Construction, prediction method and system for a key component life prediction model of a wheel set

Abstract

The present invention provides a method and system for constructing and predicting a life prediction model for key components of a wheel set, including: determining that the model input parameters are ambient temperature, driving speed, and driving mileage, and the output parameter is the remaining wear amount of the wheel diameter; obtaining the current ambient temperature, driving speed, and driving mileage of the wheel set in real time, and obtaining the turning data of the wheel set as sample data; performing preprocessing on the above sample data; respectively constructing models for the preprocessed sample data, and finally selecting the model with the highest verification sample accuracy as the optimal prediction model for the current wheel set. And the remaining life of the wheel is evaluated and predicted by combining the wear data and development trend of the geometric parts of the wheel with the driving mileage, and the prediction accuracy is high.

Description

Technical Field

[0001] The present invention belongs to the technical field of prediction, and particularly relates to a method and system for constructing a life prediction model of key components of a wheel set and a prediction method thereof. Background Technique

[0002] The statements in this part only provide background technical information related to the present invention, and do not necessarily constitute prior art.

[0003] The wheel set is one of the key components of the EMU, which plays a crucial role in the actual operation of the EMU. The wheel set plays a vital role in the operation of the EMU. Wheel set failures not only endanger the train operation safety, but also cause huge impacts and damages to the railway lines and bridges, exacerbate the fatigue damage degree of the vehicle structure, and increase the maintenance cost. The traditional methods for inspecting wheel set failures are: one is static inspection, that is, when the vehicle is under regular maintenance, the wheel set is disassembled and measured with a special mechanical ruler; the other is dynamic inspection, that is, the train inspection personnel use manual operation methods such as hammering, touching, and seeing to inspect the in-service vehicles, resulting in the failure to detect the tread failures of the in-service wheel sets in a timely manner, and missed inspections and misinspections often occur. Therefore, it is crucial to detect the state of the wheel set.

[0004] Under normal wheel-rail matching and wheel wear conditions, both the wheel flange and the tread will be worn. Especially the wear of the wheel flange, as the running time increases, the flange thickness gradually decreases, and the flange becomes thinner, which will pose a threat to the safe operation of the train. Therefore, turning operations are required to restore the flange thickness. The wheel turning is a maintenance method that restores the flange thickness at the cost of losing part of the tread diameter. The reduction of the flange thickness and the tread diameter caused by wear, as well as the reduction of the tread diameter caused by turning, are all key factors affecting the service life of the wheel. The schematic diagram of the influence of wear and turning on the tread of the EMU wheel is as follows Figure 1 shown.

[0005] The remaining life of the wheel set is mainly determined by parameters such as the amount of each turning of the wheel set and the mileage of the vehicle. The remaining life is generally estimated by the due date of the wheel diameter value. The due date of the wheel diameter value is determined by the last turning time and the remaining running time, and the determination of the remaining running time is the key, which is determined by the remaining running kilometers. Next, the focus is on determining the remaining running kilometers, which is mainly determined by the remaining wear amount of the wheel diameter after the last turning of this axle and the average wear rate of the wheel diameter per 10,000 kilometers at the most recent turning. The more the wear amount, the shorter the remaining service life of the wheel set. Therefore, to study the evaluation of the remaining life of the wheel set, it is necessary to focus on studying the wear amount.

[0006] By analyzing the wear conditions of various parts of the wheel and considering relevant factors such as vehicle load, straight and curved sections, and different lines, the wear levels of four parts, namely the wheel tread wear amount, wheel flange wear amount, wheel inner diameter difference, and wheel hub difference, can be used to reflect the overall wear degree of the wheel, and a prediction model can be established. Then, based on the wear data and development trends of several parts of the wheel, and combined with the driving mileage of the wheel, the remaining life of the wheel can be evaluated and predicted. Currently, commonly used prediction algorithms include: model-based methods, data-driven methods, and fusion prediction methods, as shown in Figure 2 as follows.

