CN105389406A - Reliability assessment method for entire vehicle design based on unit weighted cumulative number of failures - Google Patents

Abstract

The invention relates to a reliability assessment method for entire vehicle design based on unit weighted cumulative number of failures. The reliability assessment method comprises following steps: obtaining test sample data of at least first phase of multiple engineering development phases; calculating unit weighted cumulative number of failures corresponding to the at least first phase based on test sample data; modeling a grey system based on unit weighted cumulative number of failures corresponding to unit weighted cumulative numbers of failures corresponding to all engineering development phases in order to form a grey matrix; and predicting unit weighted cumulative number of failures corresponding to remaining phases of multiple engineering development phases by solving the grey matrix. The reliability assessment method for entire vehicle design based on unit weighted cumulative number of failures has following beneficial effects: problems such as small samples, poor information, complicated and unclear transfer functions of the system are solved during a durability test of an entire vehicle; and adverse effect on entire vehicle design is avoided.

Description

Complete vehicle design reliability evaluation method based on unit weighted cumulative number of faults

Technical Field

The invention relates to the technical field of automobile reliability evaluation.

Background

The broad reliability concept refers to the ability of a product to perform a specified function under specified conditions and for a specified time. With the development of reliability research, the reliability is expanded into a plurality of engineering fields such as design reliability, use reliability and reliability increase. For the automobile industry, the design reliability (inherent reliability) of a product determines the use reliability of an automobile, and is one of important indexes of the quality of the automobile.

The automobile engineering development stage can comprise an engineering sample automobile development stage, a prototype automobile development stage, a trial production automobile development stage, a non-sales automobile development stage, a saleable automobile development stage and the like. The evaluation of the reliability of the whole automobile design obviously plays an important role in the development stage of the whole automobile, and the reliability level of the automobile can be mastered and the failure rate of the after-sale market and the like can be predicted in the design stage through the evaluation and the estimation of the reliability index.

In the prior art, for the use reliability of the whole vehicle, fault data (such as mileage information) can be used as random variables to perform distribution function fitting and optimization through a probability and mathematical statistics method, an optimal distribution function F (x) is found, and the reliability level of a product can be evaluated through characteristic analysis of a curve F (x); the technical scheme of establishing a reliability comprehensive evaluation mathematical model by using the whole vehicle use data and evaluating the whole vehicle use reliability also exists.

The scheme in the prior art has obvious limitation when being applied to evaluating the reliability of the whole vehicle design, because the number of samples of the engineering sample vehicle used for the reliability test in the engineering development stage is far smaller than that of samples from after-sales markets, if a reliability evaluation model is established according to the scheme in the prior art, the problems of model distortion, non-convergence of parameter calculation and the like inevitably exist. Therefore, the statistical evaluation model based on large samples is not suitable for the reliability evaluation modeling based on the engineering development stage of small samples.

On the other hand, the grey system theory is fully researched and applied in the prior art. The grey system theory researches poor information modeling, and provides a new way for solving system problems under the condition of poor information. All random processes are regarded as grey processes which are changed in a certain range and are related to time, and the grey quantity is researched not by a large sample from the aspect of finding a statistical rule but by a data generation method, disordered original data are arranged into a generation sequence with strong regularity and then are researched.

Aiming at the problems encountered in the evaluation of the design reliability of the whole automobile, the technical personnel in the field hope to obtain an effective evaluation method which can solve the adverse effects of the problems of small samples, poor information, complex and undefined system transfer function and the like in the durability test process of the whole automobile on the design reliability evaluation of the whole automobile.

Disclosure of Invention

The invention aims to provide a complete vehicle design reliability evaluation method based on unit weighted cumulative number of faults.

In order to achieve the above purpose, the invention provides a technical scheme as follows:

a failure-based complete vehicle design reliability assessment method based on unit weighted cumulative number comprises the following steps: a) acquiring test sample data under at least a first stage in a plurality of engineering development stages, wherein the test sample data comprises fault numbers corresponding to test mileage; b) respectively calculating unit weighted cumulative numbers of the faults corresponding to at least the first stage based on the test sample data; c) performing grey system modeling according to the unit weighted cumulative number of the faults corresponding to each engineering development stage to form a grey matrix; d) predicting unit weighted cumulative numbers of faults corresponding to other stages in a plurality of engineering development stages by solving the gray matrix; and the unit weighted cumulative number of the faults corresponding to any engineering development stage is the sum of weighted products of the fault numbers corresponding to different test mileage in the engineering development stage.

The method for evaluating the reliability of the whole vehicle design can effectively evaluate the reliability of the whole vehicle design, and can solve the adverse effects of small samples, poor information, complex and undefined system transfer function and other problems in the durability test process of the whole vehicle on the evaluation of the reliability of the whole vehicle design. The prediction of the reliability of the whole automobile accords with actual after-sale data, so that clear guidance is provided for research and development of automobile manufacturers; the implementation is simple, and the popularization and the application in the industry are facilitated.

