CN103530444A - Mechanical and electrical product reliability growth pre-testing method and device - Google Patents

Abstract

The invention discloses a mechanical and electrical product reliability growth pre-testing method. According to the method, reliability of a product can be effectively tested in a short time, and high precision and accuracy are achieved. A mechanical and electrical product reliability growth pre-testing device can carry out reliability growth pre-testing in the mechanical and electrical product reliability growth pre-testing method.

Description

A kind of electronic product reliability growth know the real situation test method and device thereof

Technical field

The present invention relates to environment and fail-test technology, refer to especially a kind of electronic product reliability growth know the real situation test method and device thereof.

Background technology

The fail-test of product is one of evaluation product reliability and the important step in life-span.At present, the fail-test of China's military use product is formulated with reference to the table MIL-STD-781D of U.S. army mostly, and this standard is program, method and the appraisal procedure of the reliability statistics test of exponential distribution electronic product.

Take servo control mechanism, hydraulic product, to send out control executive module etc. be that all kinds of electronic products and the electronic product of representative has larger difference in application characteristic, failure mode etc.In prior art, for a long time the fail-test technology of electrical category product is lacked to systematic research and guidance, but simply continue to use test and the evaluation method of electronic product, far apart with the actual conditions of product.For example: for highly reliable electronic product, by the exponential distribution Censoring reliability definite test period of testing program of knowing the real situation, with the increase of fiduciary level, can be increased sharply, when fiduciary level surpasses 0.999, test period will be more than thousand hours, and this is all worthless from experimentation cost and progress.In addition, the test period that experience is so long, indivedual consume parts of some electronic product have arrived the serious consume phase, and for shorter product task time, reliability growth is now nonsensical.

Because exponential distribution is the harshest distribution, and the life-span of electronic product is obeyed Weibull distribution mostly, and it is irrational indiscriminately imitating exponential distribution fail-test scheme and carrying out electronic product fail-test.

Summary of the invention

In view of this, the object of the invention is to propose a kind of electronic product reliability growth know the real situation test method and device thereof.

Based on above-mentioned purpose, the electronic product reliability growth provided by the invention test method of knowing the real situation, is characterized in that, supposes that life of product obeys Weibull distribution, if form parameter is known, test period is determined according to following formula;

$T={\left(\frac{t^m\ln \mathrm{\β}/\ln R_L}{n}\right)}^{1/m}---\left(4\right)$

Wherein, t is task time, and m is form parameter, and β is user's risk, R _lfail-test required value, n is the product number of participating in the experiment that test adopts.

Optionally, if form parameter is known, after off-test, according to test findings, the product Q-percentile life of participating in the experiment adopts following formula to assess:

$t_{\mathrm{Rlow}}={[-\frac{2\ln R}{{\mathrm{\χ}}^2(2z+2,\mathrm{\γ})}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{t_i}^m]}^{1/m}---\left(3\right)$

Wherein, χ ² _{(2z+2, γ)}that degree of freedom is the χ of 2z+2 ²the 1-γ upside quantile distributing; t _ithe test period that represents i the product of participating in the experiment; Z represents the number of faults in process of the test.

Optionally, if form parameter is unknown, but form parameter lower limit is known, if meet the condition of formula (6), the m in formula (3) adopts form parameter lower limit to substitute:

$R_L\≥\exp [-\frac{{{\mathrm{\χ}}^2}_{(2z+2,\mathrm{\γ})}}{2}\exp (\frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{t}_i^{m_0}\ln t_i^{m_0}}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{t}_i^{m_0}}-\ln \underset{i=1}{\overset{n}{\mathrm{\Σ}}}{t}_i^{m_0})]---\left(6\right)$

M ₀for the shape of product parameter lower limit value of participating in the experiment.

Optionally, if form parameter is known, the production reliability of participating in the experiment adopts following equation to assess:

$R_{\mathrm{Low}}\left(t\right)=\exp [-\frac{t^m\·{\mathrm{\χ}}_{(2z+2,\mathrm{\γ})}^2}{2\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{t}_i^m}]---\left(9\right)$

T _ibe the task time of i product, z represents the number of faults in process of the test, χ ² _{(2z+2, γ)}that degree of freedom is the χ of 2z+2 ²the 1-γ upside quantile distributing, the product number that n representative test is tested; γ is confidence level.

Optionally, when form parameter the unknown, but form parameter lower limit is known, if meet the condition of formula (11), the m in formula (9) adopts form parameter lower limit to substitute,

$t\≤\exp \left(\frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{t}_i^{m_0}\ln t_i}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{t}_i^{m_0}}\right)---\left(11\right)$

M ₀for the shape of product parameter lower limit value of participating in the experiment.

