EP3259690A1

EP3259690A1 - Damage modelling for a component consisting of a nonlinear elastic material

Info

Publication number: EP3259690A1
Application number: EP16704458.5A
Authority: EP
Inventors: Jochen Ringswirth
Original assignee: Siemens AG Oesterreich
Current assignee: Siemens Mobility Austria GmbH
Priority date: 2015-02-18
Filing date: 2016-02-16
Publication date: 2017-12-27
Also published as: WO2016131800A1; AT516862A2

Abstract

The invention relates to a method for the computational damage modelling of a component (1) that consists of a nonlinear elastic material, characterised in that said component (1) is replaced in the model by a plurality of node points (3) interconnected by a plurality of rods (2), a mathematical material model being imposed for each rod (2) and said rods forming a first network; in that at least one cyclic load with a known load value is applied to the model of the component (1), the position and/or length of the rods (2) being altered, as a result of the load, depending on the mathematical material model of said rods; and in that a degree of damage to the rods is determined depending on the state of strain of said rods (2).

Description

description

Modeling the damage of a component from a

nonlinear elastic material

Technical area

The invention relates to a method for computational

Modeling the damage of a component from a

nonlinear elastic material. In addition, the present invention includes an associated

Computer program product. Non-linear elastic materials include components consisting of natural or synthetic elastomers as well as components containing natural or synthetic elastomers

Elastomers, in particular filled, ie filler-reinforced, elastomers. Such materials often have a tensile-pressure anomaly as well as a pronounced

Medium voltage dependence, which strength calculations with respect to fatigue strength and damage calculations very difficult. The medium voltage (or medium load) is the voltage which the component in installed state in

Rest position is subject (eg, the bias of a spring, resulting from the weight of a vehicle mounted thereon part), and then in the operating state (for example, when driving a rail vehicle) other voltages, especially cyclic tensions (eg, caused by the operation of the vehicle movements), be superimposed. The other stresses can only act as compressive stresses, only as tensile stresses or as compressive and tensile stresses. Depending on the position of the medium voltage, so whether this in the Pull or in the pressure range, the other voltages will have different effects.

Components to be modeled may be vehicle components,

in particular springs and coupling elements, from

Rail vehicles or components for other technical

Applications that consist of at least one layer of elastomeric material.

State of the art

Conventional methods of damage calculation for components made of nonlinear elastic materials, such as

Elastomer components are based, for example, on finite element methods. To do this for each

Elastomer mixture on the one hand, the geometric dimensions and on the other hand be known physical properties. The latter must be determined by testing the strength of component samples due to the nonlinearity of the elastic properties for different stress levels and medium stresses, respectively. The damage calculation based on the finite element method must then also be calculated for different load cases and uses the

Resistance to vibration tests on the component samples or must be between the results of the

Interpolate vibration resistance tests. In addition, the finite element method for the exact geometric model also requires a fine-meshed network structure of the model. The local discontinuities in the calculations caused by the interlinking of strongly deforming components, for example due to overlapping nets, require many years of experience in the field for correct interpretation the damage calculation based on the finite element method.

Presentation of the invention

It is therefore an object of the present invention to provide a method by which the modeling of the damage of a component by means of a less complex physical model of the component, at least without finite element method, can take place.

This object is achieved by a method having the features of patent claim 1, wherein it is provided

that the component in the model is replaced by a plurality of nodes which are interconnected by a plurality of rods, each rod being given a mathematical material model and the rods forming a first network,

that at least one cyclic load with a known load value is exerted on the model of the component and the load changes the position and / or length of the rods as a function of the mathematical material model of the members,

- That depending on the strain state of the rods damage to the rods is determined.

The invention thus consists in the properties of the material by - in comparison to finite elements

To model methods simpler - physical-mathematical model, without first performing vibration strength tests on real component samples. The physical properties such as stiffness gradients etc. of the real component are imposed on the material model. However, there is also the possibility of these stiffness gradients through the model to adjust, but for a pure

Damage calculation is not leading.

The damage of the component can be determined according to different aspects. Basically, the individual damages of the rods are determined, which are determined on the basis of the strain state. The individual damages are summed linearly and thus a virtual damage of the component is calculated. Furthermore, strains of individual rods, which exceed a predetermined limit, be used so that one dissolves this compound and so realized in total a reduction of the network and thus damage. In this case, each rod can be imposed on any material model, in particular a linear or nonlinear elastic behavior.

