristics of multiple temperature sensors, and further determines whether the temperature sensor to be diagnosed is faulty based on the obtained temperature reconstruction data. This improves the reliability of the temperature estimation data and the effectiveness and reliability of the fault diagnosis, which is beneficial to ensuring the safety of the target device.

[0101] In an embodiment of the present invention, in order to improve the effectiveness of temperature sensor diagnosis, fault diagnosis of the temperature sensor may also be performed based on a dynamic principal component analysis method. For details, please refer to the description in the following embodiment 2.

[0102] Example 2

[0103] See also Figure 2 As shown, Figure 2 A schematic flow chart of a temperature sensor fault diagnosis method provided in the second embodiment of the present invention is shown, which may include:

[0104] Step S201: collecting multiple groups of temperature measurement data from multiple temperature sensors for detecting the temperature of a target device, where each group of temperature measurement data corresponds to the temperature measurement data of the multiple temperature sensors at a first moment;

[0105] Step S202: Expanding multiple sets of temperature measurement data according to a preset hysteresis factor to obtain an expanded matrix of temperature measurement data from a first moment to a second moment, wherein the difference between the first moment and the second moment is equal to the preset hysteresis factor;

[0106] Step S203: For each set of temperature measurement data, reconstruct the temperature measurement data collected by the temperature sensor to be diagnosed using the temperature measurement data collected by the remaining temperature sensors among the multiple temperature sensors except the temperature sensor to be diagnosed, so as to obtain temperature reconstructed data corresponding to the temperature sensor to be diagnosed; wherein the temperature reconstructed data is obtained by the following expression:

[0107]

[0108] z i represents the temperature reconstruction data of the i-th temperature sensor, i = 1, 2, ..., m; Represents the matrix transpose of the temperature measurement data collected by 0 to i-1 temperature sensors; c represents the matrix transpose of the temperature measurement data collected by temperature sensors i+1 to m; i represents the rank-deficient element corresponding to the i-th temperature sensor; c ii Represents the rank deficiency element corresponding to the i-th row and i-th column in the rank deficiency matrix;

[0109] Step S204: Based on the temperature reconstruction data, determine the temperature estimation data using the principal component analysis method;

[0110] Step S205: Determine whether the temperature sensor to be diagnosed is faulty based on the temperature estimation data.

[0111] It should be noted that step S201 can be implemented in the same manner as step S101 in Example 1, and steps S203 to S205 can be implemented in the same manner as steps S102 to S104 in Example 1. For the sake of brevity, they will not be repeated here. For details, please refer to the description in Example 1 above.

[0112] In the embodiment of the present invention, step S202 may specifically be:

[0113] Based on the preset time lag factor h, the multiple sets of temperature measurement data from the multiple temperature sensors acquired in step S201 can be expanded to include h+1 sets of temperature measurement data collected by the multiple temperature sensors from the first moment to the second moment before the first moment. As an example, based on the temperature measurement data matrix X shown in the first embodiment above, the expanded temperature measurement data expansion matrix can be obtained:

[0114]

[0115] Where V=[x1,x2,x3,……,x m ], V(t+hn) represents a set of temperature measurement data corresponding to the nhth moments before time t.

[0116] Formula (11) can reflect the dynamic change process of temperature measurement data. Data reconstruction is performed based on each set of temperature measurement data in the temperature measurement data expansion matrix, and temperature estimation data can be obtained based on the principal component analysis method, which can effectively improve the effectiveness of temperature sensor fault diagnosis.

[0117] It should be noted that, in this embodiment, after determining the temperature estimation data based on the temperature measurement data expansion matrix, the principal component analysis method can also be predetermined. The rank deficiency matrix in , where the rank deficiency matrix C can be determined by constructing a historical temperature sample data expansion matrix based on the historical temperature sample data of multiple temperature sensors in normal operating conditions and a preset time lag factor h, and then using equations (3) to (5) to determine the corresponding rank deficiency matrix.

[0118] The above is a temperature sensor fault diagnosis method provided in the second embodiment of the present invention. It collects multiple sets of temperature measurement data from multiple temperature sensors used to detect the temperature of the target device; expands the multiple sets of temperature measurement data according to a preset time lag factor; for each set of temperature measurement data, reconstructs the temperature measurement data collected by the temperature sensor to be diagnosed using the temperature measurement data collected by the remaining temperature sensors except the temperature sensor to be diagnosed, so as to obtain temperature reconstruction data corresponding to the temperature sensor to be diagnosed; determines temperature estimation data based on the temperature reconstruction data using the principal component analysis method; and finally, determines whether the temperature sensor to be diagnosed is faulty based on the temperature estimation data. While achieving the same beneficial effects as the above embodiment, this method can reflect the dynamic change process of the temperature measurement data by performing matrix expansion, and further obtains temperature estimation data based on the dynamic principal component analysis method, effectively improving the effectiveness of temperature sensor fault diagnosis.

