Motor train unit gearbox fault detection method and related device

Abstract

The invention discloses a motor train unit gearbox fault detection method and a related device, and the method comprises the steps: obtaining a gearbox on each axle of a current motor train unit, taking each transmission connection side as the current temperature data corresponding to a measurement dimension, summing the temperature difference of each axle and other axles of the same carriage under the same measurement dimension, and calculating the temperature difference of each axle; the method comprises the following steps: acquiring a current temperature characteristic value of an axle, and judging whether the axle is an abnormal axle or not according to a magnitude relationship between the current temperature characteristic value and a temperature threshold value under the same measurement dimension so as to detect faults of a gearbox of the motor train unit, and the temperature threshold value is based on historical temperature data of the axle under each measurement dimension. And determining by using a multi-algorithm collaborative outlier comprehensive detection technology. According to the invention, the temperature of each axle in the same compartment under the same measurement dimension is compared, so that the influence caused by the train operation environment and the individual difference of the gear box is reduced; the temperature threshold value is determined through a multi-algorithm cooperative outlier comprehensive detection technology, and the occurrence of false alarm and missing alarm conditions is reduced.

Description

Technical Field

[0001] This invention relates to the field of fault detection technology, and more specifically, to a method and related apparatus for detecting faults in the gearbox of a high-speed train. Background Technology

[0002] As a core component of the train's transmission system, the gearbox of a high-speed train is prone to failure under high-speed, heavy-load, and complex operating environments. Existing methods for detecting gearbox failures mainly rely on human experience: these methods are based on experts' understanding of the physical mechanisms of gearbox transmission, lubrication, and temperature rise characteristics. Typically, they estimate the temperature range of the gearbox under normal operating conditions by establishing heat conduction equations or power loss models, and then set empirical thresholds based on this to detect failures.

[0003] However, existing detection methods are highly sensitive to model assumptions and struggle to fully account for complex train operating environments (such as climate and track conditions) and individual differences between gearboxes (such as manufacturing errors and wear levels). This makes the setting of empirical thresholds overly reliant on human experience, highly subjective, and with weak generalization ability. They are easily affected by train operating environments and individual gearbox differences, thus leading to false alarms and missed alarms. Summary of the Invention

[0004] In view of this, the present invention discloses a method and related device for detecting gearbox faults in high-speed trains. By comparing the temperatures of various axles within the same carriage under the same measurement dimension, the temperature characteristic values ​​of the axles are determined, effectively reducing the impact of the train operating environment and individual differences in gearboxes. Simultaneously, the multi-algorithm collaborative outlier detection technology integrates the advantages of multiple algorithms. Compared to methods relying on human experience, it can accurately, dynamically, and intelligently determine temperature thresholds, thereby effectively reducing false alarms and missed alarms.

[0005] A method for detecting faults in a high-speed train gearbox includes:

[0006] Obtain the current temperature data of the gearboxes on each axle of the current EMU, with each transmission connection side as the measurement dimension;

[0007] Based on the current temperature data, for each axle, the temperature difference between it and other axles in the same carriage under the same measurement dimension is summed, and the sum is used as the current temperature characteristic value of the axle.

[0008] Based on the relationship between the current temperature feature value and the corresponding temperature threshold under the same measurement dimension, it is determined whether the axle is an abnormal axle, thereby realizing the fault detection of the EMU gearbox. The temperature threshold is determined by using a multi-algorithm collaborative outlier detection technology based on the historical temperature data of the axle under each measurement dimension.

[0009] Optionally, the process of determining the temperature threshold includes:

[0010] Obtain historical temperature data for each gearbox on each axle of the EMU within a preset time period before the current moment, with each transmission connection side as the measurement dimension;

[0011] The historical temperature data is preprocessed using a linear interpolation method to remove any missing values, thus obtaining the historical target temperature data.

[0012] Based on the historical target temperature data, for each axle, the temperature difference between it and other axles in the same carriage under the same measurement dimension is summed, and the sum is used as the temperature characteristic value of the axle.

[0013] Based on the outlier detection technology of multi-algorithm collaboration, the temperature feature value of each axle is processed to obtain the temperature threshold corresponding to the axle.

[0014] Optionally, based on a multi-algorithm collaborative outlier detection technology, the temperature feature value of each axle is processed to obtain the temperature threshold corresponding to the axle, including:

[0015] The outlier detection method of local outlier factor is used to detect outliers in the temperature feature value of the axle to obtain the first outlier data point;

[0016] Anomaly detection methods based on isolated forests are used to detect outliers in the temperature characteristic values ​​of the axle, thereby obtaining outlier second data points.

[0017] The intersection of the first data point and the second data point corresponding to the axle in the same measurement dimension is obtained to obtain the intersection data.

[0018] The minimum value in the intersection data is used as the temperature threshold.

[0019] Optionally, based on the current temperature data, for each axle, the temperature difference between it and other axles in the same compartment under the same measurement dimension is summed, and the resulting sum is used as the current temperature characteristic value of the axle, including:

[0020] For each of the current temperature data, a linear interpolation method is used to preprocess the areas with missing values ​​to obtain the corresponding current target temperature data;

[0021] Calculate the temperature difference between the current target temperature data corresponding to the axle and the current target temperature data of other axles in the same carriage under the same measurement dimension;

[0022] The temperature differences corresponding to the axle are summed, and the sum is used as the current temperature characteristic value of the axle.

