CN103810368B - A kind of reliability prediction algorithm - Google Patents

Abstract

The invention discloses a kind of reliability prediction algorithm, comprise the following steps: crash rate λ that calculating device single particle effect SEE, total dose effect TID and displacement damage effect DD cause respectively_SEE、λ_TIDAnd λ_DD；According to formula λ_space=λ_TID+λ_DD+λ_SEECalculate crash rate λ that space radiation environment causes_space；According to formula λ=λ_NOspace+λ_spaceTotal crash rate λ of calculating device, wherein λ_NOspaceThe crash rate that Yes-No space radiation environment causes.The present invention considers radiation stress response, can be used for the reliability prediction of the radiosensitive components and parts of Aero-Space；Compared with the method for predicting reliability assumed based on constant crash rate, the intended result of inventive algorithm is more accurate.

Description

A kind of reliability prediction algorithm

Technical field

The present invention relates to reliability prediction technical field, be specifically related to a kind of reliability prediction algorithm.

Background technology

Existing method for predicting reliability is mainly based upon the method for predicting reliability that constant crash rate is assumed, such as GJB299 And MIL-HDBK-217, using failure number statistical method, its electronic devices and components reliability prediction result differs very with practical situation Greatly, have the disadvantage that

1) impact that can not cause environmental stress to component reliability carries out quantitative analysis；

2) user procedures control and quality control level are not considered；

3) the new failure mechanisms such as irradiation effect are not considered.

Such as, use GJB299 to carry out PRE-CALCULATING FOR RELIABILITY OF PRODUCTS when certain star power module is proved, but do not consider product Radiosensitive effect, it is contemplated that inaccurate.

Summary of the invention

(1) to solve the technical problem that

The present invention, from component failure mechanism and pattern, studies and sets up the quantitative pass of component failure rate and stress System, considers that component quality level and user procedures control and quality control level simultaneously.At reliability prediction (FIDES) model On the basis of, increase radiation stress response, can be used for the reliability prediction of the radiosensitive components and parts of Aero-Space.

(2) technical scheme

The invention provides a kind of reliability prediction algorithm, comprise the following steps:

The inefficacy that S1, respectively calculating device single particle effect SEE, total dose effect TID and displacement damage effect DD cause Rate λ_SEE、λ_TIDAnd λ_DD；

S2, according to formula λ_space=λ_TID+λ_DD+λ_SEECalculate crash rate λ that space radiation environment causes_space；

S3, according to formula λ=λ_NOspace+λ_spaceTotal crash rate λ of calculating device, wherein λ_NOspaceYes-No space radiation ring The crash rate that border causes.

Further, crash rate λ that described single particle effect SEE causes_SEEBy formula λ_SEE=λ_SEEPhysical∏_PM ∏_ProcessCalculate, wherein λ_SEEPhysicalIt is the crash rate of single particle effect physics contribution, ∏_PMIt is the matter in device manufacturing processes Amount and technical controlling factor, ∏_ProcessBe device research and development, manufacture and use during quality and technical controlling factor.

Further, ∏_PMAnd ∏_ProcessValue with reference to FIDES guide 2009, λ_SEEPhysicalPass through below equation Calculate:

${\mathrm{\λ}}_{\mathrm{SEEPhysical}}=\underset{i}{\overset{\mathrm{SEEtype}}{\mathrm{\Σ}}}\left(\frac{{\mathrm{Annual}\_\mathrm{time}}_{\mathrm{SEEtype}-i}}{8760}{\mathrm{\λ}}_{\mathrm{SEEtype}-i}\right)$

Wherein, Annual_time_SEEtype-iIt is that device powers up the working time；

λ_SEEtype-iIt is the crash rate that causes of i-th kind of single particle effect, λ_SEEtype-iComputing formula be:

${\mathrm{\λ}}_{\mathrm{SEEtype}-i}=\underset{0}{\overset{\∞}{\∫}}\underset{-1}{\overset{1}{\∫}}\underset{0}{\overset{2\mathrm{\π}}{\∫}}f(\mathrm{LET},\mathrm{\θ},\mathrm{\φ}){\mathrm{\σ}}_{\mathrm{SEEtype}-i}(\mathrm{LET},\mathrm{\θ},\mathrm{\φ})\mathrm{d\φd}\left(\cos \mathrm{\θ}\right)\mathrm{dLET}$

Wherein, σ_SEEtype-i(LET, θ, φ) is the device single-particle cross section that test obtains；

F (LET, θ, φ) is anticipated space omnidirectional LET Differential Spectrum, and LET refers to linear energy transfer, and θ is angle, and φ is long-pending Open score.