Drawings

Fig. 1 is a schematic flow chart illustrating a method for evaluating reliability of a finished automobile design according to a first embodiment of the present invention;

fig. 2 is a graph showing a relationship between a predicted value, a target value, and an actual value of a unit weight cumulative number of failures obtained in one routine endurance test and each stage of engineering development.

Detailed Description

It should be noted that, in each embodiment of the present invention, the unit weighted cumulative number UWIC of the fault is used to measure the reliability of the entire vehicle design. The unit weighted cumulative count UWIC is a result obtained by assigning values to the number of faults occurring in the test sample according to the fault levels thereof based on a conventional endurance test and performing weighted cumulative calculation. The unit weighted cumulative number is related to the accumulation of test mileage, reflecting the part failure that is indistinguishable during the endurance test. The reliability index of the vehicle indication is set asThenAnd isWherein S is the total mileage of the durability test, and MTBF is the Mean Time Between Failures (MTBF).

In order to accurately calculate the unit weighted cumulative value UWIC of the fault, the following preconditions are made in the invention:

1) in each stage of engineering development, test input and fault finding are sufficient and effective;

2) in each stage of engineering development, the calculation criteria of unit weighted cumulative numbers are consistent;

3) the change in the value of the unit weight cumulative number is continuous at each stage of engineering development.

As shown in fig. 1, a method for evaluating reliability of a finished vehicle design according to a first embodiment of the present invention includes the following steps:

and S10, obtaining test sample data in at least the first stage of the multiple engineering development stages, wherein the test sample data comprises the fault number corresponding to the test mileage.

Specifically, from the conventional endurance test data that has been performed, test sample data at least at the first stage in a plurality of stages of engineering development is acquired.

Preferably, the engineering development stage at least comprises an engineering sample vehicle development stage, a prototype vehicle development stage, a trial production vehicle development stage, a non-sales vehicle development stage and a saleable vehicle development stage.

In step S10, the first stage is, for example, a engineering sample vehicle development stage.

The test sample data includes the mileage during the test, the failure number corresponding to the mileage, and further may include the grade of each failure.

And step S11, respectively calculating unit weighted cumulative numbers of the faults corresponding to at least the first stage based on the test sample data.

Specifically, the calculation formula of the unit weighted cumulative number of the faults corresponding to each engineering development stage is as follows:

wherein L represents any engineering development stage, X_LA unit weighted cumulative number, x, representing the corresponding failure in the engineering development phase_iThe corresponding failure number when the test mileage is i is shown,_idenotes x_iThe corresponding weight coefficient. In the calculation of the unit weighted cumulative number of the corresponding fault in the first stage, L is 1, and the weight coefficient_iThe total situation of the fault occurred when the test mileage is i is evaluated and set by the tester.

Further, the weight coefficient_iThe method is obtained by performing corresponding assignment and weighted calculation on each fault which occurs when the test mileage is i according to the fault grade of the fault.

And step S12, performing grey system modeling according to the unit weighted cumulative number of the faults corresponding to each engineering development stage to form a grey matrix.

Specifically, the calculation formula for forming the gray matrix is:

X⁽⁰⁾x × M, wherein X⁽⁰⁾For a gray-scale matrix, X is a matrix composed of unit-weighted cumulative numbers of failures corresponding to each engineering development stage, i.e., X ═ X₁，X₂，…，X_L]And M is a transformation coefficient matrix, and each element in the transformation coefficient matrix corresponds to the fault percentage associated with each engineering development stage and test input respectively.

Those skilled in the art will appreciate that the test inputs during the durability test of the entire vehicle at each engineering development stage are durability test specifications including various types of load inputs, operational inputs for the driver to simulate real customer use, checks, etc., in a manner that may be summarized and described as "load conditions". The test output refers to the number of faults that occurred under various load conditions. Not all faults are associated with the tested load conditions, in combination with the actual physical meaning of the fault. For example, a failure is analyzed and found to be due to an erroneous component mounting, and thus in this state, whether a failure occurs or not is not related to the load condition of the actual test. The existence of such data is not beneficial to analyzing the reliability index of the whole vehicle, and belongs to data elements with large discreteness, so that the unit weighted cumulative number matrix X needs to be converted and transformed before the correlation analysis is performed. And setting the transformation coefficient matrix as M ═ M1, M2, M3, …, Mn ], wherein Mn represents a transformation factor corresponding to a certain engineering development stage, and based on the unit weighted cumulative number of the fault corresponding to the engineering development stage, Mn ═ unit weighted cumulative number-failure weighted cumulative number/unit weighted cumulative number under the non-load condition. Thus, each element in the transformation coefficient matrix M corresponds to a fault percentage associated with each engineering development stage and test input, respectively.