[0023] Optionally, based on the relationship between the current temperature characteristic value and the corresponding temperature threshold under the same measurement dimension, it is determined whether the axle is an abnormal axle, thereby realizing fault detection of the EMU gearbox, including:

[0024] If, at the current moment, among the current temperature characteristic values ​​of the axle under various measurement dimensions, there is at least one current temperature characteristic value that is greater than the corresponding temperature threshold under the same measurement dimension, the axle is determined to be an abnormal axle, and the gearbox of the EMU is determined to have malfunctioned.

[0025] Optionally, based on the relationship between the current temperature characteristic value and the corresponding temperature threshold under the same measurement dimension, it is determined whether the axle is an abnormal axle, thereby realizing fault detection of the EMU gearbox, including:

[0026] When at least one of the current temperature feature values ​​of the axle under each measurement dimension is greater than the corresponding temperature threshold under the same measurement dimension, the axle is determined to be an abnormal axle.

[0027] If the same axle is identified as an abnormal axle multiple times within the current fault detection cycle, and the time it is in an abnormal state exceeds the preset percentage threshold of the current fault detection cycle, then the gearbox of the EMU is determined to have malfunctioned.

[0028] Optionally, it also includes:

[0029] In the event that a fault has been determined in the gearbox of the high-speed train, the temperature characteristic values ​​of the abnormal shaft during the abnormal state will be output as early warning data.

[0030] A fault detection device for a high-speed train gearbox includes:

[0031] The data acquisition unit is used to acquire the current temperature data of the gearboxes on each axle of the current EMU, with each transmission connection side as the measurement dimension;

[0032] The temperature feature value determination unit is used to sum the temperature difference between each axle and other axles in the same carriage under the same measurement dimension based on the current temperature data, and the sum is used as the current temperature feature value of the axle.

[0033] The fault detection unit is used to determine whether the axle is an abnormal axle based on the relationship between the current temperature feature value and the corresponding temperature threshold under the same measurement dimension, thereby realizing the fault detection of the EMU gearbox. The temperature threshold is determined by using a multi-algorithm collaborative outlier detection technology based on the historical temperature data of the axle under each measurement dimension.

[0034] A computer storage medium storing at least one instruction, which, when executed by a processor, implements the above-described method for detecting faults in a high-speed train gearbox.

[0035] An electronic device, the electronic device comprising: a memory and a processor;

[0036] The memory is used to store at least one instruction;

[0037] The processor is used to execute at least one instruction to implement the above-described method for detecting faults in the gearbox of a high-speed train.

[0038] As can be seen from the above technical solution, this invention discloses a method and related device for detecting gearbox faults in high-speed trains. It acquires the current temperature data of the gearboxes on each axle of the current high-speed train, using each transmission connection side as the measurement dimension. Based on the current temperature data, for each axle, the temperature difference between it and other axles in the same carriage under the same measurement dimension is summed. The sum is used as the current temperature characteristic value of the axle. Based on the relationship between the current temperature characteristic value and the corresponding temperature threshold under the same measurement dimension, it is determined whether the axle is an abnormal axle, thus achieving gearbox fault detection in high-speed trains. The temperature threshold is determined by using a multi-algorithm collaborative outlier detection technology based on the historical temperature data of the axle under each measurement dimension. This invention determines the temperature characteristic value of the axle by comparing the temperatures of each axle in the same carriage under the same measurement dimension, effectively reducing the impact of the train operating environment and individual differences in gearboxes. Simultaneously, the multi-algorithm collaborative outlier detection technology integrates the advantages of multiple algorithms, and compared to methods relying on human experience, it can accurately, dynamically, and intelligently determine the temperature threshold, thereby effectively reducing false alarms and missed alarms. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the published drawings without creative effort.

[0040] Figure 1This is a flowchart of a method for detecting gearbox faults in high-speed trains, as disclosed in an embodiment of the present invention.

[0041] Figure 2 This is a flowchart of a method for determining a temperature threshold disclosed in an embodiment of the present invention;

[0042] Figure 3 (1) is a schematic diagram of the result obtained by using the local outlier factor algorithm to detect outliers in a normal carriage according to an embodiment of the present invention;

[0043] Figure 3 (2) is a schematic diagram of the result obtained by using the local outlier factor algorithm to detect outliers in a faulty carriage according to an embodiment of the present invention.

[0044] Figure 4 This is an isolated forest-density outlier map disclosed in an embodiment of the present invention;

[0045] Figure 5 This is a schematic diagram of the structure of a high-speed train gearbox fault detection device disclosed in an embodiment of the present invention;

[0046] Figure 6 This is a schematic diagram of the structure of an electronic device disclosed in an embodiment of the present invention. Detailed Implementation

[0047] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0048] This invention discloses a method and related apparatus for detecting gearbox faults in high-speed trains. The method acquires the current temperature data of the gearboxes on each axle of the current high-speed train, using each transmission connection side as the measurement dimension. Based on this current temperature data, for each axle, the temperature difference between it and other axles in the same carriage under the same measurement dimension is summed. The sum is used as the current temperature characteristic value of the axle. Based on the relationship between the current temperature characteristic value and the corresponding temperature threshold under the same measurement dimension, it is determined whether the axle is an abnormal axle, thus achieving gearbox fault detection in high-speed trains. The temperature threshold is determined by using a multi-algorithm collaborative outlier detection technology based on the historical temperature data of the axle under each measurement dimension. This invention determines the temperature characteristic value of the axle by comparing the temperatures of each axle in the same carriage under the same measurement dimension, effectively reducing the impact of train operating environment and individual gearbox differences. Simultaneously, the multi-algorithm collaborative outlier detection technology integrates the advantages of multiple algorithms, and compared to methods relying on human experience, it can accurately, dynamically, and intelligently determine the temperature threshold, thereby effectively reducing false alarms and missed alarms.