Further, crash rate λ that described total dose effect TID causes_TIDCalculated by below equation:

${\mathrm{\λ}}_{\mathrm{TID}}=-\frac{1}{T}\ln \{1-\mathrm{\Φ}[\frac{\ln \left(R_{\mathrm{specTID}}\left(T\right)\right)-\mathrm{\μ}}{\mathrm{\σ}}\left]\right\}$

Wherein, T is the working cycle；

Φ is the distribution function of standard normal distribution；

R_specTID(T) it is the ionizing radiation accumulated dose accumulated from work 0 moment of beginning to T moment device；

μ is logarithm normal distribution scale factor, μ=ln (R_MF-TID),Wherein n It is test specimen number, R_TIDFAIL-iThe ionizing radiation dose of accumulation when being i-th sample fails；

σ is logarithm normal distribution form factor, $\mathrm{\σ}={\left(\frac{1}{n-1}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{[\ln \left(R_{\mathrm{TIDFAIL}-i}\right)-\mathrm{\μ}]}^2\right)}^{1/2}.$

Further, for bipolar process device, crash rate λ that described total dose effect TID causes_TIDBy following public affairs Formula calculates:

${\mathrm{\λ}}_{\mathrm{TID}}=-\frac{1}{T}\ln \{1-\mathrm{\Φ}[\frac{\ln \left({\mathrm{PAR}}_{\mathrm{Fail}}\right)-\mathrm{\μ}}{\mathrm{\σ}}\left]\right\}$

Wherein, T is the working cycle；

Φ is the distribution function of standard normal distribution；

PAR_FailIt it is device ionization radiation effect sensitive parameter failure criteria；

μ is that device suffers work cumulative ionization in latter stage radiation dose to be R_specTIDTime, the logarithmic mean value of sensitive parameter,Wherein n is test specimen number, PAR_specTID-iIt it is the ionizing radiation dose of i-th sample accumulation Reach R_specTIDTime sensitive parameter value；

σ is the sensitive parameter distribution shape factor, $\mathrm{\σ}={\left(\frac{1}{n-1}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{[\ln \left({\mathrm{PAR}}_{\mathrm{specTID}-i}\right)-\mathrm{\μ}]}^2\right)}^{1/2}.$

Further, crash rate λ that described displacement damage effect DD causes_DDCalculated by below equation:

${\mathrm{\λ}}_{\mathrm{DD}}=-\frac{1}{T}\ln \{1-\mathrm{\Φ}[\frac{\ln \left(R_{\mathrm{speDD}}\left(T\right)\right)-\mathrm{\μ}}{\mathrm{\σ}}\left]\right\}$

Wherein, T is the working cycle；

Φ is the distribution function of standard normal distribution；

R_specDD(T) it is the Non-ionizing radiation dosage accumulated from work 0 moment of beginning to T moment device；

μ is logarithm normal distribution scale factor, μ=ln (R_MF-TID),Wherein n It is test specimen number, R_TIDFAIL-iThe ionizing radiation dose of accumulation when being i-th sample fails；

σ is logarithm normal distribution form factor, $\mathrm{\σ}={\left(\frac{1}{n-1}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{[\ln \left(R_{\mathrm{TIDFAIL}-i}\right)-\mathrm{\μ}]}^2\right)}^{1/2}.$

Further, crash rate λ that described device non-space radiation environment causes_NOspaceMethod for predicting with reference to FIDES guide 2009。

(3) beneficial effect

The present invention considers radiation stress response, can be used for the reliability prediction of the radiosensitive components and parts of Aero-Space；With The method for predicting reliability assumed based on constant crash rate is compared, and the intended result of inventive algorithm is more accurate.

Accompanying drawing explanation

Fig. 1 is the flow chart of inventive algorithm.

Detailed description of the invention

Below in conjunction with the accompanying drawings and embodiment, the detailed description of the invention of the present invention is described in further detail.Hereinafter implement Example is used for illustrating the present invention, but is not limited to the scope of the present invention.

Fig. 1 is the flow chart of inventive algorithm, and the present invention provides a kind of reliability prediction algorithm, comprises the following steps:

The inefficacy that S1, respectively calculating device single particle effect SEE, total dose effect TID and displacement damage effect DD cause Rate λ_SEE、λ_TIDAnd λ_DD；

S2, according to formula λ_space=λ_TID+λ_DD+λ_SEECalculate crash rate λ that space radiation environment causes_space；

S3, according to formula λ=λ_NOspace+λ_spaceTotal crash rate λ of calculating device, wherein λ_NOspaceYes-No space radiation ring The crash rate that border causes.

Further, crash rate λ that described single particle effect SEE causes_SEEBy formula λ_SEE=λ_SEEPhysical∏_PM ∏_ProcessCalculate, wherein λ_SEEPhysicalIt is the crash rate of single particle effect physics contribution, ∏_PMIt is the matter in device manufacturing processes Amount and technical controlling factor, ∏_ProcessBe device research and development, manufacture and use during quality and technical controlling factor.

