A method for determining the mean mileage between failures of a vehicle

Abstract

This invention provides a method for determining the mean mileage between failures (MTBF) of a vehicle. This method, designed from a forward development perspective at the initial stage of a project, combines historical data from previous vehicle models and user complaint cycles to create a simple and practical method for determining the mean mileage between failures (MTBF). This method has significant practical significance in automotive product development. The method includes the following steps: A. establishing a service information system for currently available vehicle models; B. calculating the average daily mileage for a user group with a predetermined percentage, Q, under a normal distribution; C. quantifying and estimating the user complaint cycle; and D. determining the mean mileage between failures for the vehicle.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle reliability development, and in particular to a method for determining the mean mileage between failures of a vehicle. Background Art

[0002] Reliability is a dimension of vehicle performance, and the vehicle reliability development process requires goal setting and achievement assessment. Quantifying reliability targets typically involves numerous indicators, such as the number of equivalent failures, equivalent failure rate, mean mileage to first failure (MTTF), mean mileage between failures (MTBF), repair time per thousand kilometers, and repair cost per thousand kilometers. MTBF is a particularly important metric. Based on the vehicle's MTBF10, it can be broken down into system and component reliability, thereby initiating a "V"-shaped performance development and verification process.

[0003] GB / T12678-1990, "Automobile Reliability Driving Test Methods," provides a point estimate formula for MTBF based on an exponential distribution. However, this method is more suitable for estimating results by planning specific reliability tests and conducting statistical analysis (classification and grading) of test process issues. For data with small sample sizes, the estimated results are often inaccurate. Summary of the Invention

[0004] The present invention aims to provide a method for determining the mean mileage between failures (MTBF) of a vehicle. This method, designed from a forward development perspective at the initial stages of a project, combines historical data from previous vehicle models and user complaint cycles to create a simple and practical method for determining the mean mileage between failures (MTBF). This method has significant practical significance in automotive product development. The method includes the following steps:

[0005] A. Build a service information system for all vehicles currently on sale: record identification information, fault information, and mileage information for each vehicle sold;

[0006] B. Calculate the average daily mileage of a user group with a preset proportion of Q under normal distribution: Based on the service information system in step A, retrieve and filter the fault information records. Fit the normal distribution function on the filtered samples to estimate the average daily mileage S of the user group with a preset proportion of Q. up ;

[0007] C. Quantify and estimate the user complaint cycle: Assume that the user complaint cycle is T, the vehicle warranty period is X years or Y kilometers, and use the MAX function to estimate the complaint cycle of the user group with a proportion of Q.

[0008]

[0009] D. Determine the average mileage between failures of the vehicle: Average mileage between failures of the vehicle

[0010]

[0011] The steps of fitting and estimating the normal distribution function in step B are as follows:

[0012] ①、Calculate basic data: Set the total number of vehicles after screening as n, the sample as S, and the vehicles are arranged in order according to the sales date and marked as L1, L2, L3, ..., L n , the driving days are recorded as D1, D2, D3, ..., D n The fault mileage is recorded as R1, R2, R3, ..., R n , calculate the basic sample through the ROUND function

[0013]

[0014] ② The basic sample is fitted with the general normal distribution function to estimate the average daily mileage S of the user group with a proportion of Q up ; The general normal distribution function in step ② is S~N(μ, σ 2 ), the calculation method is: calculate the mean μ through the mean function; calculate the standard deviation σ through the standard deviation function, and set the maximum value of the average daily mileage calculation to S max , the basic sample is divided into intervals of (0, ROUNDUP(Smax)] every 5 km, and the total number of intervals after division is recorded as k; the sequence (0, 5]; (5, 10]; (10, 15]; ...; (a, a+5]; ...;

[0015]

[0016] Calculate the number of vehicles in the corresponding interval and record them as n1, n2, n3, ..., n k ;

[0017] The calculation formula for the proportion of vehicles in each section is:

[0018]

[0019] The formula for calculating the cumulative proportion of vehicles in each section is:

[0020]

[0021] The formula for calculating the average daily mileage of the user group with a cumulative proportion of Q is:

[0022] S up =NORMINV(Q,μ,σ).

[0023] Furthermore, the step ② further includes a Pearson test step for the basic sample: according to the idea of ​​non-parametric hypothesis testing and the Pearson-Fisher theorem, the null hypothesis is: H0: S~N(μ, σ 2 ); the constructed normal distribution probability density and its distribution function are:

[0024]

[0025]

[0026] According to the 2 distribution table, the corresponding degrees of freedom are k-2-1, and the significance level is specified (usually 0.05), the rejection region is obtained:

[0027]

[0028] The Pearson 2 statistic is:

[0029]

[0030] If χ 2 If the statistic is not in the rejection region, it means that the original hypothesis is acceptable and the sample obeys the normal distribution; otherwise, the sample does not obey the normal distribution; the mean μ and standard deviation σ are calculated by the maximum likelihood estimation method.

