A method for predicting the quantity of spare parts for rail transit traction motors based on type III censored data

Abstract

The present invention discloses a method for predicting the quantity of spare parts for rail transit traction motors based on type III censored data, which can more accurately reflect the actual situation, improve the accuracy of spare parts quantity prediction, and enhance the security level of rail transit traction motors, and belongs to the technical field of traction motor spare parts quantity. The method has the following steps: step one, obtaining basic product data; step two, processing type III censored data; S21, establishing an initial data model; S22, inserting basic product information and time-series-based life cycle data into the initial data model through matching rules to obtain a final data model; S23, the final data model forms a type III censored data set through time-series-based cyclic calculation and type III censored data labeling; step three, fitting the life distribution curve; S31, fitting multiple statistical models using the type III censored data set; S32, obtaining a life distribution curve through a fitting method and a goodness of fit test method; and step four, spare parts prediction.

Description

Technical Field

[0001] The present invention belongs to the technical field of traction motor spare parts quantity, and in particular relates to a method for predicting the spare parts quantity of rail transit traction motors based on type III censored data. Background Art

[0002] my country has built the world's largest and most complex rail transit network. The rail transit system has become one of the indispensable infrastructures today. Its reliability and security directly affect the efficiency of passenger and freight transportation and the development of the industry, and have a significant impact on logistics and the quality of residents' travel. As one of the nine core components of rail transit, the traction motor is subjected to the test of complex operating environment, frequent start and stop, and high-speed operation, and its performance and reliability have a vital impact on the safety, stability and economy of the rail transit system. In order to ensure the normal operation of the rail transit system and reduce the downtime caused by equipment failure, in addition to ensuring the high reliability level of the traction motor, the requirements for its security level are also getting higher and higher. Rail transit companies need to reserve a certain number of traction motor spare parts online to achieve rapid recovery from faults. The rationality of spare parts types directly affects maintenance efficiency and operating costs. Excessive spare parts inventory will occupy a lot of funds and storage space, while insufficient spare parts may lead to untimely maintenance and affect the normal operation of vehicles. Therefore, scientifically and rationally predicting the demand for traction motor spare parts improves its security level and ensures the important link of efficient operation of the rail transit system.

[0003] The prediction of spare parts quantity for rail transit traction motors refers to the prediction of the number of spare parts required in a certain period of time in the future based on historical fault data, operation data and other relevant factors through certain methods and models to guide spare parts planning. Accurate spare parts quantity prediction can not only effectively reduce spare parts inventory costs, but also improve equipment maintenance efficiency and the availability level of rail transit systems. At present, common spare parts quantity prediction methods include: empirical judgment method, quantitative prediction method and qualitative prediction method. The empirical judgment method mainly relies on the empirical judgment of after-sales service engineers, and often has greater uncertainty for new products or lack of historical data. These spare parts quantity prediction methods have certain limitations when applied to rail transit traction motors. First, the operating environment of traction motors is complex and the failure modes are diverse, which makes it difficult for traditional quantitative prediction models to accurately capture the law of changes in their failure rates. Secondly, type III censored data is relatively common in rail transit systems, which poses a challenge to the integrity and continuity of the data, making it difficult for traditional prediction methods to be directly applied. In addition, the spare parts demand for traction motors is affected by many factors, including equipment aging, maintenance strategies, operating conditions, etc. The combined effect of these factors increases the complexity of prediction.

[0004] Figure 1The spare parts reserve plan for existing rail transit traction motors is based on the spare parts quantity of the previous year and the fluctuation of the online quantity of products this year. After-sales service engineers use the spare parts usage of each railway bureau / section last year as the input of the spare parts quantity of this year, and adjust the quantity based on personnel experience. The adjustment basis includes: online product quantity, batch preventive maintenance, spare parts type (easy-to-consume parts, parts that affect the mission, high-value parts, etc.) for classification decision-making.