Optionally, when form parameter m the unknown, but form parameter lower limit m ₀known, the m in formula (4) adopts form parameter lower limit m ₀substitute.

Optionally, the reliability growth that the is applied to electronic product test of knowing the real situation.

Optionally, before product is carried out to fail-test, determine product task equivalent time; According to the Reliability Distribution index of development departments, determine fail-test required value R _l.

Further, the invention provides a kind of electronic product reliability growth test unit of knowing the real situation, adopt the test method of knowing the real situation of described any one electronic product reliability growth above to carry out the reliability growth test of knowing the real situation.

As can be seen from above, the electronic product reliability growth provided by the invention test method of knowing the real situation, can carry out effective verification experimental verification to the reliability of product at short notice, has higher precision and accuracy simultaneously.The electronic product reliability growth providing by the present invention and embodiment know the real situation test method and device thereof, rational confirmed test time, solve high reliability verification experimental verification and the test period contradiction between long; Can avoid finite lifetime long due to test period and consumable products; During test findings assessment, adopt formula (4) to carry out reliable life assessment, adopt formula (9) to carry out Reliability assessment, assessment result has better credibility.

Accompanying drawing explanation

Fig. 1 is the electronic product reliability growth of the embodiment of the present invention test method schematic flow sheet of knowing the real situation.

Embodiment

For making the object, technical solutions and advantages of the present invention clearer, below in conjunction with specific embodiment, and with reference to accompanying drawing, the present invention is described in more detail.

Because life-span of electronic product is general, obey Weibull distribution, so the embodiment of the present invention adopts Weibull distribution fixed time test scheme to carry out the reliability test of knowing the real situation, within the time of setting, complete the test of knowing the real situation of the reliability of product.The embodiment of the present invention is carried out the applied research of reliability growth technology for different electronic products, first from the failure mechanism of different electronic products, start with, product bug situation is investigated and data analysis, utilize the existing related data of product, adopt bayesian theory to determine form parameter, the failure message of statistics electronic product, tentatively determines the estimated value of the life-span distribution parameter of product according to prior imformation; Then the mode that adopts integrated environment stress to combine with electric stress, simulate actual stand and simultaneous comprehensive environmental effects is tested product, in the short time, cheaply under condition, expose electronic product and at design, technique, components and parts, select etc. the weak link of aspect.Test findings is assessed, and improved product reliability by well designed, to finding as early as possible design defect and the fault mode of product in the design phase, improve the reliability of product.

It is example that certain aircraft servo control mechanism is take in the present invention, at the experimental program of embodiment, mainly comprises the steps:

Step 1: determine initial parameter: initial parameter mainly comprises task time, fail-test required value, user's risk etc.The reliability of servo control mechanism of the take test of knowing the real situation is example, carrying out reliability while knowing the real situation test, first analyzes aircraft mission profile, determines servo control mechanism task equivalent time t, hereinafter to be referred as task time.The reliability testing requirements value R that knows the real situation _laccording to the Reliability Distribution index of development departments, determine.Consumer's risk β is generally not more than 0.3.

Step 2: confirmed test time:

In certain embodiments, when test period is determined according to exponential distribution, Q-percentile life t _lowrepresented that single product breaks down the needed time, the one-sided confidence lower limit of Q-percentile life that confidence level is γ is as formula (1), and this formula (1) has reflected when confidence level is γ, the minimum life of the single product of participating in the experiment,

$t_{\mathrm{Low}}=[-\frac{2\ln R_L}{{{\mathrm{\χ}}^2}_{(2z+2,\mathrm{\γ})}}T]---\left(1\right)$

Wherein, χ ² _{(2z+2, γ)}that degree of freedom is the χ of 2z+2 ²the 1-γ upside quantile that (card side) distributes, z is supposition number of faults, γ is 1-β in the present embodiment.

The initial trial time is definite by zero failure situation, and the computing method of total time on test are shown in formula (2),

$T=t\frac{\ln \mathrm{\β}}{\ln R_L}---\left(2\right)$

In formula: R _lrepresent the reliability testing requirements value of knowing the real situation; T represents total time on test; T represents task time, i.e. given life value; β represents user's value-at-risk, general β≤30%.

In a preferred embodiment, test period is definite according to Weibull distribution, and the one-sided confidence lower limit of Q-percentile life that confidence level is γ is as formula (3), and formula (3) has reflected when confidence level is γ, the minimum life of the single product of participating in the experiment,

$t_{\mathrm{Rlow}}={[-\frac{2\ln R_L}{{{\mathrm{\χ}}^2}_{(2z+2,\mathrm{\γ})}}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{t_i}^m]}^{1/m}---\left(3\right)$

The initial trial time is definite by zero failure situation, and the computing method of total time on test are shown in following formula:

$T={\left(\frac{t^m\ln \mathrm{\β}/\ln R_L}{n}\right)}^{1/m}---\left(4\right)$

R _lrepresent the reliability testing requirements value of knowing the real situation; T represents total time on test; T represents task time, i.e. given life value; β represents user's value-at-risk, general β≤30%; M represents form parameter; N represents the product number that test adopts.In the present embodiment, form parameter adopts bayesian theory to determine.In other embodiments, the initial trial time also can according to a failure condition, two failure conditions or more the situation of multiple faults determine.