[0007] Since the model-based method requires determining the distribution law, which is difficult to achieve in actual rail transit operations, and the model requires a failure physics model, which is generally a mechanism model obtained in an experimental environment and is greatly affected by various factors in actual operation, resulting in a significant reduction in the prediction effect of the final model.

[0008] From the perspective of mechanism research, the mechanism model requires precise mathematical formula principle proof. Under the current technical conditions, some models cannot be established or there is no breakthrough in mechanism research, leading to the monitoring of some fault-occurring objects.

[0009] Currently, the Prognostics and Health Management (PHM) system for EMUs has been connected to multiple vehicle types and opened for application to multiple railway bureaus. In recent years, it has prevented more than a hundred potential faults in EMU operation on average every year, playing an important role in the daily operation monitoring and fault emergency handling of EMUs.

[0010] During the popularization and application process of the PHM system, users have put forward higher requirements for the predictive maintenance of operation items in the actual application process of the PHM system. It is urgent to carry out research on the predictive maintenance of key components in operation and maintenance of the PHM system, put forward reasonable predictive maintenance suggestions, and guide the formulation of reasonable and economical maintenance plans and maintenance tasks.

[0011] Therefore, the technical problems solved by the present invention include:

[0012] How to analyze a large amount of data related to the wheel set, reduce the subjectivity of manual participation, and make the wheel set life prediction result more objective and accurate. Summary of the Invention

[0013] To solve the above problems, the present invention proposes a method for constructing a life prediction model of key components of a wheel set. The present invention constructs a life prediction model of key components of a wheel set based on a regression prediction algorithm of machine learning.

[0014] According to some embodiments, the present invention adopts the following technical solutions:

[0015] In a first aspect, a method for constructing a life prediction model for key components of a wheel set is disclosed, including:

[0016] Determine that the model input parameters are ambient temperature, driving speed, and driving mileage, and the output parameter is the remaining wear amount of the wheel diameter;

[0017] Obtain the current ambient temperature, driving speed, and driving mileage of the wheel set in real time, and obtain the turning data of the wheel set as sample data;

[0018] Preprocess the above sample data;

[0019] Construct models for the preprocessed sample data respectively, and finally select the model with the highest verification sample accuracy as the optimal prediction model for the current wheel set.

[0020] In a further technical solution, the turning data is stored in a database and regularly transmitted offline from each service station to the corresponding turning details table in the database.

[0021] In a further technical solution, the preprocessing of the sample data at least includes removing dirty data in the sample data, filling in missing data, and standardizing the data;

[0022] Preferably, remove outliers with certain errors or overlimits during the transmission of speed, temperature, and mileage, as well as incomplete turning data;

[0023] Preferably, fill in the turning data based on the difference algorithm according to the transmission frequencies of speed, temperature, and mileage;

[0024] Preferably, standardize the temperature, speed, mileage, and wheel diameter values so that the data is under a unified metric.

[0025] In a second aspect, a method for predicting the life of key components of a wheel set is disclosed, including:

[0026] Obtain the optimal prediction model of the wheel set based on the above method;

[0027] Obtain the driving speed, ambient temperature, and driving mileage of the wheel set in real time and preprocess them, select the corresponding prediction model, output the remaining wear amount of the wheel diameter, realize the real-time prediction of the remaining wear amount of the wheel diameter, and estimate the remaining life of the wheel set in real time based on the real-time predicted remaining wear amount of the wheel diameter.

[0028] In a further technical solution, after obtaining the optimal prediction model of the wheel set, solve the wheel diameter wear rate per 10,000 kilometers, then calculate the remaining operating kilometers and remaining operating days, and finally calculate the expiration date of the wheel diameter value.

[0029] For a further technical solution, when solving the wheel diameter wear rate per 10,000 km, the wheel diameter wear rate per 10,000 km in the most recent wheel turning = (the wheel diameter value after the previous wheel turning - the wheel diameter value before the most recent wheel turning) / the operating mileage within the most recent wheel turning cycle;

[0030] The above wheel diameter values all refer to (the left wheel diameter value + the right wheel diameter value) / 2 of the same wheel set;

[0031] Preferably further, obtain the wear rates for all wheel turning times of this wheel, find the maximum and minimum values of the wheel diameter wear rates for all wheel turning times, and finally obtain the maximum, minimum and the most recent average values of the wear rates.