[0005] The current spare parts plan based on the spare parts quantity of the previous year is inaccurate, has large quantity deviation fluctuations, and is not adjusted in time manually. This is mainly affected by the following factors: 1) Fluctuations in spare parts demand. Online faulty parts replacement, preventive maintenance, life cycle, warranty period, etc. of products will lead to fluctuations in spare parts demand. For example, new rail transit products are ordered year by year and continuously delivered online. The number of online products changes every day, and the demand for spare parts continues to fluctuate. When some products are out of warranty, users designate the company to repair online failures of some components, and the corresponding amount of spare parts needs to be reduced. Before or after preventive maintenance of the product, the fluctuation trend of the failure rate before and after maintenance needs to be considered. When the product is out of warranty, the company does not need spare parts when an online failure occurs. Therefore, only using the number of online products in the previous year as the basis for this year's spare parts will lead to a large deviation in spare parts demand.

[0006] 2) The accuracy and efficiency of manually sorting life cycle data are low. Through the collection and sorting of product operation data and life distribution fitting, it is found that the failure rate at the whole machine level and component level will change over time (non-constant failure rate), and there are differences between the life distribution and the exponential distribution. Especially when the reliability level of the component drops sharply (such as: the component enters the wear and tear failure period, batch failures caused by changes in the operating environment, etc.), the number of spare parts will fluctuate greatly compared with the previous year. The existing spare parts forecast based on the product life distribution curve requires manual sorting of product life cycle data, which has low analysis efficiency and is prone to errors. The company has many product models, and manual sorting work cannot support spare parts quantity forecasting.

[0007] 3) The deviation between the predicted value and the actual value of spare parts is not considered. If only historical data is used to predict the spare parts quantity, there will inevitably be a deviation between the predicted value and the actual value, and the longer the prediction time node is from the current time, the greater the prediction deviation may be.

[0008] Due to the influence of various factors, planning and adjusting the spare parts quantity only through personnel experience cannot support spare parts planning and management.

[0009] Inaccurate spare parts planning will lead to spare parts waste. Excessive online spare parts require storage, personnel and spare parts purchase costs. At the same time, inappropriate online component storage will lead to component life consumption, thereby reducing its online operating life and the reliability level of the whole machine; insufficient online spare parts will cause frequent spare parts dispatching and procurement activities, and the national railway transportation volume is large, and spare parts dispatching will generate a lot of dispatching and transportation costs and vehicle recovery time, which seriously affects the availability level of the whole vehicle. Summary of the invention

[0010] The purpose of the present invention is to provide a method for predicting the spare parts quantity of rail transit traction motors based on type III missing data, which can more accurately reflect the actual situation, improve the accuracy of spare parts quantity prediction, and enhance the security level of rail transit traction motors, which is of great significance for ensuring the efficient operation of trains.

[0011] The present invention is achieved by adopting the following technical solutions: A method for predicting the quantity of spare parts for rail transit traction motors based on type III censored data comprises the following steps: Step 1: Obtain basic product data Step 2: Type III censored data processing S21, classify spare parts demand according to the operation and maintenance characteristics of rail transit traction motors, that is, classify algorithm model strategies, and then establish an initial data model.

[0012] S22, inserting product basic information and time-series-based life cycle data into the initial data model through matching rules to obtain the final data model.

[0013] S23, the final data model forms a type III censored data set through time-series-based cyclic calculation and type III censored data marking.

[0014] Step 3: Fitting the life distribution curve S31, using the type III censored data set to fit a variety of statistical models.

[0015] S32, obtaining a life distribution curve through a fitting method and a goodness of fit test method.

[0016] Step 4: Spare parts forecasting S41, use the life distribution curve to calculate the spare parts prediction value and provide support to external business systems.

[0017] Further preferably, in step one, basic product data is obtained based on the enterprise digital platform, and the basic product data includes product shipment data, product failure data, basic product information, product repair and manufacturing data, product BOM data and data on necessary replacement parts for repair and manufacturing.

[0018] Further preferably, the spare parts demand classification in S21 includes faulty parts replacement, preventive maintenance, life cycle, and warranty period.

[0019] Further preferably, the enterprise digital platform is various business systems including TC, ERP, QMS, MRO, and big data platform.

[0020] Further preferably, the algorithm model strategy classification in S21 includes product level and component level, within warranty period and out of warranty period, and necessary replacement parts and non-necessary replacement parts.

[0021] Further preferably, step one also includes using big data technology to preliminarily integrate scattered data from decentralized business systems, and the scattered data is converted into basic product data after the preliminary integration.