[0031] Furthermore, for samples that do not obey the normal distribution, statistical software tools are used to reselect new fitting functions.

[0032] Furthermore, the screening of fault information records in step B is as follows: among the fault information records of the same vehicle, only the fault information record with the longest maintenance mileage is retained.

[0033] Furthermore, the identification information includes vehicle model, production date, sales information, and vehicle identification number, and the fault information includes repair date, damaged part name, and damaged part number.

[0034] Furthermore, the fault information also includes the service station name, fault description, troubleshooting process, system classification, amount involved, and maintenance personnel.

[0035] By constructing a service information system for models currently on sale, the present invention can combine historical data of past models and user complaint cycles, perform fitting according to a normal distribution function, calculate the average daily mileage of users of a certain percentage, quantify and estimate the user complaint cycle, and finally, combine the average daily mileage of users and the user complaint cycle to easily determine the average mileage between failures of the entire vehicle. After determining the average mileage between failures of the entire vehicle, the reliability of the entire vehicle can be decomposed according to the Weibull distribution function, thereby deriving the reliability target requirements of the subdivided systems and components. Reliability design can then be carried out, and through reliability verification at each level from components to systems to the entire vehicle, combined with the ideas of reliability engineering management, repeated iterations, rectification and optimization can be carried out to achieve a "V"-shaped reliability positive development. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 Flowchart of the present invention.

[0037] Figure 2 The figure shows the distribution of the number of vehicles and the cumulative percentage of vehicles versus the average daily mileage. DETAILED DESCRIPTION

[0038] The following describes an embodiment of the present invention with reference to the accompanying drawings, and further describes in detail the specific embodiments of the present invention, such as the shape, structure, relative positions and connection relationships between the various components involved, the functions and working principles of the various components.

[0039] the following Figure 1 The steps of the present invention are further described with reference to an example of a certain type of automobile product development process:

[0040] 1. Build a service information system for models currently on sale

[0041] First, we built a service information system for vehicles currently on sale to ensure that every vehicle sold at the terminal is recorded. This system allows us to query the following information, as shown in Table 1.

[0042]

[0043] Table 1 Service information of models on sale

[0044] 2. Calculate the average daily mileage of users under normal distribution

[0045] Secondly, extract the information from Table 1 and add a column for "Daily Average Mileage Calculation". At the same time, filter out vehicles that meet the requirements. The screening principles are as follows:

[0046] 1) If the same vehicle is repaired multiple times, the most recent repair record (i.e. the record with the longest mileage) will be retained;

[0047] 2) Eliminate invalid maintenance records, such as sales date earlier than production date, abnormal mileage (average daily mileage of 0 or more than 1000km), and other obvious errors.

[0048] In this example, the actual number of original data retrieved is 12356. After screening, the valid data is 10492; n = 10492; according to the vehicle sales date, they are numbered L1, L2, L3, ..., L 10492 .

[0049] By using the maximum likelihood estimation method, we can calculate the mean μ = 32.98 km and the standard deviation σ = 15.85 km. The maximum value of the average daily mileage is S max =172.14km.

[0050] The above values ​​are divided into intervals every 5 km: (0, 5]; (5, 10]; (10, 15]; ...; (165, 170]; (170, 173]; There are 35 sections in total. The number of vehicles in each section and their cumulative proportion and average daily mileage distribution are as follows: Figure 2 shown.

[0051] Fit the normal distribution function and calculate the upper limit S of the daily average mileage of the user group with a proportion of Q up ;

[0052] In this embodiment, Q is selected as 90%; S up =NORMINV(0.9,μ,σ)=53.3km; that is, 90% of the user group has an average daily mileage of less than 53.3km.

[0053] More rigorously, conduct hypothesis testing: the theoretical frequency np i Intervals smaller than 5 were merged and integrated into 18 intervals;

[0054] According to χ 2 Distribution table, corresponding to 15 degrees of freedom, specifying the significance level α (usually 0.05), and obtaining the rejection region

[0055]

[0056] Pearson χ 2 The statistics are:

[0057]

[0058] χ 2 If the statistic is not within the rejection region, the null hypothesis is acceptable and the sample follows a normal distribution. For samples that do not follow a normal distribution, statistical software tools can be used to reselect a new fitting function to make the estimation results more accurate.

[0059] 3. Quantify and estimate user complaint cycles

[0060] Secondly, after comparing competing vehicles and conducting market research, the warranty for the example model is set at 3 years or 100,000 kilometers (whichever comes first). The acceptable driving cycle for the user group, which accounts for Q, is T days.