[0119] Another aspect of the embodiments of the present invention further provides a traction system, for details please refer to the description in the following embodiment three.

[0120] Example 3

[0121] See also Figure 3 As shown, Figure 3 A schematic structural diagram of a traction system provided by an embodiment of the present invention is shown, which may include:

[0122] The memory 31 and the processor 32 , wherein the memory 31 stores a computer program, and when the computer program is executed by the processor 32 , the temperature sensor fault diagnosis method as described in the first or second embodiment above is implemented.

[0123] In some embodiments, the traction system may further include: a traction transformer 33, a traction converter 34, an oil temperature sensor 35 for detecting the oil temperature in the traction transformer, and a water temperature sensor 36 for detecting the cooling water in the traction converter.

[0124] The traction transformer 33 and traction converter 34 in the traction system can be used as target devices. The temperature sensor diagnosis method provided by the present invention can be used to perform fault diagnosis on the oil temperature sensor 35 and / or water temperature sensor 36 to ensure the safety of the traction system operation.

[0125] In some embodiments, see Figure 4 As shown, Figure 4 A schematic diagram of the traction system cooling structure provided by an embodiment of the present invention is shown. The composite radiator is a heat exchange device in the composite cooling tower oil-water cooling system. Two traction converters 34 can be provided, corresponding to two composite cooling towers. The cooling water from the traction converter 34 enters the upper layer of the composite radiator, while the oil from the traction transformer 33 enters the lower layer of the composite radiator. Heat from the traction converter 34 and traction transformer 33 is exchanged with air in the composite radiator, cooling the water and oil to the required temperature. The cooled water and oil then enter the traction converter 34 and traction transformer 33, respectively, to cool them. The composite radiator of the composite cooling tower is connected to the traction converter 34 and traction transformer 33 via pipelines. The composite radiator consists of two separate fluid branches. Through the operation of a fan, oil pump, and water pump, heat is exchanged between the oil, water, and air in the composite radiator.

[0126] The cooling fan in the composite cooling tower is provided with variable frequency and variable voltage power by the converter. The train control system automatically adjusts its operating voltage and frequency to change the speed of the cooling fan according to the oil temperature of the traction transformer 33 branch and the water temperature of the main converter branch of the traction converter 34, so that the air volume is suitable for the cooling needs of the traction converter 34 and the traction transformer 33 at that time. The operating speed of the water pump and the oil pump can remain constant.

[0127] Each traction converter 34 can be equipped with two water temperature sensors 36, installed at the inlet and outlet of the water pipe of the traction converter 34, respectively, to measure the inlet and outlet water temperatures within the traction converter 34. Two oil temperature sensors 35 can be provided, installed at the top of the traction transformer 33, to measure the oil temperature within the traction transformer 33. The temperature sensor fault diagnosis method provided by the embodiment of the present invention can realize real-time diagnosis of each temperature sensor in the traction system.

[0128] In some embodiments, a set of temperature measurement data of 6 temperature sensors in the traction system can be obtained, and then X=[T IW1 , TOW1 , T IW2 , T OW2 , T MT1 , T MT2 ], where T IW1 、T IW2 represents the temperature of the water inlet of the traction converter 34 detected by the different water temperature sensors 36, T OW1 、T OW2 The temperature of the water outlet of the traction converter 34 detected by the different water temperature sensors 36; T MT1 、T MT2 represents the internal oil temperature of the traction transformer 33 detected by different oil temperature sensors 35. Based on the acquired temperature measurement data, the temperature sensor fault diagnosis method provided by the embodiment of the present invention is used to perform fault diagnosis, which is conducive to improving the effectiveness of temperature sensor fault diagnosis and ensuring the safe operation of the traction system.

[0129] In another aspect of the embodiments of the present invention, a storage medium is provided, in which a computer program is stored. When the computer program is executed by a processor, the temperature sensor fault diagnosis method described in the first or second embodiment can be implemented.

[0130] The processes, functions, methods, and / or software described above may be recorded, stored, or fixed in one or more computer-readable storage media, which include program instructions that are implemented by a computer so that a processor executes the program instructions. The storage media may also include program instructions, data files, data structures, etc., alone or in combination. The storage media or program instructions may be specifically designed and understood by those skilled in the art of computer software, or the storage media or instructions may be generally known and available to those skilled in the art of computer software. Examples of computer-readable media include: magnetic media, such as hard disks, floppy disks, and magnetic tapes; optical media, such as CDROMs and DVDs; magneto-optical media, such as optical disks; and hardware devices specifically configured to store and execute program instructions, such as read-only memory (ROM), random access memory (RAM), flash memory, etc. Examples of program instructions include machine code (e.g., code generated by a compiler) and files containing higher-level code that can be executed by a computer using an interpreter. The described hardware devices may be configured to function as one or more software modules to perform the operations and methods described above, and vice versa. In addition, the computer-readable storage medium may be distributed among networked computer systems, so that computer-readable codes or program instructions may be stored and executed in a decentralized manner.