[0022] Further preferably, at the product level, an initial data model is established with the whole machine as the smallest computing unit, and the initial data model is filled by matching product failure data and product shipment data from the perspective of the whole machine, and the filled final data model is arranged in time sequence and checked one by one; at the component level, an initial data model is established for each component of the product, and product failure data and product shipment data are matched from the perspective of component material coding, and the components are accurately matched according to the quantity of components, component numbers, and replacement part numbers to form a quality chain to fill the initial data model, and the filled final data models are arranged in time sequence and checked one by one.

[0023] Further preferably, the product shipment data includes product model, product number, repair level, repair sequence, and factory date; the product failure data includes life stage, discovery time, product model, product number, new factory date, failure location, failure number, replacement part number, and responsibility information; the product basic information includes daily operating hours, operating speed, annual operating days, maintenance cycle, maintenance interval, product series, and product model; the repair data includes product model, repair stage, and repair sequence; the product BOM data includes product model, material code, material number, and material name; the repair and replacement parts data includes product model, part code, repair level, and replacement part cycle.

[0024] Further preferably, step 2 also includes S20 before S21, product basic data processing, data cleaning and final integration of the product basic data, the final data integration logic is the series connection of life cycle data with product / component material code as the unique identifier and the association of similar data; the data cleaning logic is to eliminate duplicate data, fill in missing data, eliminate data noise, correct data format, correct data type, and standardize data format.

[0025] Further preferably, the process further includes step 5, external business system application S51, the external business system collects and feedbacks the actual number of faults.

[0026] S52, comparing the actual number of failures with the predicted value of spare parts to obtain a correction factor.

[0027] S53, the correction factor is used as a dynamic feedback parameter of the life distribution curve to dynamically adjust the life distribution curve and correct the spare parts prediction value.

[0028] S54, the external business system feeds back the collected real data to the product basic data and continues to update and iterate the life distribution curve.

[0029] Compared with the prior art, the present invention has the following beneficial effects: The present invention improves the efficiency and accuracy of data acquisition and ensures the comprehensiveness and efficiency of the acquired data by introducing type III censored data processing technology.

[0030] The present invention improves the accuracy and reliability of spare parts quantity prediction by establishing a data model and algorithm suitable for processing type III censored data.

[0031] The present invention enhances the application effect of prediction results in spare parts management by establishing a data feedback mechanism and a method for dynamically adjusting the model, ensuring that the prediction results can guide spare parts procurement and inventory management in real time.

[0032] The present invention establishes a comprehensive prediction model by combining the maintenance strategy of the rail transit system to ensure that the prediction results are highly consistent with the actual maintenance needs.

[0033] Through the above technical means, the present invention improves the scientificity and practicality of the prediction of spare parts quantity for rail transit traction motors, enhances the accuracy and efficiency of spare parts management, thereby effectively reducing the operational risks and costs caused by insufficient or excessive spare parts, and promoting the green and sustainable development of the rail transit industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0035] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0036] Figure 1 The present invention shows the spare parts reserve plan of the rail transit traction motor in the prior art.

[0037] Figure 2A schematic diagram showing the overall flow chart of the present invention. DETAILED DESCRIPTION

[0038] In order to more clearly understand the above-mentioned objectives, features and advantages of the present invention, the scheme of the present invention will be further described below. It should be noted that the embodiments of the present invention and the features in the embodiments can be combined with each other without conflict.

[0039] In the description, it should be noted that the terms "first" and "second" are only used for descriptive purposes and should not be understood as indicating or implying relative importance. It should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected" and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be internal communication between two elements. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to specific circumstances.

[0040] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein; it is obvious that the embodiments in the specification are only part of the embodiments of the present invention, rather than all of the embodiments.

[0041] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0042] Embodiment 1, a method for predicting the quantity of spare parts for rail transit traction motors based on type III censored data, step 1, obtaining basic product data; The method for acquiring the basic product data is to acquire product shipment data, product failure data, product basic information, product repair and manufacturing data, product BOM data and repair and manufacturing necessary replacement parts data based on the enterprise digital platform.

[0043] The enterprise digital platform includes business systems such as TC, ERP, QMS, MRO, and big data platform.