[0049] See Figure 1 The present invention discloses a flowchart of a method for detecting faults in a high-speed train gearbox, the method comprising:

[0050] Step S101: Obtain the gearboxes on each axle of the current EMU, and use the current temperature data corresponding to each transmission connection side as the measurement dimension.

[0051] A high-speed train has multiple carriages, such as carriages 2, 4, 5, and 7. Each carriage has multiple axles, such as axle 1, axle 2, axle 3, and axle 4. Each gearbox has multiple transmission connection sides, namely: the large gear motor side (GM side), the large gear wheel side (GW side), the small gear motor side (PM side), and the small gear wheel side (PW side).

[0052] In this embodiment, each transmission connection side is used as a measurement dimension. The gearboxes on each axle of the current train are obtained through temperature sensors on the train. The current temperature data corresponding to each transmission connection side as the measurement dimension is obtained, that is, the current ambient temperature of the gearboxes on each axle of each carriage is obtained on the GM side, GW side, PM side and PW side.

[0053] Step S102: Based on the current temperature data, for each axle, sum the temperature difference between it and other axles in the same carriage under the same measurement dimension, and use the sum as the current temperature characteristic value of the axle.

[0054] Each axle has four drive connection sides: GM side, GW side, PM side, and PW side. In practical applications, it is necessary to determine the corresponding temperature characteristic value for each drive connection side.

[0055] Taking the determination of the temperature characteristic value of axle 1 of car 2 on the GM side as an example, assuming that the other axles in the same car as axle 1 of car 2 include: axle 2 of car 2, axle 3 of car 2, and axle 4 of car 2, then:

[0056] Temperature characteristic value of GM side of axle 1 of vehicle 2 = Temperature of GM side of axle 1 of vehicle 2 - Temperature of GM side of axle 2 of vehicle 2 + Temperature of GM side of axle 1 of vehicle 2 - Temperature of GM side of axle 3 of vehicle 2 + Temperature of GM side of axle 1 of vehicle 2 - Temperature of GM side of axle 4 of vehicle 2.

[0057] Step S103: Based on the relationship between the current temperature characteristic value and the corresponding temperature threshold under the same measurement dimension, determine whether the axle is an abnormal axle, thereby realizing the fault detection of the EMU gearbox.

[0058] This invention pre-determines the corresponding temperature thresholds based on historical temperature data of the axle in each measurement dimension using a multi-algorithm collaborative outlier detection technology. By comparing the current temperature feature value of the axle in any measurement dimension with the corresponding temperature threshold in the same measurement dimension, it determines whether the axle is an abnormal axle, thereby achieving fault detection in the gearbox of the high-speed train.

[0059] This invention employs a multi-algorithm collaborative outlier detection technique to process historical temperature data and determine corresponding temperature thresholds. It integrates the detection results of different outlier algorithms to form a "collective decision." For example, a weighted voting mechanism can be used to assign different weights to different outlier detection methods, improving the accuracy of temperature threshold determination. Cross-validation can also be used: if multiple outlier algorithms simultaneously determine a temperature point as an anomaly, the temperature threshold can be determined based on the commonly identified anomalous temperature points, helping to increase the confidence level of the temperature threshold and reduce false positives.

[0060] In practical applications, if the current temperature characteristic value of an axle in any measurement dimension is greater than the corresponding temperature threshold in the same measurement dimension, the axle is usually identified as an abnormal axle. Applying this determination method to all axles of the current high-speed train enables fault detection in the train's gearbox.

[0061] In summary, this invention discloses a method for detecting gearbox faults in high-speed trains. It acquires the current temperature data of the gearboxes on each axle of the current high-speed train, using each transmission connection side as the measurement dimension. Based on this current temperature data, for each axle, the temperature difference between it and other axles in the same carriage under the same measurement dimension is summed. The sum is used as the current temperature characteristic value of the axle. Based on the relationship between the current temperature characteristic value and the corresponding temperature threshold under the same measurement dimension, it is determined whether the axle is an abnormal axle, thus achieving gearbox fault detection in high-speed trains. The temperature threshold is determined by using a multi-algorithm collaborative outlier detection technology based on the historical temperature data of the axle under each measurement dimension. This invention determines the temperature characteristic value of the axle by comparing the temperatures of each axle in the same carriage under the same measurement dimension, effectively reducing the impact of train operating environment and individual gearbox differences. Simultaneously, the multi-algorithm collaborative outlier detection technology integrates the advantages of multiple algorithms, and compared to methods relying on human experience, it can accurately, dynamically, and intelligently determine the temperature threshold, thereby effectively reducing false alarms and missed alarms.

[0062] In one embodiment, the fault detection method for a high-speed train gearbox also provides a process for determining a temperature threshold, see [link to relevant documentation]. Figure 2 The present invention discloses a flowchart of a method for determining a temperature threshold, the method comprising:

[0063] Step S201: Obtain historical temperature data of the gearboxes on each axle of the EMU within a preset time period before the current moment, with each transmission connection side as the measurement dimension.