Step 3: according to step 1,2 definite initial parameters and initial time, carry out the reliability test of knowing the real situation, record trouble number, number of faults refers to the product number z' breaking down in the present embodiment.In the present embodiment, the reliability of the step 3 fingering row aircraft servo control mechanism test of knowing the real situation.

Step 4: test findings assessment.In the present embodiment, by step 4, carry out aircraft servo control mechanism reliability assessment, step 4 can comprise following two aspects.

First aspect, Weibull distribution aircraft servo control mechanism life appraisal.

The one-sided confidence lower limit formula of the Q-percentile life that is γ according to exponential distribution confidence level is known, for Weibull distribution stochastic variable, and when form parameter m is known, its given fiduciary level R _lq-percentile life t _rconfidence level be γ one-sided confidence lower limit is formula (3) have:

P(t _R≥t _Rlow)＝γ （5）

In engineering reality, form parameter m is normally unknown, therefore in engineering reality, cannot try to achieve Q-percentile life t according to formula (4) _rconfidence lower limit t _rLow, but in many cases, can know that m is more than or equal to a certain constant m ₀.Be form parameter lower limit m ₀known.Can prove, as given fiduciary level R _lwhile meeting formula (6), its Q-percentile life t _rconfidence level be γ one-sided confidence lower limit is formula (7).

$R_L\≥\exp [-\frac{{{\mathrm{\χ}}^2}_{(2z+2,\mathrm{\γ})}}{2}\exp (\frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{t}_i^{m_0}\ln t_i^{m_0}}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{t}_i^{m_0}}-\ln \underset{i=1}{\overset{n}{\mathrm{\Σ}}}{t}_i^{m_0})]---\left(6\right)$

$t_{\mathrm{Rlow}}={[-\frac{2\ln R_L}{{{\mathrm{\χ}}^2}_{(2z+2,\mathrm{\γ})}}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{t_i}^{m_0}]}^{1/m_0}---\left(7\right)$

In the present embodiment, total time on test is determined according to Weibull distribution, χ ² _{(2z+2, γ)}that degree of freedom is the χ of 2z+2 ²the 1-γ upside quantile distributing; N represents the product number of participating in the experiment, and as a specific embodiment, product number is 1 or 2; t _ithe test period that represents i the product of participating in the experiment, z represents the number of faults in process of the test, m is form parameter, if m is unknown, if meet the condition of formula (6), the m in formula (4) can adopt form parameter lower limit m ₀replace, can obtain formula (7).

Have:

P(t _R≥t _Rlow)≥γ （8）

As can be seen here, the reliable life assessment method that the embodiment of the present invention adopts has higher confidence level.

Second aspect, Weibull distribution aircraft servo control mechanism Reliability assessment.

For Weibull distribution stochastic variable t, when form parameter m is known, its give task time t the confidence level of fiduciary level R (t) be γ one-sided confidence lower limit is

$R_{\mathrm{Low}}\left(t\right)=\exp [-\frac{t^m\·{\mathrm{\χ}}_{(2z+2,\mathrm{\γ})}^2}{2\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{t}_i^m}]---\left(9\right)$

Have:

P(R(t)≥R _Low(t))＝γ （10）

Equally, in engineering reality, form parameter m is normally unknown, but can know that m is more than or equal to a certain constant m ₀.Be form parameter lower limit m ₀known.Adopt m ₀replace m, when given t task time meets formula (11), the one-sided confidence lower limit that the confidence level of its fiduciary level R (t) is γ is formula (12).

$t\≤\exp \left(\frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{t}_i^{m_0}\ln t_i}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{t}_i^{m_0}}\right)---\left(11\right)$

$R_{\mathrm{Low}}\left(t\right)=\exp \left[\frac{t^{m_0}\·{\mathrm{\χ}}_{(2z+2,\mathrm{\γ})}^2}{2\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{t}_i^{m_0}}\right]---\left(12\right)$

Have:

P(R(t)≥R _Low(t))≥γ （13）

Further, the invention provides a kind of electronic product reliability growth test unit of knowing the real situation, this device adopts the electronic product reliability growth provided by the present invention test method of knowing the real situation to test, and carries out following steps during test:

Determine initial parameter: initial parameter mainly comprises aircraft servo control mechanism task time, fail-test required value, user's risk etc.; According to above-mentioned formula (3), calculate test period; According to described initial parameter and test period, test, obtain number of faults; According to number of faults and test period, the life-span of aircraft servo control mechanism and reliability are assessed.