[0044] The product shipment data includes product model, product number, repair level, repair sequence, and factory date; the product failure data includes life stage, discovery time, product model, product number, new factory date, failure location, failure number, replacement part number, and responsibility information; the product basic information includes daily operating hours, operating speed, annual operating days, maintenance cycle, maintenance interval, product series, and product model; the repair data includes product model, repair stage, and repair sequence; the product BOM data includes product model, material code, material quantity, and material name; the repair and replacement parts data includes product model, part code, repair level, and replacement part cycle.

[0045] Step 2: Type III censored data processing; S20, product basic data processing, data cleaning and final integration of product basic data.

[0046] The final data integration logic is to concatenate life cycle data and associate similar data using product / component material codes as unique identifiers.

[0047] The data cleaning logic is to remove duplicate data, fill in missing data, eliminate data noise, correct data format, correct data type, and standardize data format.

[0048] S21, classify spare parts demand according to the operation and maintenance characteristics of rail transit traction motors, that is, classify algorithm model strategies, and then establish an initial data model.

[0049] The spare parts demand classification includes faulty parts replacement, preventive maintenance, life cycle, and warranty period.

[0050] The algorithm model strategy classification includes product level and component level, within the warranty period and out of the warranty period, and necessary replacement parts and non-necessary replacement parts.

[0051] The initial data model includes a fault information storage linked list, a replacement parts storage list, a maintenance time node time series, etc.

[0052] S22, inserting product basic information and time-series-based life cycle data into the initial data model through matching rules to obtain the final data model.

[0053] At the product level, an initial data model is established with the entire machine as the smallest computing unit. The initial data model is filled with product failure data and product shipment data from the perspective of the entire machine, and the filled final data model is arranged in chronological order and checked one by one. At the component level, an initial data model is established for each component of the product. Product failure data and product shipment data are matched from the perspective of component material coding. Components are accurately matched according to component quantity, component number, and replacement part number to form a quality chain to fill the initial data model, and the filled final data models are arranged in chronological order and checked one by one.

[0054] S23, the final data model forms a type III censored data set through time-series-based cyclic calculation and type III censored data marking.

[0055] Working principle: Based on the operation and maintenance characteristics of rail transit motor products, algorithm model strategy classification is carried out, data model is established and matching rules are imported; the output forms a type III missing data set that is highly consistent with the actual operating conditions of rail transit traction motors.

[0056] Step 3, life distribution curve fitting; S31, using the type III censored data set to fit a variety of statistical models.

[0057] The life distribution curves of component products and parts in the Type III censored data set are used to fit a variety of statistical models. During the life distribution fitting process, statistical methods such as maximum likelihood estimation (MLE) are used to estimate the parameters. Through stepwise optimization and goodness of fit test methods, a life distribution curve with a high degree of fit is obtained.

[0058] Various statistical models include exponential distribution, Weibull distribution, and lognormal distribution.

[0059] S32, obtaining a life distribution curve through a fitting method and a goodness of fit test method.

[0060] Types of life distribution curves include PDF, CDF, HF, SF, and CHF.

[0061] Working principle: The Type III censored data set is further used to fit the product life distribution curve. The fitting process comprehensively considers the failure mode, maintenance cycle, use environment and other complex factors of rail transit traction motors (represented by the parameters input in the statistical model), and finally generates various life distribution curves that can accurately reflect the actual performance of the product. These life distribution curves not only provide a reliable basis for the prediction of future spare parts demand, but also support the in-depth application of life analysis and risk assessment.

[0062] Step 4: Spare parts forecast; S41, use the life distribution curve to calculate the spare parts prediction value and provide support to external business systems.

[0063] In the second embodiment, the product basic data is collected and stored by the enterprise's various business platforms (TC, ERP, QMS, MRO, big data platform and other business systems) for the original data generated during the entire life cycle of each product / component. For these scattered data, the software interface development technology between systems is used, and the big data platform is used as the data middle platform to collect and initially integrate the data of each business system, as the product basic data for the prediction of the spare parts quantity of type III missing data. For these scattered data, the software interface development technology between systems is used, and the big data platform is used as the data middle platform to collect and integrate the data of each business system.