[0064] The value of the preset time period before the current moment can be determined according to actual needs, such as one year.

[0065] For example, obtain historical temperature data for the past year from the gearboxes on each axle (such as axle 1, axle 2, axle 3, and axle 4) of each carriage (such as carriage 2, carriage 4, carriage 5, and carriage 7) of the EMU, as well as from each transmission connection side (large gear motor side, large gear wheel side, small gear motor side, and small gear wheel side).

[0066] Step S202: The historical temperature data is preprocessed using a linear interpolation method to remove any missing values, thereby obtaining the historical target temperature data.

[0067] Linear interpolation is a numerical method for estimating the value of unknown data points based on the linear relationship between known data points. Its core principle is to construct a straight line using two adjacent known points and then use the equation of this straight line to approximate the unknown value between the two points.

[0068] Historical temperature data may contain missing values ​​due to equipment malfunctions, data transmission errors, or human error. Linear interpolation methods can use known temperature data points to estimate the temperature values ​​of missing points by establishing a linear relationship, thereby filling in the gaps in the dataset and making the temperature dataset more complete.

[0069] Step S203: Based on historical target temperature data, for each axle, sum the temperature difference between it and other axles in the same carriage under the same measurement dimension, and use the sum as the temperature characteristic value of the axle.

[0070] In this embodiment, each axle's gearbox has historical target temperature data corresponding to each transmission connection side. For each axle, the temperature difference between it and other axles in the same carriage under the same measurement dimension is first calculated, and then these temperature differences are summed. The sum is used as the temperature characteristic value of that axle. By comparing the temperature of each axle in the same carriage under the same measurement dimension to determine the temperature characteristic value of the axle, this method can effectively reduce the impact of train operating environment and individual gearbox differences.

[0071] Step S204: Based on the multi-algorithm collaborative outlier detection technology, the temperature feature value of each axle is processed to obtain the temperature threshold corresponding to the axle.

[0072] Different outlier detection algorithms have their own applicable scenarios and advantages. For example, statistical methods (such as Z-score) make certain assumptions about the data distribution and can detect outliers well when the data conforms to or is approximately normally distributed. Distance-based methods (such as the K-nearest neighbor algorithm) do not rely on data distribution assumptions and determine outliers by calculating the distance between samples. Multi-algorithm collaborative outlier detection technology can combine the advantages of these algorithms, analyze axle temperature characteristics from different perspectives, more accurately identify true outliers, and thus obtain more reasonable temperature thresholds.

[0073] In practical applications, the outlier detection algorithms included in the multi-algorithm collaborative outlier detection technology can be: outlier detection methods based on local outlier factors, outlier detection methods based on isolated forests, statistical methods (such as Z-score), distance-based methods (such as the K-nearest neighbor algorithm), and so on.

[0074] The following section uses the outlier detection method for local outliers and the anomaly detection method for isolated forests as examples to illustrate the specific implementation process of step S204:

[0075] (1) The outlier detection method of local outlier factor is used to detect outliers in the temperature characteristic value of the axle and obtain the first outlier data point.

[0076] The Local Outlier Factor (LOF) is a density-based anomaly detection algorithm that identifies outliers by comparing the density differences between a data point and its neighborhood. Its core concept is that the local density of an outlier is significantly lower than that of its neighbors. In other words, the core idea of ​​the LOF is "density comparison." It calculates the k-neighborhood of a sample, using "locally accessible density" to measure the density of the sample and its neighborhood, and then uses the density ratio of the sample to its neighbors (i.e., the LOF value) to determine anomalies. When the sample density is much lower than the surrounding samples (LOF value much greater than 1), it is considered a local anomaly. It can quantify the degree of anomaly, but it relies on distance calculations and is sensitive to local density differences.

[0077] Taking the LOF algorithm with a neighbor count (n_neighbors) of 20 as an example, the proportion of outliers in the dataset is 0.01. Figure 3(1) is a schematic diagram of the results obtained by using the Local Outlier Factor algorithm to detect outliers in normal carriages, and Figure 3(2) is a schematic diagram of the results obtained by using the Local Outlier Factor algorithm to detect outliers in faulty carriages. The three coordinate axes in the three-dimensional space of Figure 3(1) and Figure 3(2) represent three feature value dimensions (such as the GM side feature value, GW side feature value, and PM side feature value of carriage 10). Each data point corresponds to a set of monitoring values ​​of these three dimensions, which together constitute the "feature space" of the equipment operating status. The closer the color is to dark red, the larger the LOF value is, and vice versa. The LOF value is used to measure the "outlier degree" of the data point - the larger the LOF value, the higher the deviation of the data point from the surrounding normal data points, and the more likely it is to be an "outlier"; when the LOF value is close to 1, the data point is more in line with the distribution pattern of normal data.

[0078] Figure 3(1) shows the characteristic value distribution of a normal carriage. It can be seen that the characteristic values ​​are all below 25. Figure 3(2) shows the characteristic value distribution of a faulty carriage. The characteristic values ​​of the abnormal points are mostly above 30, and the maximum value can reach 60.

[0079] (2) The outlier detection method of isolated forest is used to detect outliers in the temperature feature value of the axle and obtain the second data point of the outlier.

