CN108776294B - Circuit board service life evaluation method based on self-adaptive strategy - Google Patents
Circuit board service life evaluation method based on self-adaptive strategy Download PDFInfo
- Publication number
- CN108776294B CN108776294B CN201810558455.3A CN201810558455A CN108776294B CN 108776294 B CN108776294 B CN 108776294B CN 201810558455 A CN201810558455 A CN 201810558455A CN 108776294 B CN108776294 B CN 108776294B
- Authority
- CN
- China
- Prior art keywords
- failure
- circuit board
- component
- layer
- sampling
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/28—Testing of electronic circuits, e.g. by signal tracer
Abstract
The invention discloses a circuit board life evaluation method based on a probability failure physical model and a multilayer Monte Carlo simulation method, which can obviously improve the efficiency of circuit board life evaluation work. The method comprises the following steps: 1 estimating a time distribution parameter before failure based on a physical failure model. And 2, constructing a component set with significant influence based on the component failure time. Firstly, sampling all components on a circuit board, determining a component set which has obvious influence on the service life of the circuit board, and secondly, selecting other-level component sets according to a method capable of maximizing the simulation efficiency of the multilayer Monte Carlo. And 3, evaluating the average service life of the circuit board based on a multilayer Monte Carlo method. The method comprises the steps of setting initial sample number, updating the sample number of a core set and an outer set, judging a simulation termination condition and outputting a result.
Description
Technical Field
The invention provides a circuit board life evaluation method based on a probability failure physical model and a multilayer Monte Carlo simulation method. The method is suitable for evaluating the service life of a circuit board containing a large number of components in the design stage, and improves the method of adopting fixed sampling number for all the components in the original simulation process by dividing the circuit board with a significant influence component set and a non-significant influence component set on the board-level service life and adaptively adjusting the sampling times of the components, thereby realizing the significant improvement of the simulation efficiency under the same precision constraint. The invention belongs to the field of reliability and system engineering.
Background
The technological level of modern electronic components is rapidly developed, high-quality components have the characteristics of high reliability and long service life, and circuit boards formed by a plurality of components have the same characteristics. These characteristics lead to the fact that the evaluation of the service life of the circuit board by using the traditional reliability prediction method (such as GJB/Z299C-2006 electronic equipment reliability prediction handbook), reliability test and other means is neither economical nor efficient.
Currently, reliability prediction methods based on failure physical models have been increasingly applied at home and abroad. The method starts from the failure mechanism of the electronic component and the packaging thereof, and can comprehensively consider the influence of various stresses including heat, vibration, electricity and the like on the service life of the component. On the basis of the existing failure physical model, the cost of performing reliability test on components can be greatly reduced, the test time is effectively saved, and the rapid iteration process of circuit board design is realized.
However, the life evaluation method based on the failure physical model needs to perform multiple simulations, and the simulation process is extremely costly in terms of computing resources and time. How to improve the simulation efficiency and reduce the consumption of computing resources is the first problem to be solved by applying the method. The method can realize the effect of self-adaptively adjusting the sampling times and rapidly converging under the same precision based on the probability failure physical model and the multilayer Monte Carlo simulation method, and has obvious efficiency improvement compared with the common life evaluation simulation method.
Disclosure of Invention
The invention provides a circuit board life evaluation method based on a probability failure physical model and a multilayer Monte Carlo simulation method. Aims and solves the problems that: the efficiency of the circuit board service life evaluation method based on the probability failure physical model is improved. Firstly, simulating all components and packages thereof on a circuit board for multiple times, determining components, Plated Through Holes (PTHs) and package sets which have obvious influence on the service life of the circuit board, and simultaneously constructing a plurality of component sets which take the component sets with obvious influence as cores and have progressive inclusion relations; secondly, setting an initial sampling sample number, sampling a plurality of component sets constructed in the previous step, and calculating the number of samples needing to be newly added in each level set according to a formula described below; and finally, judging whether the simulation process is finished according to the variance of the overall sample, and outputting a simulation result.
The invention relates to a circuit board life evaluation method based on a probability failure physical model and a multilayer Monte Carlo simulation model, which mainly comprises the following three parts:
a first part: and estimating a time distribution parameter before failure based on the failure physical model.