In the matrix X composed of the unit weighted cumulative numbers of failures corresponding to the engineering development stages, some elements are calculated in the above step S11, and some elements are unknown and remain to be found in the following step S13.

And step S13, through solving the gray matrix, predicting the unit weighted cumulative number of the faults corresponding to the rest stages in the engineering development stages.

In the whole vehicle durability test process, the occurrence of a specific fault has the characteristics of nonnegativity, uncertainty, randomness and the like. The unit weighted cumulative UWIC sequence of faults can therefore be regarded as a random sequence with uncertainty. From the information theory point of view, the uncertain information itself has randomness, and the random signal is the carrier of the information. Therefore, the unit weighted cumulative number UWIC sequence has high correlation with the gray level sequence, and an uncertain system can be converted into a quasi-definite system by establishing a gray level model for the UWIC sequence.

Specifically, in this step, the basic equation of the gray first order univariate model is:

x⁽⁰⁾(k)+az⁽¹⁾(k)＝b，(1)

wherein a is a development coefficient, b is a gray effect amount, and Z⁽¹⁾Is a gray matrix X⁽¹⁾Generation order of the adjacent mean ofAnd (4) columns. When modeling gray data, the gray sequence represents the characteristic values of the uncertainty system of "small samples" and "poor information" with "part of information known and part of information unknown". The physical meaning of equation (1) is that the quantitative relationship of two sequences of randomly perturbed numbers is described by a linear system. Grey sequence X⁽⁰⁾The mathematical model cannot be solved directly by the internal disturbance of the system, therefore, the sequence Z is defined⁽¹⁾Is a sequence X⁽⁰⁾The mean value of (a) generates a sequence, namely:

Z⁽¹⁾＝(z⁽¹⁾(1)，z⁽¹⁾(2)，…，z⁽¹⁾(n)), wherein,

$z^{\left(1\right)}\left(k\right)=\frac{1}{2}(x^{\left(1\right)}\left(k\right)+x^{\left(1\right)}(k-1)),$ wherein,

$x^{\left(1\right)}\left(k\right)={\mathrm{\Σ}}_{i=1}^k{x}^{\left(0\right)}\left(i\right),(k=\mathrm{1,2},\·\·\·,n).$

the whitening equation of equation (1) is:

$\frac{{\mathrm{dx}}^{\left(1\right)}}{\mathrm{dt}}+{\mathrm{ax}}^{\left(1\right)}=b---\left(2\right)$

the time response sequence of equation (1) is:

${\hat{x}}^{\left(1\right)}(k+1)=(x^{\left(1\right)}\left(1\right)-\frac{b}{a})e^{-\mathrm{ak}}+\frac{b}{a},---\left(3\right)$

wherein k is 1, 2, … n,

represents a solution of equation (2) wherein,

$X^{\left(1\right)}=(x^{\left(1\right)}\left(1\right),x^{\left(1\right)}\left(2\right),\·\·\·,x^{\left(1\right)}\left(n\right)),x^{\left(1\right)}\left(k\right)={\mathrm{\Σ}}_{i=1}^k{x}^{\left(0\right)}\left(i\right),(k=\mathrm{1,2},\·\·\·,n).$

the physical meaning of the time response sequence is: after the gray sequence is processed by eliminating discrete items, the trend of the change caused by the disturbance of the data is described by a mathematical method, and the purpose is to disclose the internal change rule. The processed gray data will form a curve in a two-dimensional coordinate system with the characteristics of exponential regularity, smoothness, guidability and the like.

Based on the data rule, the solution obtained by the differential equation of the composed sequenceThe method is obtained by data sequence accumulation operation, and a predicted value can be obtained by carrying out inverse operation on the data sequence accumulation operation, wherein the predicted value is as follows:

${\hat{x}}^{\left(0\right)}(k+1)={\mathrm{\α}}^{\left(1\right)}{\hat{x}}^{\left(1\right)}(k+1)={\hat{x}}^{\left(1\right)}(k+1)-{\hat{x}}^{\left(1\right)}\left(k\right)---\left(4\right)$

through solving the equations of the model, the unit weighted cumulative number UWIC of the faults in other engineering development stages can be estimated.

According to a further improved implementation manner of the above embodiment, step S13 specifically includes the following steps:

s130, solving the gray matrix, and predicting unit weighted cumulative numbers of the faults corresponding to the rest stages in the engineering development stages;

s131, calculating residual errors of unit weighted accumulation numbers of the faults corresponding to the rest stages;

s132, if the residual error is larger than the set threshold value, adjusting the modeling parameters of the gray system, modeling the gray system again, and continuing to execute the steps S130, S131 and S132; if the residual is smaller than the set threshold, the step S13 is ended.