[0032] For a further technical solution, when calculating the remaining operating kilometers and remaining operating days, the remaining operating kilometers (average / maximum / minimum) = the remaining wear amount of the wheel diameter after the most recent wheel turning of the axle / (the average wheel diameter wear rate per 10,000 km / the maximum wear rate / the minimum wear rate at the most recent wheel turning);

[0033] Preferably further, the remaining wear amount of the wheel diameter after the most recent wheel turning of the axle = the wheel diameter after the most recent wheel turning - the inner diameter of the wheel;

[0034] The average daily operating mileage within the most recent wheel turning cycle = (the cumulative running mileage of the vehicle at the most recent wheel turning - the cumulative running mileage of the vehicle at the previous wheel turning) / (the date of the most recent wheel turning - the date of the previous wheel turning);

[0035] Preferably further, the remaining operating time = the remaining operating kilometers / the average daily operating mileage within the most recent wheel turning cycle.

[0036] For a further technical solution, when calculating the due date of the wheel diameter value, the predicted due date of the wheel diameter value of this wheel = the most recent wheel turning time of this wheel + the remaining operating time of this wheel, and three due dates are obtained, namely the farthest due date, the nearest due date and the average due date.

[0037] In a third aspect, a wheel set key component life prediction system is disclosed, including:

[0038] A prediction model acquisition module, configured to: obtain the optimal prediction model of the wheel set based on the above method;

[0039] A remaining life real-time estimation module, configured to: obtain the wheel set driving speed, environmental temperature and driving mileage in real time and perform preprocessing, select the corresponding prediction model, output the remaining wear amount of the wheel diameter, realize the real-time prediction of the remaining wear amount of the wheel diameter, and perform real-time estimation of the remaining life of the wheel set based on the real-time predicted remaining wear amount of the wheel diameter.

[0040] Compared with the prior art, the beneficial effects of the present invention are:

[0041] Since the assessment of the remaining life of the wheel set is mainly estimated generally through the due date of the wheel diameter value. Therefore, the specific process of the assessment of the remaining life of the wheel set of the present invention is as follows: by predicting the wheel diameter value after turning, selecting relevant factors such as mileage, temperature, speed, etc., establishing a prediction model related to the wheel diameter value after turning, predicting the future wheel diameter value after turning, and calculating the nearest due date, average date, and farthest date of the wheel diameter value reaching the limit according to the wheel diameter value after turning, the average wear rate per 10,000 kilometers, average value, maximum value, and minimum value of all turning times, so as to realize the assessment of the remaining life of the wheel set.

[0042] The machine learning algorithm of the present invention uses a large amount of data for data analysis, mines its internal characteristic laws, can find the essential implicit laws from the data level, greatly reduces the subjectivity of manual participation, and the calculation results can be more objective and accurate.

[0043] According to the life prediction of the key components for operation and maintenance based on PHM of the present invention, reasonable predictive suggestions can be put forward to guide the formulation of reasonable and economical maintenance plans and maintenance tasks.

[0044] The advantages of the additional aspects of the present invention will be partially given in the following description, partially will become obvious from the following description, or will be understood through the practice of the present invention.

[0045] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following specific preferred embodiments are given in conjunction with the accompanying drawings and are described in detail as follows. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The accompanying drawings forming a part of the present invention are used to provide a further understanding of the present invention. The schematic embodiments and descriptions thereof of the present invention are used to explain the present invention and do not constitute an improper limitation to the present invention.