As can be seen from above, the reliability growth provided by the invention test method of knowing the real situation, according to formula (4) the confirmed test time, can carry out effective verification experimental verification to the reliability of product at short notice, has higher precision and accuracy simultaneously.The electronic product reliability growth providing by the present invention and embodiment know the real situation test method and device thereof, rational confirmed test time, solve high reliability verification experimental verification and the test period contradiction between long; Can avoid finite lifetime long due to test period and consumable products; By actual tests, can prove, during outcome evaluation, adopt formula (4) or formula (7) to carry out reliable life assessment, adopt formula (9) or formula (12) to carry out Reliability assessment, assessment result has better credibility.

Those of ordinary skill in the field are to be understood that: the foregoing is only specific embodiments of the invention; be not limited to the present invention; within the spirit and principles in the present invention all, any modification of making, be equal to replacement, improvement etc., within all should being included in protection scope of the present invention.

Claims (8)

1. the electronic product reliability growth test method of knowing the real situation, is characterized in that, test period is determined according to following formula;

$T={\left(\frac{t^m\ln \mathrm{\β}/\ln R_L}{n}\right)}^{1/m}---\left(4\right)$

Wherein, t is task time, and m is form parameter, and β is user's risk, R _lthe reliability testing requirements value of knowing the real situation, n is the product number of participating in the experiment that test adopts.

2. the reliability growth according to claim 1 test method of knowing the real situation, is characterized in that, if form parameter is known, the product Q-percentile life of participating in the experiment adopts following equation to assess:

$t_{\mathrm{Rlow}}={[-\frac{2\ln R_L}{{\mathrm{\χ}}^2(2z+2,\mathrm{\γ})}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{t_i}^m]}^{1/m}---\left(3\right)$

χ ² _{(2z+2, γ)}that degree of freedom is the χ of 2z+2 ²the 1-γ upside quantile distributing; t _ithe task time that represents i the product of participating in the experiment; Z represents the number of faults in process of the test.

3. the reliability growth according to claim 2 test method of knowing the real situation, is characterized in that, if form parameter is unknown, but form parameter lower limit is known, if meet the condition of formula (6), the m in formula (3) adopts form parameter lower limit to substitute:

$R_L\≥\exp [-\frac{{{\mathrm{\χ}}^2}_{(2z+2,\mathrm{\γ})}}{2}\exp (\frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{t}_i^{m_0}\ln t_i^{m_0}}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{t}_i^{m_0}}-\ln \underset{i=1}{\overset{n}{\mathrm{\Σ}}}{t}_i^{m_0})]---\left(6\right)$

M ₀for the shape of product parameter lower limit value of participating in the experiment.

4. the reliability growth according to claim 1 test method of knowing the real situation, is characterized in that, if form parameter is known, the production reliability of participating in the experiment adopts following equation to assess:

$R_{\mathrm{Low}}\left(t\right)=\exp [-\frac{t^m\·{\mathrm{\χ}}_{(2z+2,\mathrm{\γ})}^2}{2\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{t}_i^m}]---\left(9\right)$

T _ibe the task time of i product, z represents the number of faults in process of the test, χ ² _{(2z+2, γ)}that degree of freedom is the χ of 2z+2 ²the 1-γ upside quantile distributing, the product number that n representative test is tested; γ is confidence level.

5. the reliability growth according to claim 4 test method of knowing the real situation, is characterized in that, when form parameter is unknown, but form parameter lower limit is known, if meet the condition of formula (11), the m in formula (9) adopts form parameter lower limit to substitute,

$t\≤\exp \left(\frac{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{t}_i^{m_0}\ln t_i}{\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{t}_i^{m_0}}\right)---\left(11\right)$

M ₀for the shape of product parameter lower limit value of participating in the experiment.

6. the reliability growth according to claim 1 test method of knowing the real situation, is characterized in that, if form parameter m is unknown, but form parameter lower limit m ₀known, the m in formula (4) adopts form parameter lower limit m ₀substitute.

7. the reliability growth according to claim 1 test method of knowing the real situation, is characterized in that, knows the real situation before test the product of participating in the experiment being carried out to reliability, according to the product mission profile of participating in the experiment, determines the product task equivalent time of participating in the experiment; According to the Reliability Distribution index of development departments, determine the reliability testing requirements value R that knows the real situation _land consumer's risk β.

8. the electronic product reliability growth test unit of knowing the real situation, adopts the electronic product reliability growth of any one in the claim 1-7 test method of knowing the real situation to carry out the reliability growth test of knowing the real situation.

Images