[0064] In the third embodiment, if the basic data of the product is missing, redundant, heterogeneous, non-standard, or biased, it will affect the quality and accuracy of the data model calculation. Therefore, the basic data can be cleaned first, that is, data verification, missing data filling, duplicate data removal, and data noise filtering can be performed to obtain a reliable data set. The cleaned data is still stored in an independent data set to reduce the coupling of subsequent algorithm data structure filling. According to the characteristics of the spare parts demand for rail transit traction motors, the algorithm model strategies are classified from different dimensions of spare parts demand, such as replacement of faulty parts, preventive maintenance, life cycle, and warranty period, and are divided into product level (whole machine level) and component level. By classifying the algorithm model strategies and filling the initial data model, data models of different demand dimensions can be obtained, and then type III censored data sets of different demand dimensions can be obtained, which provide support for subsequent life distribution curve fitting and spare parts prediction.

[0065] Life distribution curve fitting is a key link in product reliability analysis, which is directly related to the accuracy and applicability of spare parts prediction. Using the aforementioned type III censored data set, the optimal life distribution curve of the product or component is constructed through fitting methods, goodness of fit tests, and other means. Specifically, the following steps are used to fit the life distribution curve: the type III censored data set obtained after type III censored data processing is fitted for a variety of statistical models (including but not limited to exponential distribution, Weibull distribution, lognormal distribution, etc.). In the process of life distribution curve fitting, statistical methods such as maximum likelihood estimation (MLE) are used to estimate parameters. Through step-by-step optimization and goodness of fit test methods (BIC, AICc, etc.), a life distribution curve with a high degree of fit is obtained.

[0066] Embodiment 5, based on embodiment 1, step 5 is added, external business system application, Figure 2 is a flow chart of this embodiment.

[0067] S51, the external business system collects and feedbacks the actual number of faults.

[0068] S52, comparing the actual number of failures with the predicted value of spare parts to obtain a correction factor.

[0069] S53, the correction factor is used as a dynamic feedback parameter of the life distribution curve to dynamically adjust the life distribution curve and correct the spare parts prediction value.

[0070] S54, the external business system feeds back the collected real data to the product basic data and continues to update and iterate the life distribution curve.

[0071] Working principle: Based on the enterprise digital platform, obtain basic product data, clean the basic product data required by the algorithm, eliminate redundant data, supplement missing data, etc.

[0072] Then, according to the operation and maintenance characteristics of rail transit traction motors, the spare parts demand dimension is carried out, and the corresponding algorithm strategy is selected to establish the initial data model. The basic product information and time-series-based life cycle data are inserted into the initial data model through matching rules. Time-series-based cyclic calculation and type III missing data marking are carried out to form a type III missing data set.

[0073] Then, the type III censored data set is used to perform data fitting to form the product’s PDF, CDF, SF and other life distribution curves.

[0074] The life distribution curve is used to calculate the spare parts quantity for future time periods and provide external system support. Finally, the external business system is waited for to collect and feedback the actual number of failures, which is compared with the predicted value to form a correction factor as the dynamic feedback parameter of the life distribution curve. The life distribution curve is dynamically adjusted to correct the spare parts prediction value of subsequent products. At the same time, the subsequent product basic data will also update the iterative life characteristic curve to achieve accurate data acquisition, efficient collaborative feedback and dynamic update and iteration of rail transit motor spare parts prediction.

[0075] The advantage of adding step five is to continuously monitor the difference between the actual number of failures and the predicted value of spare parts. Through the dynamic correction factor, it is fed back to the spare parts prediction model (i.e., the life distribution curve) to achieve iterative optimization and continuous improvement of spare parts prediction, thereby improving the accuracy of spare parts prediction value and the actual application effect. In summary, this method not only improves the accuracy of rail transit traction motor spare parts quantity prediction, but also realizes efficient control of spare parts management through dynamic data update and model iteration, ensuring the safe and stable operation of the rail transit system.

[0076] In the sixth embodiment, the fitted life distribution curve can be used to evaluate the remaining life of the rail transit traction motor or key components and predict the probability of future failure. This process not only provides a reliable basis for the accurate prediction of the spare parts quantity, but also provides scientific guidance for the preventive maintenance of the equipment.