The failure physical model of each component is the basis for developing the probability failure physical model evaluation of the circuit board, and the evaluation process comprises the following three steps:
step 1: determining the geometric parameters, technological parameters, material parameters, packaging parameters and the like of the components, setting the size (such as length, width and height) dispersity of the components, and setting the dispersity of the parameters of factors such as heat, vibration and the like which influence the service life of the components.
Step 2: and executing a primary Monte Carlo simulation process, and substituting parameters required by the failure physical model into the model to obtain the failure time of the component in each simulation process, so that failure time sequences of the component under different failure mechanisms can be obtained after multiple rounds of simulation.
And step 3: and (3) performing distribution fitting on the failure time sequence obtained in the step (2) to obtain the failure time (service life) distribution type and distribution parameters of each component.
A second part: a set of components having a significant impact is constructed based on component failure times.
The lifetime of a circuit board is typically determined by the components it contains that have a short failure time. The average failure time of all the devices can be obtained by pre-sampling all the devices, so that the significant degree sequence of the influence of the devices on the service life of the circuit board can be obtained according to the sequence from small to large of the average time before failure. And then, obtaining a component set which has a significant influence on the service life of the circuit board by using a method described below, and evaluating the service life of the circuit board by only sampling the component set, wherein the evaluation value has a large error. Therefore, the precision of the service life evaluation of the circuit board can be further corrected by combining the sampling of the rest component sets. In the method, a component set with progressive inclusion relation is the basis of multilayer Monte Carlo simulation, and the construction process comprises the following three steps:
step 1: the number of sets L having progressive inclusion relationships is determined. In general, the components are collected into a set CLThe following progressive relationship is satisfied:
and is provided with
Wherein m ═ CLL is the total number of components
Step 2: a set of components for layer 0 is determined.
Respectively sampling the failure time of each device, arranging the average values of the obtained samples from small to large, and taking the average value beforeAs a component set at layer 0.
And step 3: and determining the component assemblies of other layers.
In order to maximize the multi-layer Monte Carlo simulation sampling efficiency, the component with the largest failure time difference with the previous layer in the rest components can be extracted to form a component set of the current layer, and the formula is as follows:
the number of elements of the current layer is
l denotes the current number of layers, Tl-1For the sampling time of the previous level, k denotes index identification of the remaining devices, j denotes sampling index identification,
the value of the time sample at the jth sample of the kth device representing the remaining devices.
And a third part: the average life of the circuit board was evaluated based on a multi-layer monte carlo method.
By applying multilayer Monte Carlo simulation, the average life of a circuit board is under the constraint of a given precision epsilon, the number of layers and the number of samples of the multilayer Monte Carlo simulation are adjusted in a self-adaptive manner until the given precision is reached, and the basic process comprises the following five steps:
step 1: setting an initial layer number
Initial number of samples N for each layerl。
Step 2: for component sets of each layerSampling, and obtaining the sampling result Y of each layer according to the following formulal
And step 3: calculate the variance Var (Y) of each layer samplel)(YlFor the sample values of each layer), while calculating the following equation,
if there is
If it is true, then order
Then turning to step 5; otherwise NlAnd (4) switching to the step 4 without changing.
And 4, step 4: and sampling the newly added samples in each layer, and then turning to the step 3.
And 5: if Yl| is greater than or equal to epsilon, let
Turning to step 3; if YlIf | < epsilon, terminating the simulation process and returning the estimation result
Drawings
FIG. 1 shows a circuit board life evaluation method architecture based on a probabilistic failure physical model and a multi-layer Monte Carlo simulation method
FIG. 2 design phase circuit board model
FIG. 3 device failure time estimation flow
Detailed Description
Description of the embodiments: a circuit board life evaluation method based on a probability failure physical model and a multilayer Monte Carlo simulation method is provided, the complete flow of the method is shown in FIG. 1, and the specific implementation mode is explained as follows:
a first part: and estimating a time distribution parameter before failure based on the failure physical model.
