CN108496168A - The enhancing of dummy model - Google Patents

Abstract

Disclose a kind of method for automatically generating the enhancing model of physical assemblies.This method includes：Input model is read in processor, the nominal operation pattern for the physical assemblies that input model description is modeled by input model；Input model is parsed to generate its parsing using processor；The parsing of input model is analyzed using processor；And the enhancing model of physical assemblies is automatically written from input model based on processor described in the analysis and utilization, at least one replacement operation pattern different from nominal operation pattern for the physical assemblies which describes the nominal operation pattern for the physical assemblies that (i) is modeled and (ii) is modeled.Suitably, it is analyzed according at least one of first method and second method method.First method is used to detect whether input model is influenced by one or more failures, including：(i) catastrophic, (ii) flow of power, and (iii) parameter type failure；Particular aspects of the second method for the dynamic behaviour of matching component, and these suitably modified matched aspects are with the dynamic of faults mechanism.

Description

The enhancing of dummy model

The U.S. Provisional Patent Application No.62/273,719 submitted this application claims on December 31st, 2015 and 2015 12 The U.S. Provisional Patent Application No.62/273 that the moon is submitted on the 31st, 767 priority are whole by quoting by above-mentioned two applications It is incorporated into this.

Background technology

Subject of the present invention relates generally to network physical component and/or system modelling fields.In conjunction with object-oriented and/or Such as other similar Modeling Calculation machine language of Modelica etc. find specific but non-exclusive correlation.Therefore this explanation Book sometimes specifically quotes it.It should be appreciated, however, that be applied equally to other similar for the various aspects of present subject matter Application and/or environment.

The modeling language of such as Modelica is commonly used to generate the physical set for design and analysis and/or test purpose The dummy model of part and/or system and/or it is simulated.For example, the physical system and/or component that are modeled can be suitable It is complicated or comparatively simple, and may include such as circuit or motor vehicles or other physical systems.For example, the Modelica that increases income Java standard library (MSL) is in the model that such as domain of electronics, machinery and fluid includes about 1280 components.Network physical component is established In or dependent on calculate and physical assemblies synergistic effect, such as found in vehicle, assembly line and Other Engineering system hard The system controller of control software is realized on part platform.For example, the libraries Modelica_LinearSystems2 include continuous and discrete The model of controller.MSL itself includes the model for ratio, the different variants of integral and difference (PID) controller.These Modelica models etc. are widely used in various projects more and more.In the disclosure, term " physical assemblies " is used for table Show any of " physics " and " network physical component " or both.

However, in fact, the usually description such as MSL models and/or only expression are modeled a kind of operation mode of component, i.e. quilt The nominal or correct behavior of the component of modeling.That is, these models will not usually describe mistake or other non-nominal behaviors. For example, fault mode can be short circuit or open circuit, or the motor that can not be run or slipper brake etc..Any amount of difference Different mechanism (for example, abrasion, fatigue of materials, corrosion etc.) may lead to the failed operation for being modeled component.Usually, it may be desirable to survey It tries and/or analyzes in fault mode or the component of other non-nominal mode operations, such as diagnostic purpose.However, due to normal Rule MSL models etc. do not describe such failure and/or other non-nominal modes usually, thus in such failure and/or The diagnosis for being modeled component run under other non-nominal modes and/or other tests may be problematic.

In addition, damage process model is indicated since its service environment is as the parameter such as load, material properties and size Function and the damage caused by the inefficacy mechanism that is in active state in system component.In specific system configuration and use Under the conditions of operational reliability simulation may excessively take and/or need a large amount of manpower.

It is Saha et al. in the U.S. Patent application submitted on the 15th of August in 2013 with the relevant work of this specification No.13/967,503, entire contents are clearly incorporated to herein by reference.

On the whole, it is disclosed that for automatic and/or semi-automatically generation include the component of non-nominal behavior pattern The new and/or improved method and/or system or device of model.In addition, it also allows to study component faults in system level Influence, be possibly used for improving the system design about robustness, elasticity and reliability.

Invention content

It is to introduce and the relevant concept of present subject matter to provide the brief description.The content of present invention is not intended to identify The essential feature of theme claimed is intended to be used to determine or limit the range of theme claimed.Below The embodiment of description is not intended to exhaustion or limits the invention to precise forms disclosed in described in detail below.But it selects Others skilled in the art are made to may recognize that and understand principle and the practice of present subject matter with these embodiments are described.

According to one embodiment, a kind of method for automatically generating the enhancing model of physical assemblies is provided.This method Including：Input model is read in processor, the nominal behaviour for the physical assemblies that input model description is modeled by input model Operation mode；Input model is parsed to generate its parsing using processor；The parsing of input model is analyzed using processor；And base The enhancing model of the physical assemblies from input model is automatically written in the processor described in the analysis and utilization, which retouches It states the nominal operation pattern for the physical assemblies that (i) is modeled and physical assemblies that (ii) is modeled is different from nominal operation pattern At least one replacement operation pattern.Suitably, divided according at least one of first method and second method method Analysis.First method is used to detect whether input model is influenced by one or more failures, including：(i) catastrophic, (ii) Flow of power (power flow), and (iii) parameter type failure；Second method is used for the specific of the dynamic behaviour of matching component Aspect, and these suitably modified matched aspects are with the dynamic of faults mechanism.

According to another embodiment, a kind of system for automatically generating the enhancing model of physical assemblies is provided.The system Including：Processor is used for：Input model is read, which describes the nominal of the physical assemblies modeled by input model Operation mode；The input model is parsed to generate its parsing；Analyze the parsing of input model；And it is automatically write based on the analysis Enter the enhancing model of the physical assemblies from input model, the nominal behaviour for the physical assemblies that enhancing model description (i) is modeled At least one replacement operation pattern different from nominal operation pattern for the physical assemblies that operation mode and (ii) are modeled.Properly Parsing is analyzed in ground according at least one of first method and second method method.First method is for detecting input model Whether influenced by one or more failures, including：(i) catastrophic, (ii) flow of power, and (iii) parameter type event Barrier；Particular aspects of the second method for the dynamic behaviour of matching component, and these suitably modified matched aspects are to reflect event The dynamic of barrier mechanism.

After reading and understanding this specification, many advantages and benefit of subject matter disclosed herein are general for this field Logical technical staff will be apparent.It should be understood, however, that while indicating preferred and/or other embodiment, various realities The detailed description for applying example and specific example provides by way of illustration, not of limitation.In the situation for the spirit for not departing from the present invention Under, many changes and modifications can be made within the scope of the invention, and the present invention includes all such modifications.

Description of the drawings

Fig. 1 is the schematic diagram for showing to be suitable for putting into practice the exemplary model converting system of the various aspects of present subject matter.

Fig. 2 is the stream for the illustrative methods for executing model conversion processing for showing various aspects according to the inventive subject matter Cheng Tu.

Fig. 3 shows that the exemplary workflow of the various aspects according to the inventive subject matter using MBD analyses, MBD therein are divided It analyses for determining that the task of depending on uses the component fatigue life with design configurations.

Fig. 4 shows that the graphical representation of simulation test route (course), path height (are single with inch along vertical axis Position) and route distance along trunnion axis (in inches).

Fig. 5 A and 5B show that groove intersects the sample acceleration time history with 1 meter of wall depression of order.

Fig. 6 shows the example von mises stress profile of exemplary 4 × 4 × 4 holder.

The holder that the function as speed and stiffness of support is shown respectively in Fig. 7, Fig. 8 and Fig. 9 is maximum in the x, y, and z directions Acceleration.

The maximum that Figure 10, Figure 11 and Figure 12 are shown as the function of car speed and electric power assembly mass ratio coefficient is horizontal To, the percentage error of longitudinal direction and normal acceleration.