[0077] By analyzing the morphology and parameter characteristics of the life distribution curve, the failure trend of products or parts at different operating stages can be identified, thereby optimizing the spare parts procurement and reserve strategy. Specifically, spare parts prediction and feedback are achieved through the following steps: the initial basic prediction value of spare parts quantity is obtained through the life distribution curve, the actual spare parts usage (i.e., the actual number of failures) is obtained within a certain time range, and the correction factor is formed as the dynamic feedback parameter of the prediction model (i.e., the life distribution curve) by comparing it with the predicted spare parts quantity. The prediction model is dynamically adjusted, and the actual failure data will be stored in the database of each business system. In the subsequent spare parts prediction process, the type III censored data set will continue to be affected, thereby dynamically updating the life distribution curve and spare parts prediction value. The corrected spare parts prediction value is used to guide the actual business. Through this continuous feedback of the actual number of failures and adjustment of spare parts prediction values, the prediction accuracy of spare parts prediction values ​​can be further improved.

[0078] The method for predicting the quantity of spare parts for rail transit traction motors based on type III censored data proposed in the present invention has the following advantages: 1. Improve data processing efficiency and quality: By systematically processing the data of the entire life cycle of rail transit traction motors and cleaning and integrating the basic data of various products, the redundancy and irregularities in the traditional data processing process are eliminated, thereby significantly improving the efficiency and accuracy of data processing.

[0079] 2. Enhance the adaptability and accuracy of the prediction model: This method takes into account complex factors such as replacement of faulty parts, preventive maintenance, life cycle, warranty period, etc. of rail transit traction motors, so that the prediction model can more accurately reflect the actual usage and failure trends of the product, significantly improving the accuracy of spare parts quantity prediction.

[0080] 3. Realize dynamic feedback and model iteration: The present invention realizes continuous optimization of the prediction model by introducing a dynamic feedback mechanism between the prediction results and the actual operation data. By comparing the actual fault data with the predicted value, a correction factor is formed, and the model is continuously corrected and iterated, so that the spare parts prediction results can better meet the actual needs and reduce the risk of excess or shortage of spare parts.

[0081] 4. Improve the safety and reliability of rail transit systems: Based on the fitted life distribution curve, the present invention can accurately predict the future failure probability of rail transit traction motors or their key components, and support the implementation of preventive maintenance strategies. Through accurate prediction of spare parts quantity, the risk of equipment downtime caused by insufficient spare parts supply can be effectively reduced, ensuring the safe and stable operation of the rail transit system.

[0082] 5. Promote enterprise digital management and resource optimization: This invention combines the enterprise digital platform and big data technology to integrate spare parts quantity prediction into the overall management system of the enterprise, and realizes the data connection and optimized configuration of various business systems. Through a scientific and reasonable spare parts reserve strategy, it not only reduces inventory costs, but also improves resource utilization efficiency, providing strong support for the sustainable development of the enterprise.

[0083] The present invention is used to solve the technical problems that the existing technical solutions lack the ability to efficiently acquire and process data, lack the ability to accurately evaluate complex scenarios such as the operating environment, failure mode, and maintenance cycle of rail transit traction motors, and lack the linkage and feedback mechanism between the prediction results and the prediction process. The specific performance is as follows: 1) Data acquisition is time-consuming and has low accuracy and utilization. Traditional methods often require a lot of manual operations during data acquisition, and the accuracy and utilization of data are difficult to guarantee.

[0084] 2) Insufficient adaptability of prediction models: Existing prediction models are difficult to accurately capture the complex failure modes and operating environments of rail transit traction motors, resulting in inaccurate prediction results.

[0085] 3) Insufficient application of prediction results: Traditional methods often lack an effective feedback mechanism when applying prediction results, resulting in poor results in practical applications.

[0086] 4) Insufficient coordination between maintenance strategies and prediction results: Existing methods often ignore the impact of maintenance strategies on spare parts demand, resulting in a disconnect between prediction results and actual demand.

[0087] In summary, the present invention not only overcomes the shortcomings of the prior art such as insufficient data processing, model adaptability and lack of dynamic feedback mechanism, but also has significantly improved the accuracy of spare parts quantity prediction, the practicality of the model and the reliability of the system, and can effectively support the full life cycle management of rail transit traction motors.

[0088] The above is only a specific implementation of the present invention, which enables those skilled in the art to understand or implement the present invention. Although detailed descriptions are given with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments, and they should all be covered by the protection scope of the claims.