The failure physical model of each component is the basis for developing a failure physical model evaluation method of the circuit board based on probability, and the construction process comprises the following steps:
step 1: the design diagram of the circuit board and the components thereon is shown in fig. 2, the material and the packaging property of each component are determined, the distribution type and the distribution parameter of the components on the property values with dispersion such as length, width and height caused by the process are set, the components are influenced by heat and vibration in the working environment, and the random variation parameters of the heat and the vibration can also be set.
The properties of dispersion among the board subunits are as follows:
table 1 Attribute item with scatter
Step 2: and (3) executing a primary Monte Carlo simulation process, substituting parameters required by the failure physical model in each simulation process, iterating for 1000 times to obtain a time sequence before failure of each device, wherein the iteration times meet the fitting requirement of the time sequence before failure, and the iteration times can be adjusted according to the fitting effect in the step (3). The flow chart is schematically shown in FIG. 3.
And step 3: and (3) performing distribution fitting on the time sequence before failure to obtain the failure distribution type and distribution parameters of each device, wherein the results are shown in table 2.
TABLE 2 time distribution fitting results before failure of components
A second part: and constructing a component set with progressive inclusion relation.
The part mainly comprises a method for selecting elements in layer 0 and non-layer 0 sets in multilayer Monte Carlo simulation. The method comprises the following 3 steps:
step 1: first, the number of sets having progressive inclusion relation is determined, and the number of sets can be set to be 3 according to the number of devices on a board layer.
Step 2: a set of components for layer 0 is determined.
When the number m of the devices on the current board layer is 12, each device is pre-sampled, the number N' of the pre-sampled samples is 50, the average failure time of the pre-sampling of the devices can be obtained, the average failure time is arranged according to the ascending order of the average failure time, and a table 3 can be obtained, and the number of the devices is obtained before the average failure time
The device is taken as a component assembly of the 0 th layer and is marked as C0Three-transistor, patch transistor, capacitor, resistor }:
TABLE 3 average failure time of Pre-sampling
| Serial number | Device bit number | Mean time to failure (hours) |
| 1 | Triode transistor | 1.1*E6 |
| 2 | Patch triode | 1.20*E6 |
| 3 | Capacitor with a capacitor element | 1.21*E6 |
| 4 | Resistance (RC) | 1.24*E6 |
And step 3: and determining the component assemblies of other layers.
In order to maximize the simulation efficiency of the multi-layer Monte Carlo sampling, the device with the largest failure time difference with the previous layer in the residual device set can be extracted to form the device set of the current layer. The set of all combinations of the devices is expressed as C, and the calculation result in the step 2 is used according to the following formula
The calculated values were arranged in descending order as shown in table 4:
table 4 sampling results of the remaining devices
Before selection
An element, and a set C of the upper layer0Form a set C1The circuit comprises a transistor, a patch transistor, a capacitor, a resistor, a relay, a patch resistor and a DIP switch.
The set of third layers may be a whole set of devices or may still be selected according to the rules described above.
And a third part: and estimating the average service life of the circuit board.
By applying the multilayer Monte Carlo simulation method, the number of layers and the number of samples of the multilayer Monte Carlo simulation are adjusted in a self-adaptive manner under the constraint that the average service life of the circuit board is equal to 0.01 until the given precision is reached. The whole process comprises the following steps:
step 1: setting an initial layer numberInitial number of samples N for each layerl＝100。
Step 2: and sampling the devices of each layer to obtain a sampling value of each layer.
And step 3: calculate the variance Var (Y) of each layer samplel)(YlFor the sample values of each layer), while calculating the following equation,
if there is
If it is true, then order
Then turning to step 5; otherwise NlAnd (4) switching to the step 4 without changing.
And 4, step 4: and sampling the newly added samples in each layer, and then turning to the step 3.