Further, ∏_PMAnd ∏_ProcessValue with reference to FIDES guide 2009, λ_SEEPhysicalPass through below equation Calculate:

${\mathrm{\λ}}_{\mathrm{SEEPhysical}}=\underset{i}{\overset{\mathrm{SEEtype}}{\mathrm{\Σ}}}\left(\frac{{\mathrm{Annual}\_\mathrm{time}}_{\mathrm{SEEtype}-i}}{8760}{\mathrm{\λ}}_{\mathrm{SEEtype}-i}\right)$

Wherein, Annual_time_SEEtype-iIt is that device powers up the working time；

λ_SEEtype-iIt is the crash rate that causes of i-th kind of single particle effect, λ_SEEtype-iComputing formula be:

${\mathrm{\λ}}_{\mathrm{SEEtype}-i}=\underset{0}{\overset{\∞}{\∫}}\underset{-1}{\overset{1}{\∫}}\underset{0}{\overset{2\mathrm{\π}}{\∫}}f(\mathrm{LET},\mathrm{\θ},\mathrm{\φ}){\mathrm{\σ}}_{\mathrm{SEEtype}-i}(\mathrm{LET},\mathrm{\θ},\mathrm{\φ})\mathrm{d\φd}\left(\cos \mathrm{\θ}\right)\mathrm{dLET}$

Wherein, σ_SEEtype-i(LET, θ, φ) is the device single-particle cross section that test obtains；

F (LET, θ, φ) is anticipated space omnidirectional LET Differential Spectrum, and LET refers to linear energy transfer, and θ is angle, and φ is long-pending Open score.

Further, crash rate λ that described total dose effect TID causes_TIDCalculated by below equation:

${\mathrm{\λ}}_{\mathrm{TID}}=-\frac{1}{T}\ln \{1-\mathrm{\Φ}[\frac{\ln \left(R_{\mathrm{specTID}}\left(T\right)\right)-\mathrm{\μ}}{\mathrm{\σ}}\left]\right\}$

Wherein, T is the working cycle；

Φ is the distribution function of standard normal distribution；

R_specTID(T) it is the ionizing radiation accumulated dose accumulated from work 0 moment of beginning to T moment device；

μ is logarithm normal distribution scale factor, μ=ln (R_MF-TID),Wherein n It is test specimen number, R_TIDFAIL-iThe ionizing radiation dose of accumulation when being i-th sample fails；

σ is logarithm normal distribution form factor, $\mathrm{\σ}={\left(\frac{1}{n-1}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{[\ln \left(R_{\mathrm{TIDFAIL}-i}\right)-\mathrm{\μ}]}^2\right)}^{1/2}.$

Further, for bipolar process device, crash rate λ that described total dose effect TID causes_TIDBy following public affairs Formula calculates:

${\mathrm{\λ}}_{\mathrm{TID}}=-\frac{1}{T}\ln \{1-\mathrm{\Φ}[\frac{\ln \left({\mathrm{PAR}}_{\mathrm{Fail}}\right)-\mathrm{\μ}}{\mathrm{\σ}}\left]\right\}$

Wherein, T is the working cycle；

Φ is the distribution function of standard normal distribution；

PAR_FailIt it is device ionization radiation effect sensitive parameter failure criteria；

μ is that device suffers work cumulative ionization in latter stage radiation dose to be R_specTIDTime, the logarithmic mean value of sensitive parameter,Wherein n is test specimen number, PAR_specTID-iIt it is the ionizing radiation dose of i-th sample accumulation Reach R_specTIDTime sensitive parameter value；

σ is the sensitive parameter distribution shape factor, $\mathrm{\σ}={\left(\frac{1}{n-1}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{[\ln \left({\mathrm{PAR}}_{\mathrm{specTID}-i}\right)-\mathrm{\μ}]}^2\right)}^{1/2}.$

Further, crash rate λ that described displacement damage effect DD causes_DDCalculated by below equation:

${\mathrm{\λ}}_{\mathrm{DD}}=-\frac{1}{T}\ln \{1-\mathrm{\Φ}[\frac{\ln \left(R_{\mathrm{speDD}}\left(T\right)\right)-\mathrm{\μ}}{\mathrm{\σ}}\left]\right\}$

Wherein, T is the working cycle；

Φ is the distribution function of standard normal distribution；

R_specDD(T) it is the Non-ionizing radiation dosage accumulated from work 0 moment of beginning to T moment device；

μ is logarithm normal distribution scale factor, μ=ln (R_MF-TID),Wherein n It is test specimen number, R_TIDFAIL-iThe ionizing radiation dose of accumulation when being i-th sample fails；

σ is logarithm normal distribution form factor, $\mathrm{\σ}={\left(\frac{1}{n-1}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{[\ln \left(R_{\mathrm{TIDFAIL}-i}\right)-\mathrm{\μ}]}^2\right)}^{1/2}.$

Further, crash rate λ that described non-space radiation environment causes_NOspaceMethod for predicting with reference to FIDES guide 2009。

With specific embodiment, the present invention is illustrated below:

For satellite space radiation Sensitive Apparatus, with reference to FIDES Reliability Modeling, setting up space radiation environment can By property index λ_spaceModel and computational methods.