[0080] The core concept of Isolation Forest is "fast isolation". It constructs multiple isolation trees by randomly selecting features and splitting thresholds. It takes advantage of the fact that outliers are easier to quickly separate and isolate because they are few in number and have special features. It judges anomalies by the "path length" required for a sample to be isolated in the forest - the shorter the path, the higher the probability of an anomaly. It does not rely on data density or distance calculations and pursues efficient detection.

[0081] Taking the outlier detection method in isolated forests with a maximum number of estimators (max_samples) of 30 as an example, the proportion of outliers in the total sample (contamination) is 0.01. The outliers are as follows: Figure 4 The isolated forest-density outlier map shown in the image represents normal data identified using the isolated forest outlier detection method, and outlier data identified using the same method. The outlier threshold is approximately between 30 and 35.

[0082] (3) Take the intersection of the first and second data points corresponding to the axle in the same measurement dimension to obtain the intersection data.

[0083] By taking into account the calculation results of the Local Outlier Factor (LOF) and the Isolated Forest algorithm (i.e., the first data point and the second data point), the intersection data (such as 30 and 35) is obtained.

[0084] (4) Take the minimum value in the intersection data as the temperature threshold.

[0085] In this embodiment, the minimum value in the intersection data is used as the final determined temperature threshold.

[0086] In one embodiment, step S102 may specifically include:

[0087] (1) For each current temperature data, a linear interpolation method is used to preprocess the places where there are missing values ​​to obtain the corresponding current target temperature data.

[0088] (2) Calculate the temperature difference between the current target temperature data of the axle and the current target temperature data of other axles in the same carriage under the same measurement dimension.

[0089] (3) Sum all the temperature differences corresponding to the axle and use the sum as the current temperature characteristic value of the axle.

[0090] Each axle has four drive connection sides: GM side, GW side, PM side, and PW side. In practical applications, it is necessary to determine the corresponding temperature characteristic value for each drive connection side.

[0091] Taking the determination of the temperature characteristic value of axle 1 of car 2 on the GM side as an example, assuming that the other axles in the same car as axle 1 of car 2 include: axle 2 of car 2, axle 3 of car 2, and axle 4 of car 2, then:

[0092] Temperature characteristic value of GM side of axle 1 of vehicle 2 = Temperature of GM side of axle 1 of vehicle 2 - Temperature of GM side of axle 2 of vehicle 2 + Temperature of GM side of axle 1 of vehicle 2 - Temperature of GM side of axle 3 of vehicle 2 + Temperature of GM side of axle 1 of vehicle 2 - Temperature of GM side of axle 4 of vehicle 2.

[0093] In one embodiment, step S103 may specifically include:

[0094] If, at the current moment, among the current temperature characteristic values ​​of the axle under various measurement dimensions, there is at least one current temperature characteristic value that is greater than the corresponding temperature threshold under the same measurement dimension, the axle is determined to be an abnormal axle, and the gearbox of the EMU is determined to have malfunctioned.

[0095] Each axle has four drive connection sides: GM side, GW side, PM side, and PW side. Each drive connection side serves as a measurement dimension. In practical applications, it is necessary to determine the corresponding temperature characteristic value and temperature threshold for each measurement dimension.

[0096] Therefore, it is necessary to compare the current temperature characteristic value of the axle in each measurement dimension with the corresponding temperature threshold in the same measurement dimension. If the current temperature characteristic value is greater than the corresponding temperature threshold, for example, if the current temperature characteristic value of the GM side is greater than the temperature threshold of the GM side, then it is determined that there is a temperature abnormality on the GM side.

[0097] If at least one of the current temperature characteristic values ​​of the axle in each measurement dimension is greater than the corresponding temperature threshold in the same measurement dimension, the present invention determines that the axle is an abnormal axle, and at this time it is determined that the EMU gearbox has failed.

[0098] In summary, this invention effectively avoids false alarms and missed alarms caused by sudden temperature changes or abnormal temperature sensors through multi-point detection.

[0099] Based on the above discussion, it can be seen that this invention employs a dual detection mechanism for fault detection in high-speed train gearboxes: lateral comparison and longitudinal comparison. Lateral comparison involves comparing the current temperature characteristic value of the same measuring point with a temperature threshold determined based on its own historical temperature data; longitudinal comparison involves analyzing the correlation between different points on the same axle. Therefore, this invention effectively reduces the impact of train operating environment and individual gearbox differences, and effectively reduces the occurrence of false alarms and missed alarms.

[0100] To effectively avoid errors in EMU gearbox fault detection caused by single-moment detection and improve the accuracy of EMU gearbox fault detection, this invention further adopts a continuous anomaly monitoring method to ultimately determine whether the gearbox has malfunctioned.

[0101] Therefore, in one embodiment, step S103 may further include:

[0102] When at least one of the current temperature feature values ​​of the axle under each measurement dimension is greater than the corresponding temperature threshold under the same measurement dimension, the axle is determined to be an abnormal axle.

[0103] If the same axle is identified as an abnormal axle multiple times within the current fault detection cycle, and the time it is in an abnormal state exceeds the preset percentage threshold of the current fault detection cycle, then the train's gearbox is determined to have malfunctioned.

[0104] The value of the preset percentage threshold is determined according to actual needs, such as 70%.