And 5: if Yl| is greater than or equal to epsilon, letTurning to step 3; if YlIf | < epsilon, terminating the simulation process and returning the estimation result
Claims (1)
1. A circuit board life evaluation method based on a probability failure physical model and a multilayer Monte Carlo simulation method mainly comprises the following three parts:
a first part: estimating a time distribution parameter before failure based on a failure physical model;
the failure physical model of each component is the basis for developing a failure physical model evaluation method of a circuit board based on probability, and the estimation process comprises the following three steps:
step 1: determining geometric parameters, process parameters, material parameters and packaging parameters of the components, setting the length, width and high dispersity of the components, and setting the parameter dispersity of heat and vibration factors influencing the service life of the components;
step 2: executing a primary Monte Carlo simulation process, and substituting parameters required by the failure physical model into the model to obtain the failure time of the component in each simulation process, so that a time sequence before the component fails under different failure mechanisms can be obtained after multiple times of simulation;
and step 3: performing distribution fitting on the time sequence before failure obtained in the step 2 to obtain failure distribution types and distribution parameters of each component;
a second part: constructing a component set with significant influence based on component failure time;
the service life of the circuit board is usually determined by components with short time before failure, and the average time before failure of all the components can be obtained by pre-sampling all the components, so that the significant degree sequence of the influence of the components on the service life of the circuit board can be obtained according to the sequence of the average time before failure from small to large; therefore, a large amount of estimation can be carried out on the part of devices, a component set which has obvious influence on the service life of the circuit board can be obtained by using the method described below, and the service life of the circuit board can be evaluated only by sampling the component set; in addition, the precision of the service life evaluation of the circuit board can be further corrected by combining with the rest component sets to perform a small amount of sampling, and the component set with the progressive inclusion relation in the method is the basis of multilayer Monte Carlo simulation, and the construction process comprises the following three steps:
step 1: firstly, determining the number L of sets with progressive inclusion relation; in general, the components are collected into a set CLThe following progressive relationship is satisfied:
and is provided with
Wherein m ═ CLL is the total number of the components;
step 2: determining a component set of a 0 th layer;
respectively sampling the failure time of each device, arranging the average values of the obtained samples from small to large, and taking the average value before
The device of (1) is used as a component assembly of the 0 th layer;
and step 3: determining component sets of other layers;
in order to maximize the multi-layer Monte Carlo simulation sampling efficiency, the component with the largest failure time difference with the previous layer in the remaining components can be extracted to form a component set of the current layer, and the formula is as follows:
the number of elements of the current layer isl denotes the current number of layers, Tl-1For the sampling time of the previous level, k denotes index identification of the remaining devices, j denotes sampling index identification,
a time sample value representing the remaining device at the jth device sampled at the jth time;
and a third part: evaluating the average service life of the circuit board based on a multilayer Monte Carlo method;
by applying multilayer Monte Carlo simulation, the average life of a circuit board is under the constraint of a given precision epsilon, the number of layers and the number of samples of the multilayer Monte Carlo simulation are adjusted in a self-adaptive manner until the given precision is reached, and the basic process comprises the following five steps:
step 1: setting an initial layer number
Initial number of samples N for each layerl；
Step 2: sampling the component sets of each layer, and obtaining the sampling result Y of each layer according to the following formulal
And step 3: calculate the variance Var (Y) of each layer samplel)，YlFor the sample values of each layer, the following formula is calculated at the same time,
if there is
If it is true, then orderThen turning to step 5; otherwise NlChanging to the step 4 without changing;
and 4, step 4: sampling the newly added samples in each layer, and then turning to the step 3;
and 5: if Yl| is greater than or equal to epsilon, let
Turning to step 3;
if YlIf | < epsilon, terminating the simulation process and returning the estimation result