In addition, the following steps can be carried out after the above steps are completed: and the reliability of the whole vehicle design is more fully evaluated and fed back to the vehicle manufacturers by combining the reliability index of the whole vehicle design, the number of the after-sale faults of the vehicle and the prediction of unit weighted accumulation numbers of the faults corresponding to the rest stages in a plurality of engineering development stages.

Compared with the traditional statistical model, the method overcomes the disadvantages of small sample data, can be used for evaluating the reliability of the whole vehicle design, and can obtain the calculation precision with the same level as the statistical model.

Fig. 2 is a graph showing a relationship between a predicted value, a target value, and an actual value of the unit weighted cumulative count UWIC of the failure obtained in one routine endurance test and each engineering development stage. The horizontal axis coordinate represents each engineering development stage, and the vertical axis coordinate represents the unit weighted cumulative number of the faults. Through the calculation and prediction of the unit weighted cumulative number UWIC of the fault and the relationship among the predicted value, the target value and the actual value, the quality of the design reliability of the whole vehicle can be evaluated, whether the long-term quality target of the whole vehicle is met or not can be evaluated, and the like.

The method for evaluating the reliability of the design of the whole automobile can effectively evaluate the reliability of the design of the whole automobile, and can solve the adverse effects of small samples, poor information, complex and undefined system transfer function and other problems in the durability test process of the whole automobile on the evaluation of the reliability of the design of the whole automobile.

Tests prove that the prediction of the reliability of the whole automobile by the assessment method accords with actual after-sales data, so that clear guidance is provided for research and development of automobile manufacturers; the assessment method is simple to implement and is beneficial to popularization and application in the industry.

The above description is only for the preferred embodiment of the present invention and is not intended to limit the scope of the present invention. Various modifications may be made by those skilled in the art without departing from the spirit of the invention and the appended claims.

Claims (6)

1. A failure-based complete vehicle design reliability assessment method based on unit weighted cumulative number comprises the following steps:

a) acquiring test sample data under at least a first stage in a plurality of engineering development stages, wherein the test sample data comprises fault numbers corresponding to test mileage;

b) respectively calculating unit weighted cumulative numbers of the faults corresponding to the at least first stage based on the test sample data;

c) performing grey system modeling according to the unit weighted cumulative number of the faults corresponding to each engineering development stage to form a grey matrix;

d) predicting unit weighted cumulative numbers of faults corresponding to other stages in the engineering development stages by solving the gray matrix;

and the unit weighted cumulative number of the faults corresponding to any engineering development stage is the sum of weighted products of the fault numbers corresponding to different test mileage in the engineering development stage.

2. The complete vehicle design reliability evaluation method according to claim 1, wherein the engineering development phase at least comprises an engineering sample vehicle development phase, a prototype vehicle development phase, a trial production vehicle development phase, a non-sales vehicle development phase and a saleable vehicle development phase.

3. The finished automobile design reliability evaluation method according to claim 1, wherein the formula for calculating the unit weighted cumulative number of the faults corresponding to any one engineering development stage is as follows:

wherein L represents any of said engineering development phases, X_LA unit weighted cumulative number, x, representing the corresponding failure in the engineering development phase_iThe corresponding failure number when the test mileage is i is shown,_idenotes x_iThe corresponding weight coefficient.

4. The finished automobile design reliability evaluation method according to claim 1, wherein in the step c), a calculation formula for forming the gray matrix is as follows:

X⁽⁰⁾x × M, wherein X⁽⁰⁾For the gray matrix, X is a matrix composed of unit weighted cumulative numbers of faults corresponding to each engineering development stage, M is a transformation coefficient matrix, and each element in the transformation coefficient matrix corresponds to each engineering development stage and test respectivelyThe associated failure percentage is entered.

5. The finished automobile design reliability evaluation method according to claim 1, wherein the step d) specifically comprises:

d1) solving the gray matrix, and predicting unit weighted cumulative numbers of the faults corresponding to the rest stages in the engineering development stages;

d2) calculating the residual error of the unit weighted cumulative number of the faults corresponding to the other stages;

d3) if the residual error is larger than a set threshold value, adjusting the modeling parameters of the gray scale system, modeling the gray scale system again, and continuing to execute the steps d1), d2) and d 3); and d) if the residual error is smaller than a set threshold value, ending the step d).

6. The finished automobile design reliability assessment method according to any one of claims 1 to 5, characterized by further comprising, after the step d), the steps of:

and evaluating the design reliability of the whole vehicle by combining the design reliability index of the whole vehicle, the number of the faults after the vehicle is sold and the prediction of unit weighted accumulation numbers of the faults corresponding to the rest stages in the plurality of engineering development stages.