[0061]

[0062] 4. Determine the vehicle's mean mileage between failures. Finally, the vehicle's mean mileage between failures (MTBF) can be estimated using the formula:

[0063]

[0064] In actual applications, the mean mileage between failures (MTBF) of the vehicle is set to 52,500 km, which is used to calculate vehicle reliability, decompose system reliability, and set component reliability. Component tests, system-level tests, and vehicle-level tests are then conducted to verify the achievement of the goals, and finally the results are evaluated, thus forming a complete "V"-shaped development process.

[0065] The present invention is described above by way of example in conjunction with the accompanying drawings. It is obvious that the specific design of the present invention is not limited to the above-mentioned manner. As long as various non-substantial improvements are made using the concepts and technical solutions of the present invention, or the concepts and technical solutions of the present invention are directly applied to other occasions without improvement, they are all within the scope of protection of the present invention.

Claims

1. A method for determining the mean mileage between failures of a vehicle, characterized in that The following steps are involved: A. Build a service information system for all vehicles currently on sale: record identification information, fault information, and mileage information for each vehicle sold; B. Calculate the average daily mileage of a user group with a preset proportion of Q under normal distribution: Based on the service information system in step A, retrieve and filter the fault information records. Fit the normal distribution function on the filtered samples to estimate the average daily mileage S of the user group with a preset proportion of Q. up ; C. Quantify and estimate the user complaint cycle: Assume that the user complaint cycle is T, the vehicle warranty period is X years or Y kilometers, and use the MAX function to estimate the complaint cycle of the user group with a proportion of Q. D. Determine the average mileage between failures of the vehicle: Average mileage between failures of the vehicle The steps of fitting and estimating the normal distribution function in step B are as follows: ①、Calculate basic data: Set the total number of vehicles after screening as n, the sample as S, and the vehicles are arranged in order according to the sales date and marked as L1, L2, L3, ..., L n , the driving days are recorded as D1, D2, D3, ..., D n The fault mileage is recorded as R1, R2, R3, ..., R n , calculate the basic sample through the ROUND function ② The basic sample is fitted with the general normal distribution function to estimate the average daily mileage S of the user group with a proportion of Q up ; The general normal distribution function in step ② is S~N(μ, σ 2 ), the calculation method is: calculate the mean μ through the mean function; calculate the standard deviation σ through the standard deviation function, and set the maximum value of the daily average mileage calculation to S max , the basic sample is divided into intervals of (0, ROUNDUP(Smax)] every 5 km, and the total number of intervals after division is recorded as k; the sequence (0, 5]; (5, 10]; (10, 15]; ...; (a, a+5]; ...; Calculate the number of vehicles in the corresponding interval and record them as n1, n2, n3, ..., n k ; The calculation formula for the proportion of vehicles in each section is: The formula for calculating the cumulative proportion of vehicles in each section is: The formula for calculating the average daily mileage of a user group with a cumulative proportion of Q is S up =NORMINV(Q,μ,σ).

2. The method for determining the mean mileage between failures of a vehicle according to claim 1, characterized in that The step ② also includes a Pearson test step for the basic sample: according to the idea of ​​non-parametric hypothesis testing and the Pearson-Fisher theorem, the null hypothesis is: H0: S~N(μ, σ 2 ); the constructed normal distribution probability density and its distribution function are: According to χ 2 Distribution table, corresponding to k-2-1 degrees of freedom, specifying the significance level α, and obtaining the rejection region: Pearson χ 2 The statistics are: If χ 2 If the statistic is not in the rejection region, it means that the original hypothesis is acceptable and the sample obeys the normal distribution; otherwise, the sample does not obey the normal distribution; the mean μ and standard deviation σ are calculated by the maximum likelihood estimation method.

3. The method for determining the mean mileage between failures of a vehicle according to claim 2 is characterized in that For samples that do not obey the normal distribution, statistical software tools are used to reselect new fitting functions.

4. The method for determining the mean mileage between failures of a vehicle according to any one of claims 1 to 3, characterized in that The screening of the fault information records in step B is as follows: among the fault information records of the same vehicle, only the fault information record with the longest maintenance mileage is retained.

5. The method for determining the mean mileage between failures of a vehicle according to any one of claims 1 to 3, characterized in that The identification information includes vehicle model, production date, sales information, and vehicle identification number, and the fault information includes repair date, damaged part name, and damaged part number.

6. The method for determining the mean mileage between failures of a vehicle according to claim 5, characterized in that The fault information also includes the service station name, fault description, troubleshooting process, system classification, amount involved, and maintenance personnel.

Images