Assume that total dose effect (TID), displacement damage effect (DD), single particle effect (SEE) are independent mutually, radiation stress Separate, then with temperature, mechanical vibration iso-stress:

1, universal model

Component failure rate λ:

λ=λ_NOspace+λ_space(1)

In expression formula (1),

The total crash rate of λ: device；

λ_NOspace: the crash rate that non-space radiation environment causes, its method for predicting sees FIDESguide 2009；

λ_space: the crash rate that space radiation environment causes；

λ_spaceExpression formula:

λ_space=λ_TID+λ_DD+λ_SEE(2)

In expression formula (2),

λ_TID: the crash rate that TID effect causes；

λ_DD: the crash rate that DD effect causes；

λ_SEE: the crash rate that SEE effect causes.

2, single particle effect model

λ_SEEExpression formula:

λ_SEE=λ_SEEPhysical∏_PM∏_Process(3)

In expression formula (3),

λ_SEEPhysical: the crash rate of single particle effect physics contribution；

∏_PM: the quality in components and parts manufacture process and technical controlling factor；

∏_Process: components and parts research and development, manufacture and use during quality and technical controlling factor；

∏_PMAnd ∏_ProcessObtaining value method with reference to FIDES guide 2009.

λ_SEEPhysicalExpression formula be:

${\mathrm{\λ}}_{\mathrm{SEEPhysical}}=\underset{i}{\overset{\mathrm{SEEtype}}{\mathrm{\Σ}}}\left(\frac{{\mathrm{Annual}\_\mathrm{time}}_{\mathrm{SEEtype}-i}}{8760}{\mathrm{\λ}}_{\mathrm{SEEtype}-i}\right)---\left(4\right)$

In expression formula (4),

λ_SEEtype-i: the crash rate (totally 11 kinds of single particle effects) that i-th kind of single particle effect causes, unit h^-1；

Annual_time_SEEtype-i: device powers up the working time；

λ_SEEtype-iComputing formula:

The device single-particle cross section σ that the calculating of single event rate is obtained by test_SEEtype-i(LET) with anticipated space Omnidirectional's LET Differential SpectrumThe integration of product.General formulae is represented by:

${\mathrm{\λ}}_{\mathrm{SEEtype}-i}=\underset{0}{\overset{\∞}{\∫}}\underset{-1}{\overset{1}{\∫}}\underset{0}{\overset{2\mathrm{\π}}{\∫}}f(\mathrm{LET},\mathrm{\θ},\mathrm{\φ}){\mathrm{\σ}}_{\mathrm{SEEtype}-i}(\mathrm{LET},\mathrm{\θ},\mathrm{\φ})\mathrm{d\φd}\left(\cos \mathrm{\θ}\right)\mathrm{dLET}---\left(5\right)$

Anisotropic radiation environment can be calculated by above formula.

Due to Differential Spectrum in above formulaCarry directional, the most complex, to this end, ECSS-E-ST- 10-12C gives the approximate formula of multiple simplification.

When the single-particle threshold value of device is less than 15Mev-cm²During/mg, the single particle effect that proton causes need to be considered.Proton The single particle effect estimating formula caused is:

${\mathrm{\λ}}_{\mathrm{PSEEtype}-i}=\frac{A}{4}\∫\frac{\mathrm{d\Φ}\left(E\right)}{d\left(E\right)}{\mathrm{\σ}}_{\mathrm{PSEEtype}-i}\left(E\right)\mathrm{dE}---\left(6\right)$

In expression formula (6),

λ_PSEEtype-i: the single-particle crash rate that proton causes；

Φ (E): spatial environments Proton integration is composed；

σ_PSEEtype-i(E): the single particle effect cross section that proton causes.

About device single-particle invalid cost under space radiation environment, according to " Testing andHardness Assurance Guidelines for Sigle Event Transients (SETs) in linearDevices " discussion, The SET of linear unit lost efficacy and obeyed logarithm normal distribution.Discussed below is logarithm based on device single particle effect invalid cost The hypothesis of normal distribution.Ground experiment obtains the single particle effect cross section of sample, by the single particle effect cross section of every sample and Space radiation LET during task estimates that bands of a spectrum enter formula (5), it is possible to obtain every sampleIn logarithm normal distribution Under assuming, the sub-crash rate of average single of device is:

$\stackrel{\‾}{{\mathrm{\λ}}_{\mathrm{SEEtype}-i}}=\exp \left(\mathrm{\μ}\right)---\left(7\right)$

In above formula,The sub-crash rate of average single；

μ: crash rate take the logarithm after meansigma methods,

n；Test specimen number；

The single-particle inefficacy discreet value of jth sample.

At certain confidence level c, under conditions of survival probability p, the single-particle crash rate upper limit of assessment device See expression formula (8):

${\mathrm{\λ}}_{\mathrm{SEEtype}-i_{\mathrm{UPlim}}}=\exp (\mathrm{\μ}+K_{\mathrm{TL}}(n,c,p)\×\mathrm{\σ})---\left(8\right)$

In above formula,

K_TL(n, c p): the normal distribution monolateral tolerance limit factor, can be obtained by inquiry MIL-HDBK-814 adnexa.

σ: logarithm normal distribution form factor, $\mathrm{\σ}=\sqrt{\frac{1}{n-1}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{(\ln \left({\mathrm{\λ}}_{\mathrm{SEEtype}-i_j}\right)-\mathrm{\μ})}^2};$

Note: above to single-particle failure analysis, does not consider that single particle effect safeguard procedures taked by device.

3, total dose effect Prediction Model

λ_TIDExpression formula:

When carrying out defined below, it is assumed that the TID of device lost efficacy and obeys logarithm normal distribution, if known distribution parameter μ, σ Time:

${\mathrm{\λ}}_{\mathrm{TID}}=-\frac{1}{T}\ln \{1-\mathrm{\Φ}[\frac{\ln \left(R_{\mathrm{specTID}}\left(T\right)\right)-\mathrm{\μ}}{\mathrm{\σ}}\left]\right\}---\left(9\right)$

In expression formula (9),

T: duty cycle；

The distribution function of Φ: standard normal distribution, $\mathrm{\Φ}\left(x\right)=\underset{-\∞}{\overset{x}{\∫}}\frac{1}{{\left(2\mathrm{\π}\right)}^{1/2}}{e}^{-u^2/2}\mathrm{du};$

R_specTID(T): start, from task, the ionizing radiation accumulated dose that 0 moment accumulated to T moment device；

μ: logarithm normal distribution scale factor, according to MIL-HDBK-814. μ=ln (R_MF-TID), see expression formula (10)；

σ: logarithm normal distribution form factor, is shown in expression formula (11).

$R_{\mathrm{MF}-\mathrm{TID}}=\exp \left(\frac{1}{n}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\ln \left(E_{\mathrm{TIDFAIL}-i}\right)\right)---\left(10\right)$

In expression formula (10),

N: test specimen number；

R_TIDFAIL-i: the ionizing radiation dose of accumulation during i-th sample fails.

$\mathrm{\σ}={\left(\frac{1}{n-1}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{[\ln \left(R_{\mathrm{TIDFAIL}-i}\right)-\mathrm{\μ}]}^2\right)}^{1/2}---\left(11\right)$

The derivation of expression formula (9) is as follows:

Device at the survival probability of task TID in latter stage effect is:

$\mathrm{Ps}\left(R_{\mathrm{specTID}}\left(T\right)\right)=1-\mathrm{\Φ}\left[\frac{\ln \left(R_{\mathrm{specTID}}\left(T\right)\right)-\mathrm{\μ}}{\mathrm{\σ}}\right],$

Assume device lost efficacy in duty cycle obedience constant failure rate be λ_TIDExponential, duty cycle is T, then The survival probability of the TID effect in task latter stage is:

Ps(T)=exp(-λ_TIDT),

Make Ps (T)=Ps (R_specTID(T) expression formula (9)), is then obtained.

For bipolar process device, under space radiation environment, there is ELDRS(Enhanced LowDose Rate Sensitivity) effect, usually requires that when ground experiment and uses little radiation dose rate.TID Effect Evaluation needs to adopt With PDM(Product Data Management) method, when carrying out defined below, it is assumed that device is under ionizing radiation stress The sensitive parameter of device obeys logarithm normal distribution, if known distribution parameter μ, σ, and sensitive parameter monotone decreasing during irradiation Time, λ_TIDSee expression formula (12):

${\mathrm{\λ}}_{\mathrm{TID}}=-\frac{1}{T}\ln \{1-\mathrm{\Φ}[\frac{\ln \left({\mathrm{PAR}}_{\mathrm{Fail}}\right)-\mathrm{\μ}}{\mathrm{\σ}}\left]\right\}---\left(12\right)$

In expression formula (12),

T: duty cycle；

The distribution function of Φ: standard normal distribution, $\mathrm{\Φ}\left(x\right)=\underset{-\∞}{\overset{x}{\∫}}\frac{1}{{\left(2\mathrm{\π}\right)}^{1/2}}{e}^{-u^2/2}\mathrm{du};$

PAR_Fail: device ionization radiation effect sensitive parameter failure criteria；

μ: suffer task cumulative ionization in latter stage radiation dose R at device_specTIDTime, the logarithmic mean value of sensitive parameter, see Expression formula (13)；

σ: the sensitive parameter distribution shape factor, is shown in expression formula (14).

$\mathrm{\μ}=\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\frac{1}{n}\ln \left({\mathrm{PAR}}_{\mathrm{specTID}-i}\right)---\left(13\right)$

In expression formula (13),

N: test specimen number；

PAR_specTID-i: the ionizing radiation dose of i-th sample accumulation reaches R_specTIDTime sensitive parameter value；

$\mathrm{\σ}={\left(\frac{1}{n-1}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{[\ln \left({\mathrm{PAR}}_{\mathrm{specTID}-i}\right)-\mathrm{\μ}]}^2\right)}^{1/2}---\left(14\right)$

If device is sensitive parameter monotonic increase, λ when irradiation_TIDSee expression formula (15):

${\mathrm{\λ}}_{\mathrm{TID}}=-\frac{1}{t}\ln \{1-\mathrm{\Φ}[\frac{\mathrm{\μ}-\ln \left({\mathrm{PAR}}_{\mathrm{Fail}}\right)}{\mathrm{\σ}}\left]\right\}---\left(15\right)$

4, displacement damage effect Prediction Model

λ_DDExpression formula:

When carrying out defined below, it is assumed that the DD of device lost efficacy and obeys logarithm normal distribution, if known distribution parameter μ, σ Time:

${\mathrm{\λ}}_{\mathrm{DD}}=-\frac{1}{T}\ln \{1-\mathrm{\Φ}[\frac{\ln \left(R_{\mathrm{speDD}}\left(T\right)\right)-\mathrm{\μ}}{\mathrm{\σ}}\left]\right\}---\left(16\right)$

In expression formula (16),

T: duty cycle；

The distribution function of Φ: standard normal distribution, $\mathrm{\Φ}\left(x\right)=\underset{-\∞}{\overset{x}{\∫}}\frac{1}{{\left(2\mathrm{\π}\right)}^{1/2}}{e}^{-u^2/2}\mathrm{du};$

R_specDD(T): start, from task, the Non-ionizing radiation dosage that 0 moment accumulated to T moment device；

μ: logarithm normal distribution scale factor, according to MIL-HDBK-814. μ=ln (R_MF), reference expression formula (10)；

σ: logarithm normal distribution form factor, reference expression formula (11).

5, accumulated dose and displacement damage effect evaluation model

When more than respectively illustrating known device TID and DD invalid cost parameter, for given duty cycle t and task The dosage of accumulation in latter stage, it is contemplated that the Equivalent Failure Rate (constant) between device must in office.It should be noted that for TID and DD Losing efficacy, actual crash rate increases with the dosage of accumulation and increases.

If having obtained the radiation test data of device, in given test specimen number n and confidence level c and task accumulation in latter stage Radiation dose R_specUnder, evaluate the λ of device_TIDOr λ_DD, should carry out as follows.For ease of discussing, only provide λ_TIDDerivation Step.

1) calculating device is at the survival probability p in task latter stage so that below equation is set up:

μ-K_TL(n,c,p)σ=ln(R_specTID) (17)

In expression formula (11),

K_TL(n, c, p) be the normal distribution monolateral tolerance limit factor, can be obtained by inquiry MIL-HDBK-814 adnexa.

2) λ in solving equation (18)_TIDValue:

exp(-λ_TIDt)=p (18)

For having the bipolar process device of ELDRS effect, if device sensitive parameter monotone decreasing when irradiation, task end The survival probability p of phase is expression formula (19), if sensitive parameter monotonic increase, survival probability p is expression formula (20).

μ-K_TL(n,c,p)σ=ln(PAR_Fail) (19)

μ+K_TL(n,c,p)σ=ln(PAR_Fail) (20)

6, case is calculated

Case 1: assuming that certain satellite is launched for 2012, track is GEO, 12 years projected lives, certain device equivalence in satellite capsule Shielding thickness is 4mm, utilizes spatial environments to estimate software and obtain device at the ionizing radiation dose of accumulation in task latter stage and is 69.8krad(Si).Device TID lost efficacy and obeyed logarithm normal distribution, R_MF=147.3krad(Si) (i.e. μ=11.9), σ=0.33, ask The λ of device_TID。

Given data is brought into expression formula (9) obtain:

${\mathrm{\λ}}_{\mathrm{TID}}=-\frac{1}{8760*12}\ln \{1-\mathrm{\Φ}[\frac{\ln \left(69.8\right)-11.9}{0.33}\left]\right\}=1.14\×{10}^{-7}{h}^{-1}.$

Case 2, satellite needs to select the amplifier of certain bipolar process, if the space radiation environment section of timing device such as table 1 Shown in.Carried out space radiation environment simulation test on ground, TID test uses low dose rate irradiation, test specimen 10, examination Testing accumulated dose is 69krad(Si), sensitive parameter is Ib, the Ib monotonic increase when irradiation, and failure criteria is 7000nA, test The results are shown in Table 2；DD tests, and test specimen is 11, test to all tested component failures, accumulation when each sample lost efficacy 10MeV equivalence fluence is shown in Table 3；SEE tests, test specimen number 5, finds that the main SEE of device lost efficacy for SET, root during test The cross section obtained according to test combines mission profile LET spectrum, it is contemplated that the single-particle crash rate during each sample task is as shown in table 4. Under 90% confidence level, evaluate the space radiation reliability index of device.

Table 1 space radiation environment section table example

Table 2 is irradiated to 69krad(Si) time Ib test value

Sample number into spectrum	Ib(nA)	ln(Ib)
			1#	2427	7.794
2#	2830	7.958
			3#	1765	7.476
4#	1671	7.421
			5#	1322	7.187
6#	2048	7.625
			7#	3396	8.130
8#	2682	7.894
			9#	1587	7.370
10#	1780	7.484

Table 3DD test specimen inefficacy accumulation fluence

Table 4SEE test data

Device number	SET error rate λ (h-1)	ln(λ)
			1#	0.0257	-3.66
2#	0.0239	-3.73
			3#	0.0224	-3.80
4#	0.0212	-3.85
			5#	0.0275	-3.59

1) λ_TIDCalculate

Calculate according to formula (13) $\mathrm{\μ}=\underset{i=1}{\overset{10}{\mathrm{\Σ}}}\frac{1}{10}\ln \left({\mathrm{PAR}}_{\mathrm{specTID}-i}\right)=7.633,$

Calculate according to formula (14) $\mathrm{\σ}={\left(\frac{1}{n-1}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{[\ln \left({\mathrm{PAR}}_{\mathrm{specTID}-i}\right)-\mathrm{\μ}]}^2\right)}^{1/2}=0.298,$

Bring formula (20), K into_TL(10,0.9, p)=4.10, the survival probability trying to achieve task latter stage is 99.6%, solving equation (18), λ is obtained_TID=3.8×10^-8h^-1。

2) λ_DDCalculate

Calculate according to formula (10) $\mathrm{\μ}=\underset{i=1}{\overset{11}{\mathrm{\Σ}}}\frac{1}{11}\ln \left(R_{\mathrm{specDD}-i}\right)=24.78,$

Calculate according to formula (11) $\mathrm{\σ}={\left(\frac{1}{n-1}\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{[\ln \left(R_{\mathrm{specDD}-i}\right)-\mathrm{\μ}]}^2\right)}^{1/2}=0.4,$

Bring formula (17) into and obtain K_TL(11,0.9, p)=1.7, the survival probability trying to achieve task latter stage is 85%, solving equation (18), λ is obtained_DD=1.5×10^-6h^-1。

3) λ_SEECalculate

Obtain according to formula (7) $\mathrm{\μ}=\underset{j=1}{\overset{n}{\mathrm{\Σ}}}\frac{1}{n}\ln \left({\mathrm{\λ}}_{\mathrm{SEEtype}-i_j}\right)=-3.73,$

Obtain according to formula (8) $\mathrm{\σ}=\sqrt{\frac{1}{n-1}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{(\ln \left({\mathrm{\λ}}_{\mathrm{SEEtype}-i_j}\right)-\mathrm{\μ})}^2}=0.1,$

${\mathrm{\λ}}_{\mathrm{SEEtype}-i_{\mathrm{UPlim}}}=\exp (\mathrm{\μ}+K_{\mathrm{TL}}(n,c,p)\×\mathrm{\σ})=0.0316h^{-1},(K_{\mathrm{TL}}\left(\mathrm{5,0.9,0.9}\right)=1.74),$

Bring formula (4) into and obtain λ_SEEPhysical=0.0307h^-1,

Bring formula (3) into, take ∏_PM=1.7, ∏_Process=4, obtain λ_SEE=0.2h^-1。

The above is only the preferred embodiment of the present invention, it is noted that for the ordinary skill people of the art For Yuan, on the premise of without departing from the technology of the present invention principle, it is also possible to make some improvement and replacement, these improve and replace Also should be regarded as protection scope of the present invention.

Claims (6)

1. a reliability prediction algorithm, it is characterised in that comprise the following steps:

The crash rate that S1, respectively calculating device single particle effect SEE, total dose effect TID and displacement damage effect DD cause λ_SEE、λ_TIDAnd λ_DD；

S2, according to formula λ_space=λ_TID+λ_DD+λ_SEECalculate crash rate λ that space radiation environment causes_space；

S3, according to formula λ=λ_NOspace+λ_spaceTotal crash rate λ of calculating device, wherein λ_NOspaceYes-No space radiation environment is drawn The crash rate risen；

Crash rate λ that described total dose effect TID causes_TIDCalculated by below equation:

${\mathrm{\λ}}_{TID}=-\frac{1}{T}ln\{1-\mathrm{\Φ}\[\frac{ln\left(R_{specTID}\right(T\left)\right)-{\mathrm{\μ}}_{TID-RDM}}{{\mathrm{\σ}}_{TID-RDM}}\]\}$

Wherein, T is the working cycle；

Φ is the distribution function of standard normal distribution；

R_specTID(T) it is the ionizing radiation accumulated dose accumulated from work 0 moment of beginning to T moment device；

μ_TID-RDMIt is logarithm normal distribution scale factor, μ_TID-RDM=ln (R_MF-TID), Wherein n is test specimen number, R_TIDFAIL-iThe ionizing radiation dose of accumulation when being i-th sample fails；

σ_TID-RDMIt is logarithm normal distribution form factor,

2. algorithm as claimed in claim 1, it is characterised in that crash rate λ that described single particle effect SEE causes_SEEBy public affairs Formula λ_SEE=λ_SEEPhysicalΠ_PMΠ_ProcessCalculate, wherein λ_SEEPhysicalIt is the crash rate of single particle effect physics contribution, Π_PMIt is Quality in device manufacturing processes and technical controlling factor, Π_ProcessBe device research and development, manufacture and use during quality and Technical controlling factor.

3. algorithm as claimed in claim 2, it is characterised in that Π_PMAnd Π_ProcessValue with reference to FIDES guide 2009, λ_SEEPhysicalCalculated by below equation:

${\mathrm{\λ}}_{SEEPhysical}=\underset{i}{\overset{SEEtype}{\Σ}}\left(\frac{Annual\_{\mathrm{time}}_{SEEtype-i}}{8760}{\mathrm{\λ}}_{SEEtype-i}\right)$

Wherein, Annual_time_SEEtype-iIt is that device powers up the working time；

λ_SEEtype-iIt is the crash rate that causes of i-th kind of single particle effect, λ_SEEtype-iComputing formula be:

${\mathrm{\λ}}_{SEEtype-i}=\underset{0}{\overset{\mathrm{\∞}}{\∫}}\underset{-1}{\overset{1}{\∫}}\underset{0}{\overset{2\mathrm{\π}}{\∫}}f\left(LET,\mathrm{\θ},\mathrm{\φ}\right){\mathrm{\σ}}_{SEEtype-i}\left(LET,\mathrm{\θ},\mathrm{\φ}\right)d\mathrm{\φ}d\left(\cos \mathrm{\θ}\right)dLET$

Wherein, σ_SEEtype-i(LET, θ, φ) is the device single-particle cross section that test obtains；

F (LET, θ, φ) is anticipated space omnidirectional LET Differential Spectrum, and LET refers to linear energy transfer, and θ is angle, and φ is integration Spectrum.

4. algorithm as claimed in claim 1, it is characterised in that for bipolar process device, described total dose effect TID causes Crash rate λ_TID-1Calculated by below equation:

${\mathrm{\λ}}_{TID-1}=-\frac{1}{T}ln\{1-\mathrm{\Φ}\[\frac{ln\left({\mathrm{PAR}}_{Fail}\right)-{\mathrm{\μ}}_{TID-PDM}}{{\mathrm{\σ}}_{TID-PDM}}\]\}$

Wherein, T is the working cycle；

Φ is the distribution function of standard normal distribution；

PAR_FailIt it is device ionization radiation effect sensitive parameter failure criteria；

μ_TID-PDMIt is that device suffers work cumulative ionization in latter stage radiation dose to be R_specTIDTime, the logarithmic mean value of sensitive parameter,Wherein n is test specimen number, PAR_specTID-iIt it is the ionization spoke of i-th sample accumulation Penetrate dosage and reach R_specTIDTime sensitive parameter value；

σ_TID-PDMIt is the sensitive parameter distribution shape factor,

5. algorithm as claimed in claim 1, it is characterised in that crash rate λ that described displacement damage effect DD causes_DDBy with Lower formula calculates:

${\mathrm{\λ}}_{DD}=-\frac{1}{T}ln\{1-\mathrm{\Φ}\[\frac{ln\left(R_{speDD}\right(T\left)\right)-{\mathrm{\μ}}_{DD}}{{\mathrm{\σ}}_{DD}}\]\}$

Wherein, T is the working cycle；

Φ is the distribution function of standard normal distribution；

R_specDD(T) it is the Non-ionizing radiation dosage accumulated from work 0 moment of beginning to T moment device；

μ_DDIt is logarithm normal distribution scale factor, μ_DD=ln (R_MF-DD),Wherein n is Test specimen number, R_DDFAIL-iThe Non-ionizing radiation dosage of accumulation when being i-th sample fails；

σ_DDIt is logarithm normal distribution form factor,

6. algorithm as claimed in claim 1, it is characterised in that crash rate λ that described non-space radiation environment causes_NOspace's Method for predicting is with reference to FIDES guide 2009.