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810558455.3A CN108776294B (en) | 2018-06-01 | 2018-06-01 | Circuit board service life evaluation method based on self-adaptive strategy |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810558455.3A CN108776294B (en) | 2018-06-01 | 2018-06-01 | Circuit board service life evaluation method based on self-adaptive strategy |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108776294A CN108776294A (en) | 2018-11-09 |
| CN108776294B true CN108776294B (en) | 2020-02-14 |
Family
ID=64026597
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810558455.3A Active CN108776294B (en) | 2018-06-01 | 2018-06-01 | Circuit board service life evaluation method based on self-adaptive strategy |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108776294B (en) |
Family Cites Families (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP4269982B2 (en) * | 2004-03-12 | 2009-05-27 | トヨタ自動車株式会社 | Failure diagnosis device for exhaust gas recirculation device |
| CN102103658B (en) * | 2010-12-03 | 2014-05-14 | 北京航空航天大学 | Practical electronic product service life evaluation model parameter high-accuracy extraction method |
| CN102651054B (en) * | 2012-04-10 | 2014-04-30 | 北京航空航天大学 | Probability method of electronic product service life model based on Bayesian theory |
| CN103955568A (en) * | 2014-04-17 | 2014-07-30 | 北京航空航天大学 | Physics-of-failure-based MOS (metal oxide semiconductor) device reliability simulation evaluation method |
| CN104166800A (en) * | 2014-08-11 | 2014-11-26 | 工业和信息化部电子第五研究所 | Component FMEA analysis method and system based on failure mechanisms |
| CN105675413B (en) * | 2016-01-18 | 2019-12-10 | 伟创力制造（珠海）有限公司 | method for evaluating stress life of circuit board |
| CN106199247A (en) * | 2016-06-28 | 2016-12-07 | 北京航空航天大学 | A kind of based on the star components and parts life assessment method of aging test data before installation |
| CN106295963A (en) * | 2016-07-29 | 2017-01-04 | 国网江苏省电力公司镇江供电公司 | Reliability assessment method for secondary system of intelligent substation based on the physics of failure |
| CN106529090B (en) * | 2016-12-12 | 2019-06-14 | 中国航天标准化研究所 | A kind of aerospace electron class Reliability Assessment method |
| CN106932708A (en) * | 2017-02-10 | 2017-07-07 | 电子科技大学 | Electronic Packaging welding spot fatigue analysis method |
| CN107247845A (en) * | 2017-06-14 | 2017-10-13 | 电子科技大学 | A kind of multi-chip module analysis method for reliability based on failure physical model |
| CN107563067A (en) * | 2017-09-06 | 2018-01-09 | 电子科技大学 | Analysis of structural reliability method based on Adaptive proxy model |
| CN108090266B (en) * | 2017-12-11 | 2021-01-26 | 电子科技大学 | Method for calculating related reliability of multiple failure modes of mechanical part |
-
2018
- 2018-06-01 CN CN201810558455.3A patent/CN108776294B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN108776294A (en) | 2018-11-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108446523B (en) | Method for evaluating and predicting storage life of electronic complete machine | |
| CN107220216B (en) | Approximate calculation method for Weibull spare part demand by using characteristic number | |
| US7675372B2 (en) | Circuit simulator parameter extraction using a configurable ring oscillator | |
| CN104182603A (en) | Reliability evaluation method for long-service-life and high-reliability electronic product | |
| CN105913166B (en) | Method for evaluating acceleration factor of storage life test of electromechanical complete machine product | |
| CN106446317A (en) | Mathematic model-based sealed relay storage life prediction method | |
| CN106408423A (en) | Method and system for risk assessment and method for constructing system for risk assessment | |
| US10936781B1 (en) | Method for setting parameters in design of printed circuit board, device employing method, and non-transitory storage medium | |
| CN103324787B (en) | Criticize and produce Satellite Thermal Design verification method | |
| US6880142B2 (en) | Method of delay calculation for variation in interconnect metal process | |
| CN108776294B (en) | Circuit board service life evaluation method based on self-adaptive strategy | |
| CN109684713B (en) | Bayesian-based complex system reliability analysis method | |
| CN110162379A (en) | Virtual machine migration method, device and computer equipment | |
| CN109657260B (en) | Turbine rotor system reliability distribution method considering failure correlation | |
| CN109491709B (en) | Software code controllability comprehensive evaluation method based on AHP and neural network | |
| CN110852016A (en) | Reliability simulation test model correction method | |
| US8645883B2 (en) | Integrated circuit simulation using fundamental and derivative circuit runs | |
| CN107357955B (en) | Electronic system parameter design method comprehensively considering manufacturing and temperature noise | |
| CN114091203A (en) | Method for crankshaft accelerated fatigue test | |
| CN109598371B (en) | Temperature prediction method and system for electronic equipment on aircraft | |
| CN109857581A (en) | A kind of method of optimum choice software reliability prediction | |
| CN111414724B (en) | Circuit simulation optimization method | |
| CN108710745A (en) | Electromechanical product reliability assessment method combined with manufacturing process data | |
| CN103714385A (en) | Analogue circuit fault diagnosis method based on improved type clone selection algorithm | |
| CN110826234B (en) | Simulation-based multi-stress accelerated life test scheme optimization method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |