CN102136017A - Random factor injection method applicable to dynamic system model - Google Patents

Abstract

The invention provides a random factor injection method applicable to a dynamic system model. The method comprises the following steps: 1, hierarchically decomposing a model to the target object of determine a random factor injection; 2, collecting random factors influencing the characteristics of the target object; 3, analyzing the collected random factors to determine a typical random factor; 4, establishing a random factor behavior model; and 5, packaging the behavior model to facilitate man-machine interaction. The invention provides a solution for random factor injection in a reliability simulation process; automatic selection and random injection can be performed in a modulation mode, thereby achieving good portability and improving the reliability of simulation. The method has good practical value and wide application project in the technical field of reliability engineering.

Description

A kind of enchancement factor method for implanting that is applicable to dynamic system model

Affiliated technical field

The present invention relates to a kind of enchancement factor (mainly referring to have the stray parameter and the fault of dynamic perfromance) method for implanting that is applicable to dynamic system model, it is a kind of enchancement factor implementation method in system reliability simulation analysis field, be devoted to the reasonable description and the automatic injection problem of enchancement factor dynamic perfromance in the resolution system realistic model, belong to the reliability engineering technical field.

Background technology

In the modern quality view, reliability becomes an important attribute of product gradually.Modern product becomes increasingly complex, and the software and hardware complexity increases day by day, makes that influencing the factor of product reliability gets more and more.The reliability of dynamic system is subjected to the combined influence of long period factors such as short period factors such as noise, interference and fault, wearing and tearing, structural parameters degeneration, and these influence factors are the dynamic process with randomness mostly.(Computer AidedDesign, CAD) software has improved product design efficiency, has shortened product development cycle extensively to adopt various computer-aided design (CAD)s in the product design process.By the products C AD model of setting up, in model, inject enchancement factor, and observe and write down of the influence of the enchancement factor of injection system with dynamic perfromance, be the important channel of evaluation and test product reliability.Therefore, how by products C AD Software tool, system model being injected the enchancement factor with dynamic perfromance, and then estimate reliability of products, is the problem that the reliability engineering specialty presses for solution.

Summary of the invention

The objective of the invention is: a kind of enchancement factor method for implanting that is applicable to dynamic system model is provided, it is based on Software tools such as CAD, in simulation process, realize the dynamic perfromance of enchancement factor and injection automatically, the various stray parameters that exist in the simulation system and the generating process of fault are to satisfy the needs of product reliability evaluation and test.

Technical scheme of the present invention:

The present invention is based on Software tools such as CAD, by setting up the behavior model of underlying component, inject the enchancement factor with dynamic perfromance automatically, described enchancement factor mainly is meant stray parameter and the fault with dynamic perfromance.

A kind of enchancement factor method for implanting that is applicable to dynamic system model of the present invention, these method concrete steps are as follows:

Step 1: destination object is determined: the research model object hierarchy, system is carried out Hiberarchy Decomposition, and form relatively independent subsystem, and determine wherein to need to carry out the system of enchancement factor injection as destination object;

Step 2: enchancement factor is collected: to the destination object of having determined, collecting influences the various stray parameters and the fault type of its reliability, and is gathered;

Step 3: enchancement factor analysis and modeling: the enchancement factor that gathers is carried out finishing analysis, and set up the mathematical model of describing the enchancement factor dynamic perfromance, its Changing Pattern is solidified.

Step 4: behavior model is set up: at destination object, by outsourcing business software (MATLAB and AMESim), establishing target object behavior model is with the mathematical model modularization of enchancement factor, to realize the automatic injection of enchancement factor in simulation process.

Step 5: behavior model encapsulation: to the destination object behavior realistic model of setting up, graphic user interface (the Graphic User Interface that utilizes business software (MATLAB and AMESim) to carry, GUI) encapsulate, and correlation parameter is set, be user-friendly to.

Wherein, be meant the object CAD software model of having set up at the model object described in the step 1, correspondence is " * .mdl " type file that Software tool generates among the MATLAB, and correspondence is " * .ame " type file that Software tool generates among the AMESim.Independent subsystem described in the step 1 is meant that the module of division is independent as far as possible on function and structure, can conveniently describe or modularization, has portability and maintainability preferably like this.Destination object in the step 1 requires level low more good more, and high-level fault all is that the fault by low level causes that level is low more, and modeling is easy more.For baroque subsystem, further refinement, wherein assembly mould blocking.