Specific implementation mode

In order to which clear and for the sake of simplicity, this specification should in the case where configuration or operation are not described in further detail Refer to structure and/or function element, relevant criterion, algorithm and/or agreement and other assemblies as known in the art, calculation Method, method and/or process, in addition to they according to and/or adapt to the preferred and/or other embodiments that have presented herein and repaiied Except the degree for changing or changing.In addition, by example and device and method disclosed in this specification are described in detail with reference to the drawings. The similar number in attached drawing is indicated through attached drawing for identical, similar or respective element reference unless otherwise specified,.It should recognize Know, for specific application, for disclosed and description example, arrangement, configuration, component, element, device, method, material Deng modification be that can carry out and may be desired.In the disclosure, any mark of specific material, technology, arrangement etc. Either general description related to the specific example presented or only this material, technology, arrangement etc..Detail Or it is exemplary mark be not intended to and be not construed as it is enforceable or restrictive, unless specifying in this way.With It is lower to be described in detail with reference to the attached drawings with disclosed device and method by selection example.

In general, according at least one embodiment, it is disclosed that a kind of operation for increasing with failure and/or other are non- Method, system and/or the device of the model of the physical system component of nominal mode.Suitably, automatically or half from component model The model including failure and/or other non-nominal modes is automatically generated, otherwise these component models only describe the nominal of component Or correct behavior.Note that as used herein, FAME is that (failure enhances mould to Fault-Augmented Model Extension Type extend) acronym.It should be appreciated, however, that enhancing model can be extended to as the described herein including Strictly speaking be not " failure " it is relevant but with other non-nominal or other relevant alternative patterns of replacement operation pattern.

According to a suitable embodiment, existing library of increasing income (such as Modelica java standard libraries (MSL), opened by third party The library of hair and other libraries) it is extensible to allow compact introducing fault model, while protecting when failure is not introduced into model The availability of existing nominal behavioural norm when particular instance.More specifically, for example MSL, third party develop library and/or other In similar library, the basic set for the non-nominal behavior model for expanding to nominal component model library is created.This method allows event Hinder model and is inserted into the New Parent created by various third parties and/or other libraries Modelica developer.

The processing of FAME model conversions is using JModelica Modelica resolvers frames and builds its JastAdd technologies By direct fault location to nominal component model library.The java applet for merging JastAdd and JModelica operates in provided library On, it identifies susceptible component model, suitably they is re-write to provide failure behavior, and will be modified Library be output to new position.

Here is that a simple model class of capacitor is used in Modelica, is easily opened a way (ElectricalOpen) and the influence of short-circuit (ElectricalShort) failure.

Model Capacitor " Ideal linear electrical capacitor (ideal linearity capacitor) "

extends Modelica.Electrical.Analog.Interfaces.OnePort；

Parameter Modelica.SIunits.Capacitance C (start=1) " Capacitance "；

equation

I=C*der (v)；

end Capacitor；

Aforementioned is, for example, non-reinforced or simple capacitor model.

In order to analyze the above situation, suitable embodiment is used to detecting or determining it is easy the shadow by which failure It rings, and it is written over so that these failures can be simulated under such example.In general, exemplary embodiment For searching three classes failure：(i) catastrophic, (ii) flow of power, and (iii) are parametric.Catastrophic failure would generally be by group The dynamic of part becomes entirely different component.For example, in power domain, common catastrophic failure is open circuit.Flow of power failure Influence flows in or out one or more flow of power of component.For example, in electric component, common flow of power failure is certain The short circuit of degree.Parametric failure usually reflects the offset for assuming fixed parameter.For example, in the case of capacitor, capacitance.

Once identifying the susceptible failure of component, behavior description is just added to this for the failure each identified Component.Then model may operate under any pattern in these fault modes, and for non-catastrophic failure, can refer to Determine the severity of failure.

About direct fault location, a kind of suitable method is dependent on understanding that flow of power passes through with the system of these component constructions Pattern (pattern), and have the advantages that using being quite abstracted specified model and to be faced with modeling inconsistent.This side Method is described below as " outside " method.Suitably, externalist methodology can utilize the desired parametic fault being automatically injected The manual compiler listing of statement formula specification is enhanced.This allows more specific and is accurately modeled to failure.

Another suitable method depends on the particular aspects of matching component dynamic behaviour, and these suitably modified aspects are with anti- Reflect the dynamic of fault mechanism.With externalist methodology on the contrary, this method is referred to as " inside " method.It can provide failure behavior Modeling in more detail, but be more difficult to accurately specified.

JModelica is a Modelica tools chain increased income.JModelica by Modelica language Java/ JastAdd resolvers and the various simulations with both Java and Python write-in and analysis tool composition.

JastAdd is the realization based on Java of Cnut attribute grammar (Knuth'attribute grammars), Node in middle abstract syntax tree (AST) can be with the attribute that value is defined by equation.Each attribute can be synthesis (by Be attached to node itself equation definition), or inherit (being defined by the equation in ancestor node).JastAdd is supported Reference (reference) attribute is the attribute that another node has its value in AST.This allows to weave by AST (woven) arbitrary graphic.It also allows its attribute parameterized, it means that when assessing formula, formula, which has, is tied to parameter Unbound variable.It also allows collection (collection) attribute；Multi-valued attribute, value can be by various other AST nodes Contribution.It supports circular reference attribute, it provides the mode for breaking reference cycle.

JastAdd is based on Java, so equation is implemented as Java method, attribute access is via these sides of calling Method.It also extends Java itself, especially aspect (aspect) using various constructions.The type of AST classes is supported in terms of JastAdd Between state (intertype declarations).Statement occurs from a statement in one side file between type, but It actually belongs to an AST class (such as attribute equation).JastAdd systems reading aspect file simultaneously will be stated to compile between these types It knits in suitable AST classes.(.jrag files) and commanding aspect (.jadd files) in terms of its support statement.Declaratively side Add new attribute, equation and rewriting in face；Commanding aspect only adds Java method and variable.Rewriteeing AST can also be designated； Optionally only when some condition C is true, they replace the AST nodes of type A with the node of type B.New AST subtrees can be with It is created and is designated as attribute value (" nonterminal attribute ").

In one suitable embodiment, so-called externalist methodology, JModelica resolvers are for parsing each library text Part, then check analytic tree in find each model class (with limited model non-part Modelica classes define) for With the example of lower connector class：

Modelica.Electrical.Analog.Interfaces.Pin

Modelica.Mechanics.Rotational.Interfaces.Flange_a

Modelica.Mechanics.Rotational.Interfaces.Flange_b

Modelica.Mechanics.Translational.Interfaces.Flange_a

Modelica.Mechanics.Translational.Interfaces.Flange_b

Modelica.Fluid.Interfaces.FluidPort

Modelica.Thermal.HeatTransfer.Interfaces.HeatPort

These are the examples of Modelica connector classes, it includes two variables, one be as possible type (for example, voltage Or pressure), the other is stream type (for example, electric current or mass velocity).Analysis is combined as suitable in the direct component of model class, with And the component of each derived class that model class uses.If finding any component in said modules, replacing for model class is created The parsing tree of class is changed, and the parsing tree of model class is replaced with to new parsing tree.

Finally, be used to re-create the present of model from AST in terms of JModelica " FormattedPrettyPrint " The enabled Modelica source codes of failure.

In another exemplary embodiment (that is, internalist methodology), the more general of offer such as " Fault (failure) " is utilized Each general fault or alternative patterns are modeled in terms of the subclass of class, such as " electric short circuit ".Each Fault or alternative patterns Class all provides a predicate (predicate), such as：

static boolean may_occur_with(FullClassDecl klass)；

It gives a FulIClassDecl (node that Modelica classes define is represented in JModelica AST), says Whether bright FullClassDecl, which it may be easily affected by this failure, influences (or alternative patterns may be applicable in).

Suitably, the realization of " may_occur_with " predicate is depended on as given by the FulIClassDecl nodes of AST Solution to model analysis matched with the normal form that defines of failure/alternative patterns is directed to.

In addition, if FulIClassDecl is easily affected by it, each Fault or alternative patterns subclass are suitably by it The example of itself contributes to " possibleFaults " aggregate attribute on FulIClassDecl, such as uses JastAdd " contributes (contribution) " mechanism.For example, for Fault subclass ElectricalShort, it may be as follows：

FullClassDecl contributes(ElectricalShort(this))

when(ElectricalShort.may_occur_with(this))

to FullClassDecl.possibleFaults()

for this；

Indeed, it is possible to lay equal stress on new compiler in terms of by the set addition in terms of failure to add more events Barrier/alternative patterns.

Suitably, in order to inject failure and/or alternative patterns, JModelica FullClassDecl node types are expanded Exhibition be include many adeditive attributes, such as follows：

It is then possible to re-write the AST of representative model using " rewriting " function of such as JastAdd --- Failure behavior is added to the model by FullClassDecl nodes.For example, rewriting rule may be as follows：

Suitably, this rewriting can be as the automatic application of a part of dissection process, and in appropriate circumstances, this will lead Failure/replacement clause is caused to be injected into solution to model analysis.In fact, equation rewrite method RewriteEquationsToAddFaults () uses the nominal equation of model, and they are packaged in the situation In one branch of Modelica if- equations, additional branch is then added to each possible failure/alternative patterns The equation of if- equations and the failure/alternative patterns.In one suitable embodiment, those failures/replacement equation can be with It is obtained by calling Fault and/or replacing the method on subclass example.In Modelica, if- equations are realization condition sides The common method of journey.

Finally, can for the use of such as JModelica " FormattedPrettyPrint " and/or function be used for from AST re-creates the Modelica source codes that the present failure of model enables.

For example, in the case of either method, in one suitable embodiment, it is used for failure and/or alternative patterns The global procedures being injected into model can be analogous to following (in the form of pseudocode)：

for file in recurseOver(libraryTree):

OutputFile=figureOutputName (outputDir, file)

ParsedVersion=parseFile (file)

prettyPrint(parsedVersion,outputFile)

Once being added to failure and/or alternative patterns, so that it may assess its effect to use common Modelica to simulate. In fact, enhancing model can be used for various diagnostic applications.Suitably, all enhancing models are all written with standard Modelica. Therefore, initial value, parameter and failure/alternative patterns may be located in Modelica packagings.For example, for designing and diagnosing Some purposes of the new model of purpose include but not limited to：

Give one group of primary condition, parameter value and failure, it is determined whether meet system requirements.

Give one group of primary condition, failure and component age, the conditional probability that the system of determination is met the requirements.

One group of primary condition, the age of failure, observed value and all components other than one of component are given, Determine the age of remaining components.

Diagnosis based on master pattern.

Failure mode and effect analysis (FMEA).

Fault mode and the key analysis of influence.

It is maximum to determine which component damage whether meet the requirements influence to system.

Determine MTTF.

Suitably, including each model class of failure, which defines, is replaced by new class definition, that is, it is incorporated to (subsum) Archetype class and add statement sexual behaviour to allow the Modelica model classes of simulated failure.

If it find that class model is easily influenced by one or more failures, then such will be rewritten.The enumeration type quilt of encapsulation Definition, lists the various fault modes and " nominal " pattern of class.The discrete mode parameter of this new type is defined, definition The pattern that such example is currently running.One if- equation is added, so that each operation mode can define oneself Dynamic.For example, this may be as follows：

Suitable for the appropriate branch that the equation group of each fault mode is indicated on aforementioned if- equations.Due to operating mould Formula type is Modelica parameters, and selected branch will not change during simulation, and compiler can optimize the equation.

Suitably, each new class is connected by not consuming the increased variable power consumption component of power under nominal mode The example of its power interface component.For example, if archetype class is electronic building brick, and power interface example is class Pin Example, then power consuming components appropriate by be FAME.DynamicDampers.Electrical example, wherein damping parameter sets It is set to 0.

In addition, if new class includes multiple power interfaces of same type, it may also include to connect each pair of compatibility The variable power conducting subassembly of connector assembly, is nominally set as non-conducting electric power.If the type of connector assembly is identical, Or being inherited from same type, then they are " compatibilities ".For example, if archetype class is to include connector type The electronic building brick of the example " p " of PositivePin and the example " n " of connector type NegativePin, both Pin Subtype will then have the example for connecting the two examples FAME.Bridges.Electrical, wherein setting bridge joint amount to The very small value of Modelica.Constants.eps.

The process can also suitably flatten the superclass of model into the class of rewriting, and it is visible to introduce two new outsides Component FAME_operating_mode and FAME_fault_amount and component FAME_OperatingModes is given The enumeration type of possible breakdown.

The failure of electric power throat floater is shown as by being simulated to the simple change of the two variables.For example, as it appears from the above, It can be by setting FAME_operating_mode to FAME_OperatingModes.Electrical_Short, and incite somebody to action FAME_fault_amount is set as 1 to be modeled to electric short circuit.

For example, the above-mentioned non-reinforced or simple capacitor according to the execution of externalist methodology becomes like following thing (that is, enhancing capacitor model)：

In one alternate embodiment, which reads in parametic fault table, such as it is often gone to describe and operates the phase in component Between assume the specific fault that fixed parameter changes.For each failure, this will provide fault mode, Modelica classes, The function of the special parameter component of class, the Modelica types of parameter and description as variable FAME_fault_amount The Modelica expression formulas of Parameters variation.Suitably, parametic fault by introduce prefix be " FAME_ " new continuous variable come into Row processing.The calculated value of function that the variable is set as being specified by bug list by addition equation.Therefore, initial parameter is drawn With the expression formula for being replaced by the reference new variables.

Pattern in the dynamic that so-called internalist methodology passes through detection model class works, this shows it to specific fault Sensibility.In the case where detecting this sensibility, describing the dynamic equation of the model will be to allow to carry out the failure The mode of modeling is rewritten.

In one suitable embodiment, for each failure and/or alternative patterns, sensibility/applicability normal form is referred to It is fixed, it indicates that such can be showed when being matched by model class and is out of order.In fact, these patterns can utilize language appropriate Say how mutual the component type and these components that the language description class may have be suitably to define.For example, in order to retouch The most common failure of brake sliding or clutch caused by surface abrasion is stated, the torque by clutch or brake is caused Loss, the system find the model class with two variables for indicating torque, and a variable in two of which variable is another The product of a variable and the other variable for representing component of friction.For example, this may be as follows：

Component (tau1, type=Modelica.SIunits.Torque)

Component (tau2, type=Modelica.SIunits.Torque)

Component (friction, type=Modelica.SIunits.CoefficientOfFriction)

equation(eq1,tau1,product(tau2,friction))

Wherein " tau1 " indicates that the first torque variable, " tau2 " indicate that the second torque variable, " friction " indicate friction The component of variable, " eq1 " indicate foregoing variables being mutually related equation.

Therefore, which suitably thinks that any model class for meeting this four constraints is likely to be subjected to this sliding event Barrier pattern.That is, when detecting above-mentioned constraint in model class, which is considered being subjected to specified failure mould Formula is simultaneously correspondingly rewritten.

It should be noted that distributing to the title of the variable in pattern suitably independently of distributing in Modelica source codes The title of variable.It is dynamic model will to be used to change in matching, retains normal form name binding.

Suitably, before matching, Modelica models class is flattened, i.e., all superclass are flattened, and all notes Record and connector sub-component are flattened.

In general, stylesheet language can be very simple.Such as there are two behaviour to match respectively with component and equation Make symbol component and equation.