Claims

1. A method for predicting the quantity of spare parts for rail transit traction motors based on type III censored data, characterized in that: The steps are as follows: Step 1: Obtain basic product data Step 2: Type III censored data processing S21, classify spare parts demand according to the operation and maintenance characteristics of rail transit traction motors, that is, classify algorithm model strategies, and then establish an initial data model. S22, inserting product basic information and time-series-based life cycle data into the initial data model through matching rules to obtain the final data model, S23, the final data model forms a type III censored data set through time-series-based cyclic calculation and type III censored data marking; Step 3: Fitting the life distribution curve S31, using type III censored data sets to fit a variety of statistical models, S32, obtaining a life distribution curve by a fitting method and a goodness of fit test method; Step 4: Spare parts forecasting S41, use the life distribution curve to calculate the spare parts prediction value and provide support to external business systems.

2. A method for predicting the quantity of spare parts for rail transit traction motors based on type III censored data according to claim 1, characterized in that: In step one, based on the enterprise digital platform, basic product data is obtained, wherein the basic product data includes product shipment data, product failure data, basic product information, product repair and manufacturing data, product BOM data and data on necessary replacement parts for repair and manufacturing.

3. A method for predicting the quantity of spare parts for rail transit traction motors based on type III censored data according to claim 1, characterized in that: The spare parts demand classification described in S21 includes faulty parts replacement, preventive maintenance, life cycle, and warranty period.

4. A method for predicting the quantity of spare parts for rail transit traction motors based on type III censored data according to claim 2, characterized in that: The enterprise digital platform includes various business systems such as TC, ERP, QMS, MRO, and big data platform.

5. A method for predicting the quantity of spare parts for rail transit traction motors based on type III censored data according to claim 2, characterized in that: The algorithm model strategy classification in S21 includes product level and component level, within the warranty period and out of the warranty period, and mandatory replacement parts and non-mandatory replacement parts.

6. A method for predicting the quantity of spare parts for rail transit traction motors based on type III censored data according to claim 4, characterized in that: Step one also includes using big data technology to initially integrate scattered data from decentralized business systems, and then converting the scattered data into basic product data after initial integration.

7. A method for predicting the quantity of spare parts for rail transit traction motors based on type III censored data according to claim 5, characterized in that: At the product level, an initial data model is established with the entire machine as the smallest computing unit. The initial data model is filled with product failure data and product shipment data from the perspective of the entire machine, and the filled final data model is arranged in chronological order and checked one by one. At the component level, an initial data model is established for each component of the product. Product failure data and product shipment data are matched from the perspective of component material coding. Components are accurately matched according to component quantity, component number, and replacement part number to form a quality chain to fill the initial data model, and the filled final data models are arranged in chronological order and checked one by one.

8. A method for predicting the quantity of spare parts for rail transit traction motors based on type III censored data according to claim 7, characterized in that: The product shipment data includes product model, product number, repair level, repair sequence, and factory date; the product failure data includes life stage, discovery time, product model, product number, new factory date, failure location, failure number, replacement part number, and responsibility information; the product basic information includes daily operating hours, operating speed, annual operating days, maintenance cycle, maintenance interval, product series, and product model; the repair data includes product model, repair stage, and repair sequence; the product BOM data includes product model, material code, material quantity, and material name; the repair and replacement parts data includes product model, part code, repair level, and replacement part cycle.

9. A method for predicting the quantity of spare parts for rail transit traction motors based on type III censored data according to any one of claims 1 to 8, characterized in that: Step 2 also includes S20, product basic data processing, data cleaning and final integration of product basic data before S21. The final data integration logic is to connect the life cycle data with the product / component material code as the unique identifier and associate the same type of data; The data cleaning logic is to remove duplicate data, fill in missing data, eliminate data noise, correct data format, correct data type, and standardize data format.

10. A method for predicting the quantity of spare parts for rail transit traction motors based on type III censored data according to any one of claims 1 to 8, characterized in that: Also includes step 5, external business system application S51, the external business system collects and feedbacks the actual number of faults, S52, the actual number of failures is compared with the predicted value of spare parts to obtain the correction factor, S53, the correction factor is used as a dynamic feedback parameter of the life distribution curve to dynamically adjust the life distribution curve and correct the spare parts prediction value. S54, the external business system feeds back the collected real data to the product basic data and continues to update and iterate the life distribution curve.

Images