Wherein, comprise stray parameter with dynamic perfromance and fault two classes that occur in system's operational process in the enchancement factor described in the step 2.The main path that enchancement factor described in the step 2 is collected is: for newly grinding product, mainly search historical data, the data of domestic and international like product, its all problems that exists under different situations, condition of statistical summaries; For improving product, mainly search, sum up this product historical data, all problems that statistics exists under the different operating situation.

Wherein, the enchancement factor analysis described in the step 3 be meant from the data that gathers, find out to each destination object is influential all may enchancement factor, and the effective criterion of clear and definite enchancement factor is rejected invalid enchancement factor.In the mathematical model of the description enchancement factor dynamic perfromance described in the step 3, have different the description according to the difference of enchancement factor type.For the parameter Changing Pattern, can be described with stochastic process; For the fault Changing Pattern, can be described with Markovian process.

Wherein, at the behavior model described in the step 4, also different according to the not isostructure of enchancement factor type:

For fault, there are three chief components behavior model inside, and theory diagram as shown in Figure 1.Three ingredients are respectively: the normal function unit, in order to describe destination object physical characteristics in normal working conditions; The fault characteristic unit, the physical characteristics when being in the specified fault mode state in order to describe destination object; Switch unit is in order to describe the dynamic perfromance of fault.In single artificial tasks process, switch unit lost efficacy according to destination object and restrained the life time of randomly drawing destination object, extracted current simulation time in each step-length then, dynamically judged whether to break down.If in artificial tasks, break down,, select the fault mode that takes place at random and inject realistic model, with the dynamic generating process of simulated failure then according to the transition probability of each fault mode.The mode failures even rate of supposing every kind of fault mode is respectively f1, f2...fn, and state can only transfer to a certain fault mode by normal condition, and then every kind of fault mode transition probability is obtained by following formula:

$\left\{\begin{array}{c}{\mathrm{\&lambda;}}_1=\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="HDA0000050599430000011.png"\; state="\{\{state\}\}">f</figure-callout>}1/(\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="HDA0000050599430000011.png"\; state="\{\{state\}\}">f</figure-callout>}1+\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="HDA0000050599430000011.png"\; state="\{\{state\}\}">f</figure-callout>}2+\&CenterDot;\&CenterDot;\&CenterDot;+\mathrm{fn})\\ {\mathrm{\&lambda;}}_2=\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="HDA0000050599430000011.png"\; state="\{\{state\}\}">f</figure-callout>}2/(\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="HDA0000050599430000011.png"\; state="\{\{state\}\}">f</figure-callout>}1+\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="HDA0000050599430000011.png"\; state="\{\{state\}\}">f</figure-callout>}2+\&CenterDot;\&CenterDot;\&CenterDot;+\mathrm{fn})\\ \&CenterDot;\\ \&CenterDot;\\ \&CenterDot;\\ {\mathrm{\&lambda;}}_n=\mathrm{fn}/(\mathrm{<figure-callout\; id="1"\; label="f"\; filenames="HDA0000050599430000011.png"\; state="\{\{state\}\}">f</figure-callout>}1+\mathrm{<figure-callout\; id="2"\; label="f"\; filenames="HDA0000050599430000011.png"\; state="\{\{state\}\}">f</figure-callout>}2+\&CenterDot;\&CenterDot;\&CenterDot;+\mathrm{fn})\end{array}\right.$

Suppose the corresponding incident of each fault mode, then the relative failure probability of each incident is respectively λ ₁, λ ₂..., λ _nIt is 1 straight line that imagination is got a length, and is divided into n interval, and each burst length is respectively λ ₁, λ ₂..., λ _nExtracting random number η (significant figure are Duoed at least than the mode failures even rate) between the 0-1, is 1 straight line drop point to this length, when it falls into interval λ _j-λ _J+1In time, regarded j kind fault mode as and takes place.Principle of work as shown in Figure 2.