For example, component operator has following form：

·component cname qualifier-list

Wherein, cname is allocated to the identifier of matching component (not being by the title of component declaration definition), and Qualifier-list is the comma separated list of qualifier when meeting the matching result in constraint.Such as, it is possible to specify two A qualifier is the type claimed as the fully-qualified name of Modelica classes, or the type prefix as component statement prefix。

Equation operator can for example have following form：

·equation ename cname rhs-expression

Wherein, ename is allocated to the identifier of matching component, and cname is bundled in one in component operators In a operator, and rhs-expression defines the partially restrained on the right side of matching equation.

In fact, the example of rhs-expression can only be cname, or by following primitive (primitive) structure It builds：

The product for the component or subexpression that product cname-or-exp [cname-or-exp...]-are named (when the other elements including other assemblies may include in product, they can be ignored)；

The adduction for the component or subexpression that sum cname-or-exp [cname-or-exp...]-are named (works as packet When the other elements for including other assemblies may include in adduction, they can be ignored)；

Quotient numerator-cname-or-exp denominator-cname-or-exp-molecule divided by point The expression formula that mother is formed；And/or

Difference minuend-cname-or-exp subtrahend-cname-or-exp-subtrahend subtract by The expression formula that subtrahend is formed.

The expression formula of literal modelica-literal-word Modelica

Some of cname-or-exp elements specified product cname-or-exp [cname-or-exp...]- Specific function.

Other than the pattern of sensitive/applicable class for identification, each failure and/or alternative patterns are also suitably wrapped Description containing " editing procedure ", for changing class to increase the ability for simulating the failure.For example, these editing procedures can be with such as Lower primitive write-in：

The new component of the specified Modelica types of add-component cname type-addition, inside is known as Cname (its actual Modelica title will be generated by program editor).The range of this variable will be " protected ".

Replace-component cname1cname2 [cname]-draw first title cname1 in all equations The example of variable replaces with the variable of first title cname2 references.If optional cname is designated, this yuan is only changed The equation of title reference.

The add-equation ename cname exp-new equation of addition, inside are known as ename, cname are tied up It is fixed to arrive specified exp.The operator of exp uses the identical primitive defined above for stylesheet language.

The equation of remove-equation ename-deletion metavariable ename references.

The Modelica types of the variable of type cname-Returning utilization member title cname references.

For example, the editing procedure of the clutch or brake wear failure that describe previously may have following form：

Add-component (mu-prime, type=type (mu))

Add-component (slip-factor, type=" Real (min=0.0, max=1.0) ")

replace-component(mu,mu-prime)

add-equation(eq2,mu-prime,product(mu,difference(1.0,slip-factor)))

As another example, the full failure pattern of failure Electrical_Open (open circuit) may be as follows：

Component (v, type=" Modelica.SIunits.Voltage ")

Component (i, type=" Modelica.SIunits.Current ")

equation(eq1,i,function(v))

remove-equation(eq1)

add-equation(eq2,i,literal("Modelica.Constants.eps"))

According to the various aspects of an example embodiment, if it find that class model is easily influenced by one or more failures, then should Class will be rewritten.The enumeration type of encapsulation is defined, and lists the various fault modes and " nominal " pattern of class.This new class The discrete mode parameter of type is defined, and defines the pattern that such example is currently running.The equation group of class is by for example as follows The single if equations of form are replaced：

If operating_mode==OperatingModes.Nominal then

<original set of equations>

Elif operating_mode==OperatingModes.Fault1then

<equations for dynamics of Fault1operation>

...

else

assert(False,"Invalid operating_mode detected！")；

end if；

It is as follows with the capacitor change that can simulate the reference of the upper surface of short circuit and open fault for example, being rewritten：

Please noting that the use of Modelica.Constants.eps, it is the Modelica concepts of very small value, just as Modelica.Constants.inf is the Modelica concepts of very big value.This for avoid the dummy run phase with use zero Relevant numerical instability.

Referring now to Figure 1, computer 10 or other similar processors are generally illustrated, for example, it is automatically or semi-automatic Ground executes or otherwise executes model conversion process.It is the first library 20 of nominal component model 22 to its input.For example, library 20 may include the set of model 22, wherein the description of each model 22 and/or otherwise indicate physical assemblies, such as including The operation and/or behavior of the component.Suitably, library 20 may include for example indicating to be used together in systems and/or together Model 22 as the one or more of interactive component or library 20 may include some other suitable of such model 22 Set.In fact, each model 22 in input magazine 20 may include and/or by code line (for example, with Modelica language Or another suitable language write-in) or the like indicate, the code line or the like stores hereof or with its other party Formula is stored in memory appropriate or computer or other the addressable data storage devices of processor 10.Suitably, library 20 In (but the alternatively more or all models 22) description of at least one of model 22 model or otherwise mould The nominal operation of the quasi- physical assemblies by its modeling and/or behavior, such as such nominal operation and/or behavior can correspond to The correct operation of modeled components and/or behavior.In a simple example, library 20 can only include a model 22, more In complicated example, it may include many models 22.

As another input, the second library 30 may include suitable for library 20 modeled components notable alternative patterns and/or The set of mechanism 32.For example, each alternative patterns and/or mechanism 32 in library 30 can describe and/or indicate fault mode and/or Mechanism, or operation and/or behavior some other non-nominal or alternative patterns.In fact, each alternative patterns 32 can be applicable in In one or more of library 20 modelling component, and other model components may be suitable for.For example, short circuit failure mode can be with Suitable for many different electronic building bricks (for example, such as capacitor and/or inductor), and be not suitable for mechanical component (for example, Such as brake).Suitably, each model 32 in input magazine 30 may include and/or by code line (for example, with Modelica language or the write-in of another suitable language) or the like indicate, the code line or the like is stored in file In or be otherwise stored in memory appropriate or can be stored by other data that computer or processor 10 access Equipment.It is, for example, possible to use fault mechanism expresses the potential cause of failure, and fault mode can be used for indicating modeling Component abnormal operation or behavior.For illustrative purposes, consider the lead of the axis to get rusty and disconnection relative to modeling Motor example faults mechanism.The correspondence fault mode generated by above-mentioned mechanism can be the axis for being difficult to rotate respectively (that is, compared with when motor otherwise works normally) and the motor not run only.

In one suitable embodiment, for example, can be captured for each theme component being just modeled and/or with Other modes identify fault mode and mechanism.Can these be subjected to group suitably to classify preferably together with probability of malfunction It knits, so as to consistently build corresponding behavior model.Correspondingly, it is contemplated that classification above-mentioned can form event in turn Barrier mechanism behavior framework (schema).For example, framework can be in a general sense highest abstraction hierarchy on failure definition behavior, And the behavior of lower level fault mechanism is specified using the concept of succession.This method can neatly failure definition mechanism son The behavior of class.Consider the case where fault mechanism is identified as " wearing ".This can indicate the high-level class of the different subclasses of abrasion Not, for example, grinding, impact or corrosive wear.Therefore, suitably, these subclasses (being the subclass of abrasion in this illustration) are inherited The higher level behavior with extension.For example, indicating that the abrasional behavior of material loss is extended to including such as grinding loss subclass In friction behavior.This processing the result is that the fault mode of different components can call identical fault mechanism, i.e., it is potential Failure behavior model.Moreover, these subclasses can cover the expression of the fault mechanism in different field, such as machinery, liquid It is pressure, electronics etc..

Certainly, more broadly, alternative patterns may not indicate that failed operation and/or behavior at all, and may only table Show some other non-nominal operations and/or the behavior of the component being just modeled.