[0105] Taking a current fault detection cycle of one day as an example, at any point during the current fault detection cycle, if the axle has at least one current temperature characteristic value that is greater than the corresponding temperature threshold under the same measurement dimension, the axle is determined to be an abnormal axle at that moment. If the axle is determined to be an abnormal axle multiple times during the current fault detection cycle, and the time spent in an abnormal state exceeds 70% of the current fault detection cycle, then the axle is determined to have a fault, thereby determining that the EMU gearbox has a fault.

[0106] In summary, this invention determines a train gearbox malfunction only when the same axle is repeatedly identified as an abnormal axle within the current fault detection cycle, and the duration of the axle's abnormal state exceeds a preset percentage threshold for the current fault detection cycle. This method effectively identifies situations where axles continuously deviate from their normal trend, avoids delays in early fault identification, and mitigates gearbox fault detection errors caused by detection at a single moment, significantly improving the accuracy of train gearbox fault detection.

[0107] In one embodiment, the method for detecting gearbox faults in high-speed trains further includes:

[0108] In the event that a fault has been detected in the gearbox of a high-speed train, the temperature characteristic values ​​of the abnormal shaft during the abnormal state will be output as early warning data.

[0109] This invention, when a gearbox malfunction is detected in a high-speed train, outputs warning data by showing the temperature characteristics of the abnormal axle during its abnormal state. This allows maintenance personnel to quickly pinpoint the specific axle causing the fault, avoiding a comprehensive inspection of the entire gearbox, saving significant time and effort, and improving the accuracy of fault location. For example, if the output data shows a sustained abnormal temperature increase on a particular axle, maintenance personnel can focus directly on that axle for detailed inspection and quickly find the root cause of the fault. Furthermore, different types of faults often exhibit different patterns in temperature characteristics. By analyzing the temperature characteristics of the abnormal axle, such as the rate of temperature rise and the range of temperature fluctuations, maintenance personnel can preliminarily determine the type of fault, such as bearing wear or poor gear meshing. For example, a slow temperature increase with small fluctuations may indicate insufficient bearing lubrication leading to wear; a rapid temperature increase with large fluctuations may indicate a problem with gear meshing causing severe friction. This facilitates the implementation of appropriate measures.

[0110] Corresponding to the above method embodiments, the present invention also discloses a fault detection device for a high-speed train gearbox.

[0111] See Figure 5 The present invention discloses a structural schematic diagram of a fault detection device for a high-speed train gearbox, the method comprising:

[0112] The data acquisition unit 301 is used to acquire the current temperature data of the gearboxes on each axle of the current EMU, with each transmission connection side as the measurement dimension.

[0113] A high-speed train has multiple carriages, such as carriages 2, 4, 5, and 7. Each carriage has multiple axles, such as axle 1, axle 2, axle 3, and axle 4. Each gearbox has multiple transmission connection sides, namely: the large gear motor side (GM side), the large gear wheel side (GW side), the small gear motor side (PM side), and the small gear wheel side (PW side).

[0114] In this embodiment, each transmission connection side is used as a measurement dimension. The gearboxes on each axle of the current train are obtained through temperature sensors on the train. The current temperature data corresponding to each transmission connection side as the measurement dimension is obtained, that is, the current ambient temperature of the gearboxes on each axle of each carriage is obtained on the GM side, GW side, PM side and PW side.

[0115] Temperature feature value determination unit 302 is used to sum the temperature difference between each axle and other axles in the same carriage under the same measurement dimension based on the current temperature data, and the sum is used as the current temperature feature value of the axle.

[0116] Each axle has four drive connection sides: GM side, GW side, PM side, and PW side. In practical applications, it is necessary to determine the corresponding temperature characteristic value for each drive connection side.

[0117] Taking the determination of the temperature characteristic value of axle 1 of car 2 on the GM side as an example, assuming that the other axles in the same car as axle 1 of car 2 include: axle 2 of car 2, axle 3 of car 2, and axle 4 of car 2, then:

[0118] Temperature characteristic value of GM side of axle 1 of vehicle 2 = Temperature of GM side of axle 1 of vehicle 2 - Temperature of GM side of axle 2 of vehicle 2 + Temperature of GM side of axle 1 of vehicle 2 - Temperature of GM side of axle 3 of vehicle 2 + Temperature of GM side of axle 1 of vehicle 2 - Temperature of GM side of axle 4 of vehicle 2.

[0119] The fault detection unit 303 is used to determine whether the axle is an abnormal axle based on the relationship between the current temperature characteristic value and the corresponding temperature threshold under the same measurement dimension, thereby realizing the fault detection of the EMU gearbox.

[0120] This invention pre-determines the corresponding temperature thresholds based on historical temperature data of the axle in each measurement dimension using a multi-algorithm collaborative outlier detection technology. By comparing the current temperature feature value of the axle in any measurement dimension with the corresponding temperature threshold in the same measurement dimension, it determines whether the axle is an abnormal axle, thereby achieving fault detection in the gearbox of the high-speed train.

[0121] This invention employs a multi-algorithm collaborative outlier detection technique to process historical temperature data and determine corresponding temperature thresholds. It integrates the detection results of different outlier algorithms to form a "collective decision." For example, a weighted voting mechanism can be used to assign different weights to different outlier detection methods, improving the accuracy of temperature threshold determination. Cross-validation can also be used: if multiple outlier algorithms simultaneously determine a temperature point as an anomaly, the temperature threshold can be determined based on the commonly identified anomalous temperature points, helping to increase the confidence level of the temperature threshold and reduce false positives.

[0122] In practical applications, if the current temperature characteristic value of an axle in any measurement dimension is greater than the corresponding temperature threshold in the same measurement dimension, the axle is usually identified as an abnormal axle. Applying this determination method to all axles of the current high-speed train enables fault detection in the train's gearbox.

[0123] In summary, this invention discloses a gearbox fault detection device for high-speed trains. It acquires the current temperature data of the gearboxes on each axle of the current high-speed train, using each transmission connection side as the measurement dimension. Based on this current temperature data, for each axle, the temperature difference between it and other axles in the same carriage under the same measurement dimension is summed. The sum is used as the current temperature characteristic value of the axle. Based on the relationship between the current temperature characteristic value and the corresponding temperature threshold under the same measurement dimension, it determines whether the axle is an abnormal axle, thus achieving gearbox fault detection. The temperature threshold is determined by using a multi-algorithm collaborative outlier detection technology based on the historical temperature data of the axle under each measurement dimension. This invention determines the temperature characteristic value of the axle by comparing the temperatures of each axle in the same carriage under the same measurement dimension, effectively reducing the impact of train operating environment and individual gearbox differences. Simultaneously, the multi-algorithm collaborative outlier detection technology integrates the advantages of multiple algorithms, and compared to methods relying on human experience, it can accurately, dynamically, and intelligently determine the temperature threshold, thereby effectively reducing false alarms and missed alarms.

[0124] In one embodiment, the high-speed train gearbox fault detection device may further include:

[0125] Temperature threshold determination unit, used to determine temperature threshold.

[0126] The temperature threshold determination unit can specifically be used for:

[0127] Obtain historical temperature data for each gearbox on each axle of the EMU within a preset time period before the current moment, with each transmission connection side as the measurement dimension;

[0128] The historical temperature data is preprocessed using a linear interpolation method to remove any missing values, thus obtaining the historical target temperature data.

[0129] Based on the historical target temperature data, for each axle, the temperature difference between it and other axles in the same carriage under the same measurement dimension is summed, and the sum is used as the temperature characteristic value of the axle.

[0130] Based on the outlier detection technology of multi-algorithm collaboration, the temperature feature value of each axle is processed to obtain the temperature threshold corresponding to the axle.

[0131] In one embodiment, the temperature threshold determination unit can also be used for:

[0132] The outlier detection method of local outlier factor is used to detect outliers in the temperature feature value of the axle to obtain the first outlier data point;

[0133] Anomaly detection methods based on isolated forests are used to detect outliers in the temperature characteristic values ​​of the axle, thereby obtaining outlier second data points.

[0134] The intersection of the first data point and the second data point corresponding to the axle in the same measurement dimension is obtained to obtain the intersection data.

[0135] The minimum value in the intersection data is used as the temperature threshold.

[0136] In one embodiment, the temperature characteristic value determination unit 302 can be specifically used for:

[0137] For each of the current temperature data, a linear interpolation method is used to preprocess the areas with missing values ​​to obtain the corresponding current target temperature data;

[0138] Calculate the temperature difference between the current target temperature data corresponding to the axle and the current target temperature data of other axles in the same carriage under the same measurement dimension;

[0139] The temperature differences corresponding to the axle are summed, and the sum is used as the current temperature characteristic value of the axle.

[0140] In one embodiment, the fault detection unit 303 can be specifically used for:

[0141] If, at the current moment, among the current temperature characteristic values ​​of the axle under various measurement dimensions, there is at least one current temperature characteristic value that is greater than the corresponding temperature threshold under the same measurement dimension, the axle is determined to be an abnormal axle, and the gearbox of the EMU is determined to have malfunctioned.

[0142] In one embodiment, the fault detection unit 303 can also be used for:

[0143] When at least one of the current temperature feature values ​​of the axle under each measurement dimension is greater than the corresponding temperature threshold under the same measurement dimension, the axle is determined to be an abnormal axle.

[0144] If the same axle is identified as an abnormal axle multiple times within the current fault detection cycle, and the time it is in an abnormal state exceeds the preset percentage threshold of the current fault detection cycle, then the gearbox of the EMU is determined to have malfunctioned.

[0145] In one embodiment, the high-speed train gearbox fault detection device may further include:

[0146] The output unit is used to output the temperature characteristic values ​​of the abnormal shaft during the abnormal state as early warning data when it is determined that the gearbox of the EMU has failed.

[0147] Corresponding to the above embodiments, the present invention also discloses a computer storage medium that stores at least one instruction, which, when executed by a processor, implements the steps shown in the embodiments of the high-speed train gearbox fault detection method.

[0148] Corresponding to the above embodiments, such as Figure 6 As shown, the present invention also provides a schematic diagram of the structure of an electronic device, which may include: a processor 1 and a memory 2;

[0149] The processor 1 and memory 2 communicate with each other via communication bus 3.

[0150] Processor 1, for executing at least one instruction;

[0151] Memory 2 is used to store at least one instruction;

[0152] Processor 1 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention.

[0153] Memory 2 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0154] The processor executes at least one instruction to implement the steps shown in the embodiment of the high-speed train gearbox fault detection method.

[0155] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0156] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0157] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for detecting faults in a high-speed train gearbox, characterized in that, include: Obtain the current temperature data of the gearboxes on each axle of the current EMU, with each transmission connection side as the measurement dimension; Based on the current temperature data, for each axle, the temperature difference between it and other axles in the same carriage under the same measurement dimension is summed, and the sum is used as the current temperature characteristic value of the axle. Based on the relationship between the current temperature feature value and the corresponding temperature threshold under the same measurement dimension, it is determined whether the axle is an abnormal axle, thereby realizing the fault detection of the EMU gearbox. The temperature threshold is determined by using a multi-algorithm collaborative outlier detection technology based on the historical temperature data of the axle under each measurement dimension.

2. The method for detecting faults in a high-speed train gearbox according to claim 1, characterized in that, The process of determining the temperature threshold includes: Obtain historical temperature data for each gearbox on each axle of the EMU within a preset time period before the current moment, with each transmission connection side as the measurement dimension; The historical temperature data is preprocessed using a linear interpolation method to remove any missing values, thus obtaining the historical target temperature data. Based on the historical target temperature data, for each axle, the temperature difference between it and other axles in the same carriage under the same measurement dimension is summed, and the sum is used as the temperature characteristic value of the axle. Based on the outlier detection technology of multi-algorithm collaboration, the temperature feature value of each axle is processed to obtain the temperature threshold corresponding to the axle.

3. The method for detecting faults in a high-speed train gearbox according to claim 2, characterized in that, Based on a multi-algorithm collaborative outlier detection technology, the temperature feature value of each axle is processed to obtain the temperature threshold corresponding to the axle, including: The outlier detection method of local outlier factor is used to detect outliers in the temperature feature value of the axle to obtain the first outlier data point; Anomaly detection methods based on isolated forests are used to detect outliers in the temperature characteristic values ​​of the axle, thereby obtaining outlier second data points. The intersection of the first data point and the second data point corresponding to the axle in the same measurement dimension is obtained to obtain the intersection data. The minimum value in the intersection data is used as the temperature threshold.

4. The method for detecting faults in a high-speed train gearbox according to any one of claims 1 to 3, characterized in that, Based on the current temperature data, for each axle, the temperature difference between it and other axles in the same compartment under the same measurement dimension is summed, and the resulting sum is used as the current temperature characteristic value of the axle, including: For each of the current temperature data, a linear interpolation method is used to preprocess the areas with missing values ​​to obtain the corresponding current target temperature data; Calculate the temperature difference between the current target temperature data corresponding to the axle and the current target temperature data of other axles in the same carriage under the same measurement dimension; The temperature differences corresponding to the axle are summed, and the sum is used as the current temperature characteristic value of the axle.

5. The method for detecting faults in a high-speed train gearbox according to any one of claims 1 to 3, characterized in that, Based on the relationship between the current temperature characteristic value and the corresponding temperature threshold under the same measurement dimension, it is determined whether the axle is an abnormal axle, thereby realizing the fault detection of the EMU gearbox, including: If, at the current moment, among the current temperature characteristic values ​​of the axle under various measurement dimensions, there is at least one current temperature characteristic value that is greater than the corresponding temperature threshold under the same measurement dimension, the axle is determined to be an abnormal axle, and the gearbox of the EMU is determined to have malfunctioned.

6. The method for detecting faults in a high-speed train gearbox according to any one of claims 1 to 3, characterized in that, Based on the relationship between the current temperature characteristic value and the corresponding temperature threshold under the same measurement dimension, it is determined whether the axle is an abnormal axle, thereby realizing the fault detection of the EMU gearbox, including: When at least one of the current temperature feature values ​​of the axle under each measurement dimension is greater than the corresponding temperature threshold under the same measurement dimension, the axle is determined to be an abnormal axle. If the same axle is identified as an abnormal axle multiple times within the current fault detection cycle, and the time it is in an abnormal state exceeds the preset percentage threshold of the current fault detection cycle, then the gearbox of the EMU is determined to have malfunctioned.

7. The method for detecting faults in a high-speed train gearbox according to claim 6, characterized in that, Also includes: In the event that a fault has been determined in the gearbox of the high-speed train, the temperature characteristic values ​​of the abnormal shaft during the abnormal state will be output as early warning data.

8. A fault detection device for a high-speed train gearbox, characterized in that, include: The data acquisition unit is used to acquire the current temperature data of the gearboxes on each axle of the current EMU, with each transmission connection side as the measurement dimension; The temperature feature value determination unit is used to sum the temperature difference between each axle and other axles in the same carriage under the same measurement dimension based on the current temperature data, and the sum is used as the current temperature feature value of the axle. The fault detection unit is used to determine whether the axle is an abnormal axle based on the relationship between the current temperature feature value and the corresponding temperature threshold under the same measurement dimension, thereby realizing the fault detection of the EMU gearbox. The temperature threshold is determined by using a multi-algorithm collaborative outlier detection technology based on the historical temperature data of the axle under each measurement dimension.

9. A computer storage medium, characterized in that, The computer storage medium stores at least one instruction, which, when executed by the processor, implements the high-speed train gearbox fault detection method as described in any one of claims 1 to 7.

10. An electronic device, characterized in that, The electronic device includes: a memory and a processor; The memory is used to store at least one instruction; The processor is used to execute the at least one instruction to implement the EMU gearbox fault detection method as described in any one of claims 1 to 7.

Images