For stray parameter, there are two chief components behavior model inside, and theory diagram as shown in Figure 3.Two ingredients are respectively: the normal function unit, in order to describe the characteristic of product in normal output; The dynamic perfromance unit is in order to describe the characteristic of stray parameter dynamic change.Each step-length in simulation process, the dynamic perfromance unit is randomly drawed the parameter value that meets this rule, and is injected in the realistic model according to the mathematical model of the description stray parameter dynamic perfromance that obtains in the step 3, with the generating process of simulation stray parameter.

Wherein, described in the step 5 the stray parameter that correlation parameter is meant that performance parameter that the characterization system operate as normal is set and characterization system have and the characterisitic parameter of fault be set.

The invention provides a kind of enchancement factor method for implanting based on CAD software, its advantage mainly contains:

(1) can in simulation process, simulate the various stray parameters with dynamic perfromance and the generating process of fault, increase the credibility of emulation;

Adopt layered modeling to inject when (2) enchancement factor is injected, separate between the model, have portability and maintainability preferably.

(3) model is easy to use, and carrying out does not need to understand inner structure when enchancement factor is injected, and only need carry out the parameter setting by dialog box.

Description of drawings

Fig. 1 is a fault behavior model framework chart.

Fig. 2 is a fault mode Dynamic Selection process flow diagram.

Fig. 3 is the stochastic parameter model block diagram.

Fig. 4 is certain type steering wheel structural representation.

Fig. 5 is sensor fault behavior model figure.

Fig. 6 is a direct current generator mathematical model schematic diagram.

Fig. 7 is parameter fluctuation behavior model figure.

Fig. 8 is the FB(flow block) of the method for the invention.Symbol description among the figure:

FR1 is permanent gain fault mode crash rate

FR2 is permanent deviation fault mode failures even rate

FR3 is stuck fault mode crash rate

NG is a sensor normal gain value

GC is the permanent gain coefficient value of sensor

B is the permanent deviation constant value of sensor

ST is stuck value

R _CBe the armature resistance value

L _CBe the armature inductance coefficient

C _tBe torque constant

I _aBe armature supply

T is an electromagnetic torque

J is a moment of inertia

K is a ratio of damping

C _eBe back electromotive force constant

ω is an angular speed

θ _cBe angle

MV is a mean parameter

V is a parameter variance

Embodiment

Below in conjunction with concrete case study on implementation, the enchancement factor method for implanting that is applicable to dynamic system based on CAD software of the present invention is elaborated.

Case 1: steering gear system

The present invention is an example with certain type steering wheel, and the fault filling method that has dynamic perfromance in the dynamic system model is described.Fig. 4 has shown the structural representation of certain type steering wheel.

See Fig. 8, a kind of enchancement factor method for implanting that is applicable to dynamic system model, these method concrete steps are as follows:

Step 1: destination object is determined: steering wheel is the important component part of flight control system, can be divided into controller, power amplifier, direct current generator, reducer casing, sensor by function.Wherein, sensor is main inoperative component, elects destination object as at this.

Step 2: enchancement factor is collected: collect the enchancement factor that influences sensor performance by books and network etc. and mainly contain the variation of sensor output gain value, parameter slow drift variation, the remarkable sudden change of parameter etc.

Step 3: enchancement factor analysis and modeling: the factor of collecting is carried out induction-arrangement, and the persevering gain of definite enchancement factor was lost efficacy: sensor output increases the constant ratio coefficient; Permanent deviation lost efficacy: sensor output increases the constant deviation value; Stuck inefficacy: sensor no-output.Its mathematical model is described as:

Wherein, y _OutBe the actual output of sensor value, y _InBe the input value of sensor, a _iBe constant, β is permanent change in gain scale-up factor, and Δ is that permanent deviation changes departure.On the mathematical model basis, structure behavior model block diagram, as shown in Figure 1.

Step 4: behavior model is set up: the behavior model of foundation as shown in Figure 5: the physical characteristics that passage 1 is exported for operate as normal; Passage 2 is the permanent gain fault mode physical characteristics of output down; The physical characteristics of passage 3 for exporting under the permanent deviation fault pattern; Passage 4 is the physical characteristics of output under the stuck fault mode.Switch unit adopts change-over switch, and the S-function that carries by MATLAB carries out the dynamic judgement and the selection of fault mode, the fault generating process that has dynamic perfromance with simulation, as shown in Figure 2.

Step 5: the behavior model encapsulation: the sensor behavior model is carried out self-defined encapsulation, and the parameter of encapsulation is as shown in table 1.

Table 1 sensor fault behavior model parameter encapsulation tabulation

Sequence number	Parameter name	Implication
				1	Normal?Gain	The normal gain value
2	Gain?Change	Permanent yield value
			3	Gain?Change?Failure?rate	Permanent gain mode crash rate
4	Bias	Permanent deviate

5	Bias?Failure?rate	Lateral deviation mode failures even rate
			6	Stuck	Permanent stuck stale value (0)
7	Stuck?Failure?rate	Permanent stuck mode failures even rate

Case 2: direct current motor

The present invention's direct current motor is an example, and the stray parameter implantation step is described.Direct current generator realistic model structure as shown in Figure 6.

Step 1: destination object is determined: direct current generator is a kind of device of widespread use, and its realistic model can be divided stationary part and rotor portion by functional structure.Major parameter has armature resistance, armature inductance, torque constant, moment of inertia, inverse electromotive force constant etc.Wherein, armature resistance is subject to influence such as temperature and dynamic random takes place changes in motor operation course, elect destination object as at this.

Step 2: enchancement factor is collected: collect armature resistance normal phenomenon resistance dynamic change that occurs in operational process by books and network etc., resistance burns etc.

Step 3: enchancement factor analysis and modeling: research object is carried out analysis and arrangement, determine armature resistance resistance dynamic change in operational process, and the Changing Pattern Normal Distribution.Its mathematical model is described as:

y _out(t)＝t _n(t)·σ+μ

Wherein, y _OutBe stray parameter value, t _nFor obeying the random sampling value of standardized normal distribution, μ is the normal distribution average, and σ is the normal distribution standard deviation.

Step 4: behavior model is set up: the behavior model of foundation as shown in Figure 7.In the dynamic perfromance unit, S-function coding with MATLAB carries carries out normalize to stray parameter and handles, and all produces one in each step-length and obeys the numerical value of this normal distribution and inject realistic model, with the generating process of simulation stray parameter, as shown in Figure 3.

Step 5: the behavior model encapsulation: the armature resistance behavior model is carried out self-defined encapsulation, and the parameter of encapsulation is as shown in table 2.

Table 2 parameter fluctuation behavior model parameter encapsulation tabulation

Sequence number	Parameter name	Meaning of parameters
			1	Mean?Value	Mean parameter
2	Var?iance	Parameter variance

Claims (6)

1. an enchancement factor method for implanting that is applicable to dynamic system model is characterized in that, this method is carried out according to following 5 steps:

Step 1: destination object is determined: the research model object hierarchy, system is carried out Hiberarchy Decomposition, and form relatively independent subsystem, and determine wherein to need to carry out the system of enchancement factor injection as destination object;

Step 2: enchancement factor is collected: to the destination object of having determined, collecting influences the various stray parameters and the fault type of its reliability, and is gathered;

Step 3: enchancement factor analysis and modeling: the enchancement factor that gathers is carried out finishing analysis, and set up the mathematical model of describing the enchancement factor dynamic perfromance, its Changing Pattern is solidified;

Step 4: behavior model is set up: at destination object, by outsourcing business software MATLAB and AMESim, establishing target object behavior model is with the mathematical model modularization of enchancement factor, to realize the automatic injection of enchancement factor in simulation process;

Step 5: behavior model encapsulation: to the destination object behavior realistic model of setting up, utilizing business software is that the graphic user interface that MATLAB and AMESim carry is Graphic User Interface, GUI encapsulates, and correlation parameter is set, and is user-friendly to.

2. a kind of enchancement factor method for implanting that is applicable to dynamic system model according to claim 1, it is characterized in that: be meant the object CAD software model of having set up at the model object described in the step 1, correspondence is " * .mdl " type file that Software tool generates among the MATLAB, and correspondence is " * .ame " type file that Software tool generates among the AMESim; The module that independent subsystem described in the step 1 is meant division is independent as far as possible on function and structure, can conveniently describe, modularization, has portability and maintainability preferably like this; Destination object in the step 1 requires level low more good more, and level is low more, and modeling is easy more; For baroque subsystem, further refinement, wherein assembly mould blocking.

3. a kind of enchancement factor method for implanting that is applicable to dynamic system model according to claim 1 is characterized in that: comprise stray parameter with dynamic perfromance and fault two classes that occur in system's operational process in the enchancement factor described in the step 2; The approach that enchancement factor described in the step 2 is collected is: for newly grinding product, search historical data, the data of domestic and international like product, its all problems that exists under different situations, condition of statistical summaries; For improving product, search, sum up this product historical data, all problems that statistics exists under the different operating situation.

4. a kind of enchancement factor method for implanting that is applicable to dynamic system model according to claim 1, it is characterized in that: the enchancement factor analysis described in the step 3 be meant from the data that gathers, find out to each destination object is influential all may enchancement factor, and the effective criterion of clear and definite enchancement factor, reject invalid enchancement factor; In the mathematical model of the description enchancement factor dynamic perfromance described in the step 3, have different the description according to the difference of enchancement factor type; For the parameter Changing Pattern, be described with stochastic process; For the fault Changing Pattern, be described with Markovian process.

5. a kind of enchancement factor method for implanting that is applicable to dynamic system model according to claim 1 is characterized in that: at the behavior model described in the step 4, also different according to the not isostructure of enchancement factor type:

For fault, there are three ingredients behavior model inside, is respectively: the normal function unit, in order to describe destination object physical characteristics in normal working conditions; The fault characteristic unit, the physical characteristics when being in the specified fault mode state in order to describe destination object; Switch unit is in order to describe the dynamic perfromance of fault; In single artificial tasks process, switch unit lost efficacy according to destination object and restrained the life time of randomly drawing destination object, extracted current simulation time in each step-length then, dynamically judged whether to break down; If in artificial tasks, break down,, select the fault mode that takes place at random and inject realistic model, with the dynamic generating process of simulated failure then according to the transition probability of each fault mode; If the mode failures even rate of every kind of fault mode is respectively f1, f2...fn, and state can only transfer to a certain fault mode by normal condition, and then every kind of fault mode transition probability is obtained by following formula:

$\left\{\begin{array}{c}{\mathrm{\&lambda;}}_1=f1/(f1+f2+\&CenterDot;\&CenterDot;\&CenterDot;+\mathrm{fn})\\ {\mathrm{\&lambda;}}_2=f2/(f1+f2+\&CenterDot;\&CenterDot;\&CenterDot;+\mathrm{fn})\\ \&CenterDot;\\ \&CenterDot;\\ \&CenterDot;\\ {\mathrm{\&lambda;}}_n=\mathrm{fn}/(f1+f2+\&CenterDot;\&CenterDot;\&CenterDot;+\mathrm{fn})\end{array}\right.$

If the corresponding incident of each fault mode, then the relative failure probability of each incident is respectively λ ₁, λ ₂..., λ _nIt is 1 straight line that imagination is got a length, and is divided into n interval, and each burst length is respectively λ ₁, λ ₂..., λ _nExtracting random number η between the 0-1, is 1 straight line drop point to this length, when it falls into interval λ _j-λ _J+1In time, regarded j kind fault mode as and takes place; For stray parameter, there are two ingredients behavior model inside, is respectively: the normal function unit, in order to describe the characteristic of product in normal output; The dynamic perfromance unit is in order to describe the characteristic of stray parameter dynamic change; Each step-length in simulation process, the dynamic perfromance unit is randomly drawed the parameter value that meets this rule, and is injected in the realistic model according to the mathematical model of the description stray parameter dynamic perfromance that obtains in the step 3, with the generating process of simulation stray parameter.

6. a kind of enchancement factor method for implanting that is applicable to dynamic system model according to claim 1 is characterized in that: described in the step 5 the stray parameter that correlation parameter is meant that performance parameter that the characterization system operate as normal is set and characterization system have and the characterisitic parameter of fault be set.

Images