Anyway, the library 40 that processor 10 generates and/or otherwise output enhances component model 42.In fact, Output library 40 in each model 42 may include code line and/or by code line (for example, with Modelica language or another kind Suitable language write-in) or the like indicate, the code line or the like is output in file or is otherwise stored in In memory appropriate or other data storage devices that can be accessed by computer or processor 10.Compatibly, according to by counting The model conversion processing that calculation machine and/or processor 10 execute is exported for the nominal component model 22 of each input from library 20 Enhance component model 42, reinforcing member model 42 includes not only the nominal operation of modeled components and/or description and/or the table of behavior Show, further include be suitable for be modeled the operation of component and/or the description of the one or more of behavior or whole alternative patterns and/or It indicates.Suitably, the corresponding nominal component model 22 of input of the output enhancing component model 42 in library 40 is completely compatible, such as For their interface, input, output and parameter.Thus, for example, enhancing component model 42 can be merely plugged into and/or replace Dehorn claims component model 22, for various assemblies operation and/or behavior failure and/or other alternative patterns test and/ Or analysis.

In one exemplary embodiment, computer and/or processor 10 can execute computer program etc. (for example, depositing Storage is in memory appropriate or other the addressable data storage devices of computer/processor 10) turned with executing above-mentioned model Change process.Fig. 2 is flow chart the step of showing a desired embodiment of model conversion processing 100.Optionally, by calculating The foregoing routine that machine/processor 10 executes can be write by general, class-based, object-oriented computer programming language Enter, for example, Java etc., combines JModelica platforms etc. and Java/JastAdd resolvers etc..Suitably, program It is run in input magazine 20, and for example by being identified and each replacement in the variable, equation and/or other content of model 22 32 associated given pattern of pattern, to identify that it is nominal that various failures easily by library 30 and/or alternative patterns 32 those of influence Component model 22.That is, if the variable of setting models 22, equation and/or other content and given alternative patterns 32 Special style match (for example, in given tolerance), then the alternative patterns 32 are deemed applicable to given model 22. Therefore, model 22 be substantially rewritten as include applicable alternative patterns 32 description, and export new enhancing model 42.

Referring now more particularly to Fig. 2, model conversion handles 100 suitably since step 102, and is read from library 20 A model in modulus type 22.

In step 104, analytic modell analytical model 22, and in step 106, parsed model is checked and/or analyzed to search Specified content wherein associated with alternative patterns 32, pattern and/or clue.In fact, given model 22 is typically included A line or lines of code, including for example define the component being just modeled and its nominal operation and/or the equation of behavior, parameter, Variable and/or other content.According to the content, it can detect and/or distinguish and failure and/or other non-nominal or alternative patterns Associated clue and/or patterns match (for example, in a degree of tolerance) clue and/or pattern.In order to illustrate, Content may include showing that setting models 22 describe element, clue or the pattern of the electrical component of such as capacitor.In fact, model 22 may include equation or parameter or other content, therefrom can determine that model 22 is actually capacitor.Show to continue this Example, can be with it is known that such component, i.e., common electronic building brick, more specifically capacitor, be easy by specific event The influence of barrier and/or specific alternative patterns, such as short circuit or short circuit open circuit etc..That is, the content of model 22 fully matches With 32 associated pattern of corresponding failure and/or alternative patterns.

In determination step 108, if not finding to match the pattern and/or clue that those are found enough, 100 points are handled Branch arrives step 110, else if finding pattern, then handles 100 and proceeds to step 112.

In step 110, model 22 is held essentially constant.That is, model 22 can keep intact substantially.In this feelings Under condition, the correspondence model 42 exported in library 40 will be essentially identical with input model 22.

On the contrary, in step 112, model 22 is rewritten and/or compiles new model 42, not only to describe original to include Nominal operation pattern in model 22, and alternative patterns associated with match pattern are described.

Indeed, it is possible to understand, such processing suitably is carried out for each model 22 of input magazine 20, such as Sequentially or in parallel.Further, it is to be appreciated that in fact, each model 22 can be easy by the zero described in library 30 Or the influence of one or more failures and/or alternative patterns 32.That is, the content of model 22 can mismatch and various events Barrier and/or other 32 associated different modes of alternative patterns, or matching and various failures and/or other 32 phases of alternative patterns Associated one or more different mode.

Suitably, when rewriting (or editor) model 22 is to include alternative patterns (that is, realizing corresponding enhancing model 42), It by the description of each applicable alternative patterns and/or code injection or can otherwise be inserted into model.For example, should Description or code can be obtained from applicable alternative patterns 32, and may include indicating and/or simulating discussed replacement mould The one or more equatioies and/or function of formula.In fact, being replaced by comprising the definition of each model class of failure or alternative patterns New class definition, such as be incorporated to archetype class and add behavior to allow the Modelica moulds of simulated failure or alternative patterns Type class.In addition, enhancing model 42 be provided with controlling mechanism and/or suitable code with allow selection wherein model 42 will be directed to The quasi- AD HOC operated of cover half.That is, foregoing control scheme and/or code allow enhancing model 42 selectively to transport For row under any one mode, the nominal mode either from archetype 22 still passes through the replacement of rewriting/editor's addition A pattern in pattern.In this way, it is enabled for each operation mode and substitutes dynamic.

It (is used for example, the input model 22 of capacitor can look like aforementioned non-reinforced or simple capacitor model Modelica language is written).Therefore, the output enhancing model 42 generated from such input model 22 (also uses Modelica languages Speech write-in) enhancing capacitor model above-mentioned can be looked like.

More specifically, in example in front, it can be appreciated that find that four failures and/or alternative patterns are suitable for Archetype 22.Particularly, in enhancing model 42, define " enumerating " type of encapsulation, wherein be marked as Drift, The alternative patterns of Electrical_Short, Electrical_Leak and Electrical_Break together with nominal mode by It lists.The discrete mode parameter of this new type is defined, and defines the pattern that such example is currently running.Therefore, so Afterwards using and/or Writing condition equation allow the specific operation mode of dynamic select.Suitably, it is answered in each alternative patterns Equation group is expressed in the appropriate branch of the conditional equation.Since operation mode type is such as Modelica parameters, Selected branch will not change during simulation, and compiler can optimize the equation.

(as described above) in one embodiment, flow of power analysis can depend on the standard indicated in input model 22 The identification of power interface.Suitably, these are the example of Modelica connectors class and/or analog, for example, simulating electrical interface Pin classes in the Modelica packets of type.In fact, these, which are electricity interfaces, generally comprises two variables, one is " doing the best " Type, such as " voltage " or " pressure ", the other is " stream " type, for example, electric current or mass velocity.In an implementation appropriate In example, both components of each derived class used in the direct definitions component and the model class of inspection model class are analyzed.Properly Ground, including each model class definition of this example is packaged in the definition of " shell " class, for example, comprising archetype class and Including the reality of the example of the model class and each connector assembly, parameter component and constant component that are found in the primitive class The new Modelica models class of example.

In addition, system (for example, computer or processor 10) can optionally read in parametic fault and/or alternative patterns The often row of table, table describes the specific fault or alternative patterns for assuming that fixed parameter changes during component operates.It is right In each failure or alternative patterns here, table or its row provide fault mode, Modelica classes, class special parameter component, The Parameters variation that the Modelica types of parameter and description occur as the function of the variable of such as FAME_fault_amount Modelica expression formulas.Suitably, for example, the above-mentioned table etc. that can be resident is to store or be otherwise stored in computer Or the file etc. in 10 addressable memory of processor or other suitable data storage devices.

In one suitable embodiment, each new shell class is by the variable power consumable part of addition, such as its Middle nominal mode is arranged to not consume power, and the example of its electricity interface and/or connector assembly is connected to its original mould Raw power interface in the example of type class and/or connector assembly are set as not consuming power under nominal mode.Example Such as, if archetype class is electronic building brick, and electricity interface example is the example of class Pin, then power consuming components appropriate will It is the example of FAME.DynamicDampers.Electrical, wherein damping parameter is set as 0.

In addition, if shell class includes the multiple electricity interfaces and/or connector assembly of same type, it may also be wrapped Containing the variable power conducting subassembly for connecting each pair of compatible connector component, it is nominally set as non-conducting electric power.In general, if even The type for connecing device assembly is identical, or is inherited from same type, then it is assumed that their " compatibilities ".For example, if archetype class is Include the electronic building brick of the example " n " of the example " p " and connector type NegativePin of connector type PositivePin, The both subtype of Pin will then generate the example for the FAME.Bridges.Electrical for connecting the two examples, Wherein set conductance or bridge joint amount to the very small value of Modelica.Constants.eps.

As it appears from the above, enhancing model 42 includes the example for the capacitor model for being added to aforementioned damper and bridge.Suitably, Model conversion processing also flattens the superclass of model into the class of rewriting, and introduces the visible component FAME_ in two new outsides Operating_mode and FAME_fault_amount and give possible breakdown for component FAME_OperatingModes And/or the enumeration type of replacement operation pattern.Suitably, the failure and/or other alternative patterns for showing current exception can To be modeled by changing the two variables.For example, above as shown in output enhancing model 42, it can be by by FAME_ Operating_mode is equal to FAME_OperatingModes.Electrical_Short, and FAME_fault_ Amount models electric short circuit equal to 1.

As it appears from the above, by introducing new continuous variable come processing parameter failure and/or other similar alternative patterns, example Such as, prefix is " FAME_ ".Suitably, addition equation is to set by the upper of parametic fault and/or alternative patterns the variable to State the calculated value of function specified by table.Therefore, the expression for quoting the new variables is replaced by for the reference of initial parameter Formula.For example, as shown in above-mentioned example, parameter C is replaced by expression formula C* (1-FAME_fault_ (in input model 22) Amount) (in output enhances model 42).

In general, according at least one other embodiment, it is disclosed one kind and is used in combination for generating damage Parameter Map Make the method, system and/or device of the representative of break treat with model.

In general, as disclosed herein be shown in, damage process model indicates to be used as such as load, material properties due to its service environment Damage caused by the inefficacy mechanism being in active state in system component with the functions of the parameters such as size.The damage of component State is reflected services service condition and historical record derived from context model.It is exported about from context and component model Parameter and uncertainty about break treat with model parameter be characterized and be incorporated in stochastic simulation.It is suitable according to one Embodiment, the analog result of parametrization is represented as damage Parameter Map (such as table), is then directed to using the damage Parameter Map What user specified searches expected damage without running again at basic random damage using scene in the component that user specifies Reason simulation.

Can execute failure risk or the system component of fail-safe analysis another example is engine/power trains of vehicle System mounting bracket.Due to the load of the time change caused by rugged landform, vibration or impact downward driving, these branch Frame bears wear-out failure.In fact, other fault mechanisms can also be analyzed and/or can similarly analyze different system groups Part.Suitably, analysis is related to using different project analysis tool and/or software.For example, the wind of the wear-out failure of mounting bracket It assesses danger：

Use more bodies dynamic (MBD) Software Create branch booster history；

The stress in holder is obtained using finite element analysis (FEA) software；And

(PFA) software evaluation probability of malfunction and shelf life are analyzed using material Probabilistic Fatigue.

According to a suitable embodiment, the rear end workflow 200 of various aspects according to the inventive subject matter is shown in FIG. 3 (it is optionally invisible to user).In order to illustrate herein in conjunction with the seed design vector based on one group of given design parameter (SDV) workflow 200 described.However, in fact, workflow 200 can otherwise be utilized to realize that other are answered Analog result.In one suitable embodiment, rear end workflow 200 can carry out as follows：

As shown, developing MBD simulation models 202 first.In this example, the SDV of exploitation includes representing engine The spring of the holder of shell is connected to transmission component.More generally, design configurations 204 and task definition 206 can be used as defeated Enter to be given workflow 200.

In particular, for example collecting mass property from CAD (CAD) model.As indicated, CAD information be by MBD models 202 are obtained from design configurations 204.Alternatively, them can be estimated, for example it is assumed that the babinet of even density.It closes Suitable ground, MBD models are created based on the CAD model of such as SDV and its component.In the present case, pendulum spring/damping The position and rigidity of characteristic and mounting bracket are determined, and with the selection of landform/barrier profile, rational car speed Also it is determined.To sum up, obtaining design parameter 208 (such as aforementioned) from design configurations 204.As indicated, design parameter 206 is used In generating one or more parametrization Figure 30 0, for example, such as parameterizing palmograph.

As indicated, realize one group of result 210 using MBD simulation models 202 and task definition 206, i.e., according to this example, Acceleration/power time history at point-of-interest.Suitably, the result is that post-processing, such as the electric power group including currently making great efforts Part mounting bracket.According to this example, task power history is defined by being connected the power history generated from MBD simulations.For example, in mesh Before in the case of, SDV is driven：The high semi-cylindrical hills of Perryman (Perry is graceful) 3,12 ", 30 " depression of order and 96 " width groove Intersect.Export power history is simulated (for example, this reality using dynamic analysis and design system (DADS) software, such as from the DADS of SDV The power history of holder in example).Handling result 210 is again for generating parametrization Figure 30 0.In this example, parametrization Figure 30 0 is used In determining component life 302.

In this example, one kind for 2 kinds of materials and 3 kinds of Angular Dimensions and angle changing and gusset thickness is defined Angle support.It for each support Design, executes linear-static finite element (FE) and analyzes, to obtain due to holder task historical record In maximum, force and caused by maximum stress in holder.By by the maximum stress in the yield stress of material divided by the holder To calculate the allowable stress coefficient of each holder.FE analyses the result is that stiffness of support on three directions.In DADS vehicle moulds Rigidity is shown by spring in quasi- analysis.

In turn, wear-out failure simulation model is changed, to be gone through for due to simulating derived branch booster task from DADS The specified value of history and the given material of the holder that generates, allowable stress coefficient (ASF), to generate damage Parameter Map.Pass through song Line fitting holder FE analysis results obtain backing thickness t_aAnd t_gASF equations.Stiffness of support equation also carries out in a similar way It derives.

Leading-out needle is simulated from the MBD of SDV used branch booster history is simulated to damage using DADS softwares.By will be from Perryman 3,12 " high semi-cylindrical hill, 30 " depression of order and 96 " holder that the DADS simulations that wide groove intersects generate of SDV Power history is connected, to define task power history.Fig. 4 shows the graphical representation of route, and path height is along vertical axis (with English It is very little be unit) and route distance along trunnion axis (in inches).

According to the history of support engine and four holders of transmission component, because front left (FL) holder history has highest Power peak value, so selecting them to define task.According to rapidly adapting to next-generation (FANG) vehicle plan, an example task It is defined as being based on 1200 hours 28,800 miles of range abilities.In history, 28 seconds Perryman 3 are during entire continue It repeats (212826 times), and every 10 miles of repetition walls rise rank, depression of order and groove crossing event (2880 times).Fig. 5 A and Fig. 5 B Show that groove intersects the sample acceleration time history with 1 meter of wall depression of order.

In one suitable embodiment, such as cyclic loading history, geometry, material properties and the spy of given component Fixed uncertainty, Tiredness model generate the different damage Parameter Map of probability of occurrence.Tiredness model be based on due to cyclic loading and The strain of caused material and service life (ε and N) behavior.The input of Tiredness model is power history, material data and the control of component Parameter processed, such as holder in the present case.Monte Carlo (M-C) cycle is performed multiple random experiments to be directed to one Group parameter value exports the damage of desired probability of happening.In single M-C experiments, is exported and damaged by following steps.Input carries Lotus historical record is converted into load histories record using by load described in geometry or ASF to stress alternation first.It connects It, such as the range of stress and cycle-index is identified using rain-flow counting technology or algorithm.If the range of stress of a cycle More than yield strength, then overall strain ε will include elastic strain and plastic strain.Suitably, Nuo Bai (Neuber) rule is used for It converts elastic stress to overall strain range using the load-deformation curve of material.Wo Ke (Walker) relationship is subsequently used for school Positive overall strain range is to illustrate the mean strain of cycle.Damage caused by overall strain is exported using strain and life model.Most Afterwards, it is always damaged using advanced in years receive caused by (Miner) rule adds up each cycle, damage caused by find out entire power history.

Fig. 6 shows sample von mises (von Mises) answer force profile of 4 holders of 4x 4x.39.4ksi maximum stress It is since maximum, force is 18300lbs, and the yield stress of stainless steel 17-4PH is 170ksi.Therefore, ASF=170/39.4 =4.31.

In one example, stainless steel 17-4PH H900 and aluminium 6061-T6 materials are for holder damage simulation.In holder The material property used in damage simulation is from MIL-HNDBK-5J (Department Of Defense Handbooks：Metal for space launch vehicle structure Material and element (on January 31st, 2003)) and the acquisition of other suitable material data files.The stress fracture provided in table 1 is special Property is the yield strength of alloy.A and B " basis " value (come the document delivered self) of these materials is for exporting yield stress Average value, standard deviation and variation coefficient, and intensity uncertainty is described.Overall strain is with lifetime data for material model of booting Method, it is uncertain to capture data acquisition system implicit fatigue life.

The specimen breakdown Parameter Map of 1 17-4PH900 steel brackets of table

Damage Parameter Map allow system designer or design tester explore different designs select reliability as a result, And assess due to role pressure and caused by calamitous wear-out failure possibility.For example, according to holder example described here, Method and/or system described here help to find the answer of problems with：

If holder is used to support the engine in SDV, wherein vehicle traveling for example travels on above-mentioned terrain profile 1200 hours, then the probability of the wear-out failure in holder is how many；And

If the size or material for holder are changed, which difference will produce

More than, it has been described that how to simulate random physical fault model, illiteracy therein in advance in the frame of Monte Carlo Special Carlow frame, which is incorporated to design uncertain (material selection and geometry) and the expection of the their entire life of component, to be made With spectrum.Suitably, it is as a result stored as damage Parameter Map, may then pass through cast material and geometric parameter and using horizontal It is indexed.According to a suitable embodiment, described method and/or system be extended to consider material character with And the variation (i.e. tolerance analysis) in manufacturing process.In addition to simulation is arranged, system will not only consider to come from design alternative (material Material and geometry) sample, it is also contemplated that material property and geometric dimension caused by the variation in manufacturing process The sample of distribution, extension process are similar to above-mentioned processing.

In fact, this multidimensional sample space means the quantity of simulation as the distributed quantity therefrom sampled exponentially increases It is long.However, this can be offset by the accelerated method developed for Monte Carlo simulation.

In one suitable embodiment, which is further adapted for simulating MBD points in desired interest point set Cloud carries out interpolation, so that it is determined that unknown design or operating parameter is integrated into the acceleration at these points of interest.More specifically, holding Row MBD simulations, to numerically be determined based on operating environment (for example, landform/road) and operating parameter (for example, speed) The load (acceleration) on point of interest in system (for example, suspended rack assembly in automobile).These subsequent load uses are damaging Handle be used for determining in model it is any it is given using rank (for example, using city driving profile by rugged Suspension failure after 100K miles of road) lower component failure probability.In fact, these simulations can be calculated and work close The setting and operation of collection type.Therefore, a small amount of point of interest and a small amount of operation profile in usual simulation system.But it is this Limited simulation test may result in deployment system and chance failure occur.Therefore, according to the system and or method used herein An exemplary embodiment, realize the processing that the MBD points closed in desired point-of-interest collection simulate the enterprising row interpolation of cloud, So that it is determined that the unknown design or operating parameter of such as vehicle is integrated into load and/or acceleration at these points of interest.It closes Suitable ground, which utilizes the linear structure dynamic of the MBD models based on mass-spring-damper device, in needle To the acceleration that is precalculated from the system point of interest of the parameter sets from parameter space uniform sampling into row interpolation.

MBD analyzes the reliability for contributing to rapid discovery larger design space.Specifically, MBD analyses are to determining task Using very useful with the component fatigue life under the conditions of design configurations.

Suitably, various aspects according to an exemplary embodiment, it is arbitrary in design configurations to determine to provide the ability Component load (for example, front-mounted engine of automobile is set in or rear engine configures) at position.By being simulated in MBD points Interpolation on cloud, user can by more system and more fully in a manner of explore entire design space, to realize it is more efficient, reliable and Steady design.

Following evidence illustrates the value for 3 axle accelerations that arbitrary point design is determined using interpolation, which comes from The multidimensional cloud of derived 3 axle accelerations of MBD simulations carried out from the point design set across test design point range.Here it is in All point designs that existing cross validation Research Hypothesis is considered belong to tracked personnel's delivery vehicle class.Calculate acceleration Point of interest is the support element that electric power assembly is connected to cabinet.Assuming that there is 4 identical holders of support electric power assembly.

In this example, 125 clouds consider that the MBD by following 3 simulation parametrizations is simulated：

Car speed in 3 landform of Perryman：The analogue value is 5,8,11,14 and 17mph；

Support vertical rigidity：The analogue value be 3.49 × 105,8.55 × 105,1.71 × 106,2.14 × 106 and 2.83 × 106lb/in；And

Electric power assembly mass ratio coefficient：The analogue value is 0.5,0.75,1,1.25 and 1.5 (x nominal values).

The holder that the function as speed and stiffness of support is shown respectively in Fig. 7, Fig. 8 and Fig. 9 is maximum in the x, y, and z directions Acceleration.

Estimate the acceleration (that is, for given analog parameter set) at known analog point using 3D spline interpolations, And it will be relative to the percentage that the error calculation of the analogue value is the analogue value.Figure 10, Figure 11 and Figure 12 be shown as car speed and The maximum transversal of the function of electric power assembly mass ratio coefficient, the percentage error of longitudinal direction and normal acceleration.It gives relatively Low error (i.e. most of errors are less than 5%), for determining arbitrary point design from the multidimensional cloud of 3 axle accelerations simulated in advance The interpolation methods of 3 axle accelerations be considered effective.

Above method and/or device are described about specific embodiment.It should be appreciated, however, that also may be used To consider specific modification and/or change.

Anyway, it is to be appreciated that, in conjunction with the certain exemplary embodiments presented herein, specific structure and/or work( Energy feature is described as being incorporated in the element and/or component of restriction.However, it is contemplated that these feature folding lines are to same or analogous Benefit, or be otherwise appropriately incorporated in other elements and/or component.It should also be understood that exemplary embodiment Different aspect can use to being appropriately selected property with realize be suitable for it is expected application other alternative embodiments, other are replaced Change embodiment it is thus achieved that the various aspects being incorporated to herein attendant advantages.

It will also be appreciated that any one or more particular tasks described here, step, processing, method, function, member Part and/or component can be realized suitably via hardware, software, firmware or combinations thereof.Particularly, processor 10 can be by counting Calculation machine or other electronic data-processing equipments are implemented, the computer or other electronic data-processing equipments be configured and/or with Other modes are supplied to execute one or more tasks, step, processing, method and/or function described here.For example, real The computer or other electronic data-processing equipments for applying processor 10 can be arranged, provide and/or be programmed with suitable code column Table (for example, source code, interpretive code, object code, direct executable code etc.) or other similar instructions or soft Part or firmware so that when being run by computer or other electronic data-processing equipments and/or being executed, one described here or Multiple tasks, step, processing, method and/or function are completed or otherwise execute.Suitably, code listing or The list of other similar instruction or softwares or firmware is implemented as and/or is recorded, stores, containing or comprising in non-transient meter In calculation machine and/or machine readable storage medium and/or on so that as being provided to by computer or other electronic data It manages equipment and/or can perform by computer or other electronic data-processing equipments.For example, suitable storage medium and/or medium It can include but is not limited to：Floppy disk, flexible disk (-sc), hard disk, tape or any other magnetic storage medium (medium) or medium (media), CD-ROM, DVD, CD or any other optical medium or medium, RAM, ROM, PROM, EPROM, FLASH- EPROM or other memories or chip or cassette memory or computer or machine or electronic data-processing equipment can be from them In any other tangible medium or medium for being read out and using.Substantially, as used in this, non-transient computer can Reading and/or machine readable media (mediums) and/or medium (media) include all calculating in addition to instant exposure signal Machine is readable and/or machine readable media and/or medium.

Optionally, any one or more particular tasks described here, step, processing, method, function, element and/or Component may be implemented and/or implement one or more all-purpose computers, special purpose computer, programming microprocessor or micro-control Device and peripheral integrated circuit element, ASIC or other integrated circuits processed, digital signal processor, such as discrete component circuit it is hard Wire electronics or logic circuit, the programmable logic device of PLD, PLA, FPGA, graphics card CPU (GPU) or PAL etc..It is logical Often, any equipment that can realize finite state machine can be used, which can realize described here each again Task, step, processing, method and/or function.

Also, it should be understood that being described herein can be in a suitable case for the particular element merged Independent component otherwise separates.Similarly, being described as the multiple specific functions carried out by a particular element can be with It is carried out by multiple and different element, multiple different element is independently acted to carry out individual function or specific Individual function can be divided and multiple and different elements by unanimously acting carry out.Alternatively, herein with its other party Formula describes and/or be shown as some elements different from each other or component can physically or functionally group in appropriate circumstances It closes.

In brief, this specification is elaborated by reference to preferred embodiment.Obviously, in reading and understanding this specification Afterwards, other people will expect modifications and changes.It is intended to invention is constructed as including all such modifications and change, as long as they fall In the range of entering appended claims or its equivalent.

Claims (17)

1. a kind of method for automatically generating the enhancing model of physical assemblies, the method includes：

Input model is read in processor, the behaviour for the physical model that the input model description is modeled by the input model The nominal mode of work；

The input model is parsed to generate its parsing using the processor；

The parsing of the input model is analyzed using the processor；And

Based on the analysis, the enhancing model of physical assemblies is automatically write from the input model using the processor, it is described The nominal mode and (ii) of the operation for the physical assemblies that enhancing model description (i) is modeled are different from the institute of the nominal mode of operation Both at least one alternative patterns of the operation of the physical assemblies of modeling；

The wherein described analysis is carried out according at least one of first method and second method method；

The wherein described first method is for detecting whether input model is easy by the shadow for including one or more failures below It rings：(i) catastrophic, (ii) flow of power, and (iii) parameter type failure；And

Wherein particular aspects of the second method for the dynamic behaviour of matching component, and these suitably modified matched aspects are with anti- Reflect the dynamic of fault mechanism.

2. according to the method described in claim 1, wherein：(i) variation in the dynamic of the representation for fault component of calamitous type So that it is the entirely different effect of what component that component, which is played with it, the failure influence of (ii) flow of power type flows to one or more One or more of a flow of power or one or more of one or more flow of power from component, and (iii) joins The offset of the hypothesis preset parameter of the failure reflection component of several classes.

3. according to the method described in claim 1, wherein including according to the first method progress analysis：

The example of at least one of electricity interface or connector class is searched in the parsing so that if it find that electricity interface Or at least one of connector class, then its relevant alternative patterns described in the enhancing model.

4. according to the method described in claim 1, wherein including according to the second method progress analysis：

Pattern is detected in the parsing, the pattern matching is associated with the alternative patterns that will be described in the enhancing model Target patterns so that associated described in the enhancing model if detecting the pattern for matching the target patterns Alternative patterns.

5. according to the method described in claim 1, wherein said write includes：

Controlling mechanism is provided, by the controlling mechanism, an operation mould in the operation mode described in the enhancing model Formula is selected in the simulation using the enhancing model.

6. according to the method described in claim 1, further comprising：

Obtain the task definition of the system for one or more components；

Obtain the design configurations for the system for indicating component；

Design parameter is extracted from the design configurations；

Information is extracted from the design parameter；

More body dynamic (MBD) simulation models are established based on the information extracted, and the MBD is run using the task definition Simulation is to realize one group of result；And

At least one Parameter Map is generated based on this group of result realized and the design parameter extracted.

7. according to the method described in claim 6, the tired and described side of the component in the wherein described Parameter Map reflection system Method further comprises：

The fatigue life of the component is determined from the Parameter Map.

8. being determined by the task since component is exposed to according to the method described in claim 7, the wherein described Parameter Map is related to (i) The condition that is reflected of justice and the stress that generates cause the probability of the wear-out failure of component, (ii) design parameter array.

9. according to the method described in claim 8, wherein MBD simulation models are run repeatedly to realize the multiple spot cloud of result, The array of point expression different designs parameter in medium cloud, and the method further includes：

Interpolation is directed to the value of result between the point of cloud.

10. a kind of system for automatically generating the enhancing model of physical assemblies, the system comprises：

Processor is used for：

Input model is read in, the input model describes the nominal mould of the operation of the physical model modeled by the input model Formula；

The input model is parsed to generate its parsing；

Analyze the parsing of the input model；And

Based on the analysis, the enhancing model of physical assemblies is automatically write from the input model, the enhancing model describes (i) The nominal mode of the operation of the physical assemblies modeled and (ii) are different from the physical assemblies of the nominal mode of operation modeled Both at least one alternative patterns of operation；

Parsing is wherein analyzed according at least one of first method and second method method.

The wherein described first method is for detecting whether input model is easy by the shadow for including one or more failures below It rings：(i) catastrophic, (ii) flow of power, and (iii) parameter type failure；And

Wherein particular aspects of the second method for the dynamic behaviour of matching component, and these suitably modified matched aspects are with anti- Reflect the dynamic of fault mechanism.

11. system according to claim 10, wherein：(i) change in the dynamic of the representation for fault component of calamitous type Change so that it is the entirely different effect of what component that component, which is played with it, the failure of (ii) flow of power type influences flow direction or comes from One or more directions of energy of component, and the offset of the hypothesis preset parameter of the failure reflection component of (iii) parameter type.

12. system according to claim 10, wherein including according to the first method progress analysis：

The example of at least one of electricity interface or connector class is searched in the parsing so that if it find that electricity interface Or at least one of connector class, then its relevant alternative patterns described in the enhancing model.

13. system according to claim 10, wherein including according to the second method progress analysis：

Pattern is detected in the parsing, the pattern matching is associated with the alternative patterns that will be described in the enhancing model Target patterns so that associated described in the enhancing model if detecting the pattern for matching the target patterns Alternative patterns.

14. system according to claim 10, wherein the processor is further used for：

Obtain the task definition of the system for one or more components；

Obtain the design configurations for indicating component system；

Design parameter is extracted from the design configurations；

Information is extracted from the design parameter；

More body dynamic (MBD) simulation models are established based on the information extracted, and the MBD is run using the task definition Simulation is to realize one group of result；And

At least one Parameter Map is generated based on this group of result realized and the design parameter extracted.

15. system according to claim 14, wherein the fatigue of component in the Parameter Map reflection system and described Method further comprises：

The fatigue life of the component is determined from the Parameter Map.

16. system according to claim 15, wherein the Parameter Map is related to (i) since component is exposed to by the task Define reflected condition and the stress that generates cause component wear-out failure probability, (ii) design parameter array.

17. system according to claim 16, wherein MBD simulation models are run repeatedly to realize the multiple spot cloud of result, The array of point expression different designs parameter in its medium cloud, and the method further includes：

Interpolation is directed to the value of result between the point of cloud.