[0131] Although the embodiments disclosed herein are as described above, the contents described herein are merely embodiments for facilitating understanding of the present invention and are not intended to limit the present invention. Any person skilled in the art may make any modifications and variations in the form and details of the embodiments without departing from the spirit and scope of the present invention. However, the scope of protection of the present invention shall remain subject to the scope defined by the appended claims.

Claims

1. A temperature sensor fault diagnosis method, characterized in that: include: Collecting multiple groups of temperature measurement data from multiple temperature sensors for detecting the temperature of the target device, each group of temperature measurement data corresponds to the temperature measurement data of the multiple temperature sensors at a first moment; For each set of the temperature measurement data, reconstructing the temperature measurement data collected by the temperature sensor to be diagnosed using the temperature measurement data collected by the remaining temperature sensors among the multiple temperature sensors except the temperature sensor to be diagnosed, so as to obtain temperature reconstructed data corresponding to the temperature sensor to be diagnosed; Determining temperature estimation data based on the temperature reconstruction data using a principal component analysis method; According to the temperature estimation data, it is determined whether the temperature sensor to be diagnosed is faulty; wherein the temperature reconstruction data is obtained by the following expression: zi represents the temperature reconstruction data of the i-th temperature sensor, i = 1, 2, ..., m; Represents the matrix transpose of the temperature measurement data collected by 0~i-1 temperature sensors; The matrix transpose representing the temperature measurement data collected by temperature sensors i+1 to m; Represents the rank-deficient element corresponding to the i-th temperature sensor; Represents the rank deficiency element corresponding to the i-th row and i-th column in the rank deficiency matrix.

2. The method according to claim 1, characterized in that For each set of temperature measurement data, before reconstructing the temperature measurement data collected by the temperature sensor to be diagnosed using the temperature measurement data collected by the remaining temperature sensors among the multiple temperature sensors except the temperature sensor to be diagnosed, the method further includes: Acquire historical temperature sample data when the multiple temperature sensors are in normal operating state; Based on the historical temperature sample data, a rank-deficient matrix corresponding to the plurality of temperature sensors in the principal component analysis method is determined.

3. The method according to claim 1, characterized in that For each set of temperature measurement data, before reconstructing the temperature measurement data collected by the temperature sensor to be diagnosed using the temperature measurement data collected by the remaining temperature sensors among the multiple temperature sensors except the temperature sensor to be diagnosed, the method further includes: The plurality of sets of temperature measurement data are expanded according to a preset time lag factor to obtain an expansion matrix of temperature measurement data from the first moment to the second moment, wherein a difference between the first moment and the second moment is equal to the preset time lag factor.

4. The method according to claim 1, wherein Based on the temperature reconstruction data, temperature estimation data is determined using a principal component analysis method, including: The temperature estimation data is determined based on the product of the temperature reconstruction data and the rank-deficient element corresponding to the temperature sensor to be diagnosed.

5. The method according to any one of claims 1 to 4, characterized in that Determining whether the temperature sensor to be diagnosed is faulty according to the temperature estimation data includes: Determine whether the temperature estimation data exceeds a preset temperature range. When the temperature estimation data exceeds the preset temperature range, determine that the temperature sensor to be diagnosed is faulty; when the temperature estimation data is within the preset temperature range, determine that the temperature sensor to be diagnosed is normal.

6. The method according to any one of claims 1 to 4, characterized in that Determining whether the temperature sensor to be diagnosed is faulty according to the temperature estimation data includes: Calculating a residual between the temperature measurement data corresponding to the temperature sensor to be diagnosed and the temperature estimation data; The test statistic corresponding to the residual is determined by the following expression: represents the detection statistic corresponding to the i-th temperature sensor at time k, j = 1, 2, ..., n; q represents the length of the data window selected when obtaining the residual; p represents time, p = k-q+1, k-q+2, ..., k; is the variance of the residual calculated based on the historical temperature sample data of the i-th temperature sensor in a normal state; Based on the detection statistic and the detection threshold, determining whether the temperature sensor to be diagnosed is faulty; When the detection statistic is greater than the detection threshold, it is determined that the temperature sensor to be diagnosed is faulty; when the detection statistic is less than or equal to the detection threshold, it is determined that the temperature sensor to be diagnosed is normal.

7. The method according to claim 6, characterized in that Before determining whether the temperature sensor to be diagnosed is faulty based on the detection statistic and the detection threshold, the method further includes: The corresponding detection threshold is determined according to the preset signal level and the data window length.

8. A storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by the processor, the temperature sensor fault diagnosis method according to any one of claims 1 to 7 is implemented.

9. A traction system, characterized in that: The method comprises a memory and a processor, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the temperature sensor fault diagnosis method according to any one of claims 1 to 7 is implemented.

10. The traction system according to claim 9, characterized in that: The traction system further includes: a traction transformer, a traction converter, an oil temperature sensor for detecting the oil temperature in the traction transformer, and a water temperature sensor for detecting the cooling water in the traction converter.

Images