[0047] Figure 1 Schematic diagram of the influence of wear and turning on the tread of the EMU wheels

[0048] Figure 2 Route map for fault detection and life prediction;

[0049] Figure 3 Schematic diagram of the prediction technical solution;

[0050] Figure 4 Technical route map for the assessment of the remaining life of the wheel set in the embodiment of the present invention;

[0051] Figure 5 Calculation route map of the due date of the wheel diameter in the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0052] The present invention will be further described below in conjunction with the drawings and embodiments.

[0053] It should be noted that the following detailed description is illustrative and aims to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention pertains.

[0054] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0055] Example 1:

[0056] In this embodiment, the life prediction of the wheel set is taken as an example for illustration, but it does not mean that the solution provided by the present invention is only applicable to the life prediction of the wheel set. It can also be applicable to the life prediction of other objects according to different prediction objects.

[0057] In this embodiment, a method for constructing a life prediction model for key components of a wheel set is disclosed, including:

[0058] Step 1: Analyze the business scenario and determine the model input and output parameters.

[0059] The remaining wear amount of the wheel diameter is related to the inner diameter of the wheel and the remaining wheel diameter of the wheel. Since the inner diameter of the wheel is a fixed value, it is necessary to find the relevant quantity related to the remaining wheel diameter of the wheel. Since the wheel wear is caused by the continuous increase of the driving mileage, different driving speeds, and environmental temperatures, etc., a prediction model between speed, temperature, driving mileage and the remaining wear amount of the wheel diameter is established. According to the daily running situation of the vehicle, the monitoring of the remaining wear amount of the wheel every day can be realized. Therefore, the determined model input parameters are temperature, speed, and driving mileage, and the output parameter is the remaining wear amount of the wheel diameter.

[0060] Step 2: Data collection and extraction.

[0061] Since the current data such as temperature, speed, and mileage are stored in the big data platform Hbase and can be obtained in real time; however, the turning data is stored in the mysql database and is regularly transmitted offline from each service station to the corresponding turning details table in mysql. A large amount of modeling data needs to be extracted from different databases. And based on the modeling data, the modeling data is divided into a training set and a validation set according to the cross-validation method.

[0062] Step 3: Data preprocessing.

[0063] Data preprocessing mainly includes removing dirty data, filling in missing data, and standardizing data in the modeling data of each table, etc.

[0064] 3-1) Removal of dirty data:

[0065] There will be certain errors or outlier values beyond the limit during the transmission of speed, temperature, and mileage; for the turning data, due to offline upload, there may be incomplete uploads, or it may be considered that there are manual input errors, etc. These data need to be deleted, otherwise it will seriously affect the accuracy of modeling.

[0066] 3-2) Filling in missing data:

[0067] For the turning data which is uploaded offline and only has the wheel diameters before and after turning during turning, there will be a situation where the modeling data is seriously insufficient. Therefore, it is necessary to complete the filling based on the difference algorithm according to the transmission frequencies of speed, temperature, and mileage to construct sufficient modeling samples.

[0068] 3-3) Data standardization processing:

[0069] For parameters such as temperature, speed, mileage, and wheel diameter values, there are multiple orders of magnitude differences at the data level, which will cause optimization interference in the construction of the prediction model and make the model unable to find the optimal solution all the time. Therefore, it is necessary to preprocess each parameter so that the model parameters are under the same metric, for example, all standardized to [-1, 1].

[0070] Step Four: Model establishment and optimization.

[0071] The modeling of this prediction object is a many-to-one prediction form. Therefore, it is necessary to select a multi-dimensional prediction algorithm for model construction. However, for the actual situation of each wheel being different and the environmental impact also being different, the influence it receives is caused by a comprehensive combination of multiple aspects. Therefore, the model selects the idea of an ensemble algorithm, and the single prediction models to be selected are: random forest, multi-dimensional linear regression, neural network, etc. First, the models are constructed separately, and finally, the model with the highest accuracy of the validation samples is selected as the optimal prediction model for this wheel.

[0072] Step Five: Real-time model prediction.

[0073] By calling the prediction model of the corresponding wheel position through real-time speed, temperature, and driving mileage, the prediction of the remaining wear amount of the real-time wheel diameter is realized, and the remaining wear amount of the wheel diameter after the last turning of this wheel is obtained.

[0074] Example Two:

[0075] Based on the above method of Example One, a prediction model is obtained, and the life prediction of the key components of the wheel set is carried out based on the output result of the model, that is, the remaining wear amount of the real-time wheel diameter. See Appendix Figure 4 、 5As shown, the life prediction process includes:

[0076] (1) Solving the wheel diameter wear rate per 10,000 km:

[0077] The wheel diameter wear rate per 10,000 km for the most recent wheel turning = (the wheel diameter value after the last wheel turning - the wheel diameter value before the most recent wheel turning) / the operating mileage (in 10,000 km) within the most recent wheel turning cycle.

[0078] All the above wheel diameter values refer to (the left wheel diameter value + the right wheel diameter value) / 2 of the same wheel pair.

[0079] Furthermore, obtain the wear rates for all wheel turning times of this wheel, and find the maximum and minimum values of the wheel diameter wear rates for all wheel turning times. Finally, obtain the maximum, minimum, and average values of the most recent wear rate.

[0080] (2) Calculating the remaining operating kilometers and remaining operating days:

[0081] The remaining operating kilometers (average / maximum / minimum) = the remaining wear amount of the wheel diameter after the most recent wheel turning of this axle / (the average wheel diameter wear rate per 10,000 km / the maximum wear rate / the minimum wear rate at the most recent wheel turning), and the remaining wear amount of the wheel diameter after the most recent wheel turning of this axle = the wheel diameter after the most recent wheel turning - the inner diameter of the wheel.

[0082] The average daily operating mileage within the most recent wheel turning cycle = (the cumulative running mileage of the vehicle after the most recent wheel turning - the cumulative running mileage of the vehicle after the last wheel turning) / (the date of the most recent wheel turning - the date of the last wheel turning) (unit: days).

[0083] The remaining operating time (days) = the remaining operating kilometers / the average daily operating mileage within the most recent wheel turning cycle.

[0084] (3) Calculating the due date of the wheel diameter value:

[0085] The expected due date of the wheel diameter value of this wheel = the most recent wheel turning time of this wheel + the remaining operating time (days) of this wheel. Of course, three due date values will be obtained, namely the farthest due date, the nearest due date, and the average due date.

[0086] During the calculation process, the above data is regularly offline transmitted from each railway bureau service station to the big data platform and stored in the corresponding wheel turning details table in the mysql database. It is extracted from the wheel turning details table through a Python program, including: affiliated bureau, service station, vehicle type, train number, carriage number, axle position, wheel position, once the left and right wheel diameters after wheel turning, the left and right wheel diameters before the most recent wheel turning, the operating mileage within the most recent wheel turning cycle, etc.

[0087] From the perspective of the entire life cycle of wheel usage, in the initial stage of wheel usage, there is little concern about its expiration date. However, as the wheel usage time increases and the turning repair times increase, resulting in wheel wear and a decrease in wheel diameter, it becomes necessary to start paying attention to the expiration date of the wheel diameter value. During this period, the furthest expiration date can be selected as the key focus. As the usage cycle of the wheel diameter increases, the average expiration date is further concerned. In the later stage, the nearest expiration date needs to be focused on because there are significant safety hazards to the wheel operation at this time, which will affect the driving safety.

[0088] The present invention uses a machine learning algorithm to evaluate the remaining life of the wheel set. By analyzing relevant factors such as the wear condition of the wheel part, a prediction model for the wear amount of the wheel set is established. The real-time collected wheel set data is input into the prediction model for the wear amount of the wheel set to achieve real-time prediction of the wear amount of the wheel set, and the remaining life of the wheel is evaluated based on the prediction result of the wear amount and relevant data such as the driving mileage.

[0089] Embodiment 3:

[0090] With the great development of big data and artificial intelligence technologies, from the data perspective, a large amount of historical data is analyzed and mined to find the laws and implicit mechanism characteristics hidden behind the data, providing strong support for the safe and economic operation monitoring of rail transit EMUs, and providing a new solution idea for the fault early warning and remaining life assessment of the PHM intelligent operation and maintenance system.

[0091] This embodiment discloses a wheel set key component life prediction system, including:

[0092] A prediction model acquisition module, configured to: obtain the optimal prediction model of the wheel set based on the above method;

[0093] A remaining life real-time estimation module, configured to: obtain the driving speed, ambient temperature, and driving mileage of the wheel set in real time and perform preprocessing, select the corresponding prediction model, output the remaining wear amount of the wheel diameter, achieve real-time prediction of the remaining wear amount of the wheel diameter, and perform real-time estimation of the remaining life of the wheel set based on the real-time predicted remaining wear amount of the wheel diameter.

[0094] The present invention adopts a data-driven prediction method and integrates a mechanism model to achieve the assessment of the remaining life of the wheel set.

[0095] Since the assessment of the remaining life of the wheel set is mainly estimated by the expiration date of the wheel diameter value. Therefore, the specific process of the assessment of the remaining life of the wheel set is as follows: By predicting the turned wheel diameter value, selecting relevant factors such as mileage, temperature, and speed, establishing a prediction model for the turned wheel diameter value, predicting the future turned wheel diameter value, and calculating the nearest date, average date, and furthest date of the wheel diameter value reaching the limit based on the turned wheel diameter value, the average wear rate per 10,000 kilometers, average value, maximum value, and minimum value of all turning repair times; realizing the assessment of the remaining life of the wheel set.

[0096] Based on the last wheel turning diameter value, the due date or interval of the wheel diameter can be calculated; by increasing the mileage, the wheel turning diameter value of the next wheel turning cycle can be predicted, and further the due date or interval of the wheel diameter can be predicted.

[0097] The present invention uses machine learning to evaluate the remaining life of the wheel set. By analyzing relevant factors such as the wear conditions of various parts of the wheel, the wear levels of four parts, namely the tread wear amount of the wheel, the flange wear amount of the wheel, the inner diameter difference of the wheel, and the hub difference of the wheel, are used to reflect the overall loss degree of the wheel. A prediction model is established, and the remaining life of the wheel is evaluated and predicted by combining the wear data and development trend of the geometric parts of the wheel with the driving mileage; accurate prediction results can be obtained.

[0098] In some embodiments, refer to the attached Figure 3 As shown, the data-driven prediction method of the present invention, that is, the regression prediction algorithm based on machine learning. Currently, relatively popular machine learning prediction algorithms include linear regression, random forest, long short-term memory network (LSTM), etc.

[0099] In some embodiments, there can be multiple solutions to using machine learning algorithms to solve the predictive maintenance of key components in the EMU PHM. It can be comprehensively considered according to multiple factors such as the actual business scenario, data characteristics, and data volume, and the optimal machine learning algorithm can be selected to solve the problem of predictive maintenance of key components in the EMU PHM. Based on machine learning algorithms, it can also be integrated with the business mechanism model to jointly solve some complex business problems, and finally realize the fault warning and remaining life assessment of key components, etc., to provide auxiliary support for the safe and economic operation of the EMU.

[0100] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0101] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, and the combination of processes and / or blocks in the flowchart and / or block diagram, can be realized by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate for implementation in the processFigure 1 one process or multiple processes and / or blocks Figure 1 means for the functions specified in one block or multiple blocks.

[0102] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufactured article including an instruction means that implements the functions in the process Figure 1 one process or multiple processes and / or blocks Figure 1 specified in one block or multiple blocks.

[0103] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, whereby the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in the process Figure 1 one process or multiple processes and / or blocks Figure 1 specified in one block or multiple blocks.

[0104] The above are only the preferred embodiments of the present invention and are not used to limit the present invention. For those skilled in the art, various changes and modifications can be made to the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

[0105] Although the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, it is not a limitation on the protection scope of the present invention. Those skilled in the art should understand that, based on the technical solutions of the present invention, various modifications or deformations that can be made without creative efforts by those skilled in the art are still within the protection scope of the present invention.

Claims

1. A method for predicting the service life of key components of a wheel set, characterized in that, Including: Obtaining the optimal prediction model of the wheel set; Obtaining the running speed, ambient temperature and running mileage of the wheel set in real time, preprocessing them, inputting them into the optimal prediction model of the wheel set, outputting the remaining wear amount of the wheel diameter, realizing the real-time prediction of the remaining wear amount of the wheel diameter, and estimating the remaining life of the wheel set in real time based on the remaining wear amount of the wheel diameter predicted in real time; After obtaining the optimal prediction model of the wheel set, solving the wheel diameter wear rate per 10,000 km, then calculating the remaining running kilometers and remaining running days, and finally calculating the due date of the wheel diameter value; When solving the wheel diameter wear rate per 10,000 km, the wheel diameter wear rate per 10,000 km in the most recent wheel turning = (the wheel diameter value after the last wheel turning - the wheel diameter value before the most recent wheel turning) / the running mileage within the most recent wheel turning cycle; The above wheel diameter values all refer to (the left wheel diameter value + the right wheel diameter value) / 2 of the same wheel set; When calculating the remaining running kilometers and remaining running days, the average remaining running kilometers = the remaining wear amount of the wheel diameter after the most recent wheel turning of the axle / the average wheel diameter wear rate per 10,000 km in the most recent wheel turning, the maximum remaining running kilometers = the remaining wear amount of the wheel diameter after the most recent wheel turning of the axle / the minimum wheel diameter wear rate per 10,000 km in the most recent wheel turning, and the minimum remaining running kilometers = the remaining wear amount of the wheel diameter after the most recent wheel turning of the axle / the maximum wheel diameter wear rate per 10,000 km in the most recent wheel turning; The average daily running mileage within the most recent wheel turning cycle = (the cumulative running mileage of the vehicle in the most recent wheel turning - the cumulative running mileage of the vehicle in the last wheel turning) / (the date of the most recent wheel turning - the date of the last wheel turning); The remaining running time = the remaining running kilometers / the average daily running mileage within the most recent wheel turning cycle; When calculating the due date of the wheel diameter value, the expected due date of the wheel diameter value of this wheel = the most recent wheel turning time of this wheel + the remaining running time of this wheel, and three due date values are obtained, namely the farthest due date, the nearest due date and the average due date.

2. The method for predicting the life of key components of a wheel set according to claim 1, characterized in that The construction method of the optimal prediction model of the wheel set includes: Determining that the model input parameters are ambient temperature, running speed and running mileage, and the output parameter is the remaining wear amount of the wheel diameter; Obtaining the current ambient temperature, running speed and running mileage of the wheel set in real time, and obtaining the wheel turning data of the wheel set as sample data; Performing preprocessing on the above sample data; Constructing models for the preprocessed sample data respectively, and finally selecting the model with the highest verification sample accuracy as the optimal prediction model of the current wheel set.

3. The method for predicting the life of key components of a wheel set according to claim 2, characterized in that, The wheel turning data is stored in the database and is regularly transmitted back to the corresponding wheel turning details table in the database offline from each service station.

4. The life prediction method for key components of a wheel set according to claim 2, characterized in that, The preprocessing of the sample data at least includes removing dirty data in the sample data, filling in missing data, and standardizing the data; Removing outliers with errors or exceeding limits during the transmission of speed, temperature and mileage, as well as incomplete wheel turning data uploaded; Completing the wheel turning data based on the difference algorithm according to the transmission frequencies of speed, temperature and mileage; Standardizing the temperature, speed, mileage and wheel diameter value so that the data is under a unified metric.

5. The method for predicting the service life of key components of a wheel set according to claim 1, characterized in that, Obtain the wheel diameter wear rate for all wheel grinding times in this round, find the maximum and minimum values of the wheel diameter wear rates for all grinding times, and finally obtain the maximum, minimum, and the average value of the most recent wheel diameter wear rate.

Images