It can be provided that for modeling the

Medium voltage dependence of the physical properties of the component further rods are provided in the nodes, which form at least a second network, wherein each additional rod a mathematical material model is imposed. This allows the adjustment of a

Medium voltage dependence, the intensity of

Medium voltage dependence by the number of

Additional networking is adjusted. It is also possible to impose any desired material model on these further rods, in particular with a linear elastic or non-linear elastic behavior.

The mathematical material models of the first network

(Main network) as well as the second network (s)

(Additional crosslinks) may be analog or different. Both the individual bars of the first network as well

If appropriate, the further bars of the second network are normally distributed randomly between the nodes in accordance with a normal distribution. Basically, of course, other distributions are possible.

The damage calculation can be simplified considerably if the three-dimensional component in the model is replaced by a two-dimensional replacement model before the formation of the first network. Then, the inventive

Damage calculation can be reduced to a mathematical problem in two dimensions and simplified accordingly. This transformation from a 3D to a 2D component or state of motion depends on the geometry of the component to be calculated. In this case, one usually divides the movements acting on the component into a tensile-pressure and a thrust-load. Due to the simplicity of the method according to the invention, it is possible that as input variables for the model of the component, only the acting resilient layer thickness of the component in mm and, when using force signals as

Loading size, the stiffness or stiffness gradients of the component to be defined (in all three spatial directions).

The method according to the invention can be used to calculate the service life in advance for a component yet to be created, namely as a function of the rigidity and the resilient layer thickness of the component. In this case, the method according to the invention is run through for different stiffnesses and / or layer thicknesses of the component. The

resulting damage or the corresponding Lifetimes are then compared with each other and it can then be a component with that rigidity and layer thickness or their combination produced in the

Model calculations the highest life has exhibited. Insofar it can be provided that by variation of the

Stiffnesses a component with a long service life is determined.

To create a tensile-pressure anomaly of the component in the model

To be able to take into account, it can be provided that the layer thickness of the component given as the input variable for

Considering a train-pressure anomaly of the component, in particular autonomously depending on the train-pressure distribution, is changed.

As a load input variable for the cyclic load can be a time-dependent path signal or knowledge of the

Stiffness curves also a time-dependent force signal can be used. Due to the simplicity of the model, not only a periodic signal, such as a

Sinusoidal, can be used, but it can

In principle, more complex signals, such as at

Measurement runs of a real rail vehicle have been determined.

Not only a uniaxial load of the component is considered, but usually multi-axial loads. This multiaxial load is combined into a load signal acting on the 2D replacement model. The

Loads can be in midloads, amplitudes and

Time steps are split.

If with the inventive method only in each case

Individual damages can be calculated, these can nevertheless total damage. In this regard, it can be provided that based on various

Loads the damage of the component for each of these loads is calculated and the total damage by means of linear damage accumulation, e.g. determined by Palmgren-Miner rule. Basically, the integration of other damage accumulations is conceivable.

The method according to the invention can be applied to a component which is a vehicle component, in particular a spring and coupling element.

The inventive method can be advantageously used for components of a rail vehicle, namely

for example, for cone springs in the primary stage as well

Secondary stage, Radsatzführungsbuchsen, sockets of

Längsmitnahmen, laminated springs, etc., but on for

corresponding components of other vehicles or equipment. Finally, the invention comprises a

Computer program product comprising a program and directly loadable into a memory of a computer, with program means for carrying out all the steps of the method according to the invention when the program is executed by the computer.

The method according to the invention makes it possible to carry out a number of hitherto impossible damage calculations, in particular for rail vehicles, and thus a considerable advantage in the various design phases, which helps to save time and effort. may include a damage assessment in the design phase of a component for a space and stiffness estimation Without precise knowledge and description of the finished component, but only with knowledge of the component to

Available space and the possible stiffness of the component (the latter are the skilled person usually from its construction activity in rail vehicle construction

known), at a very early stage of the construction, an assessment of the life of the component based on the - preferred in terms of running technology - stiffness of the component. The future component can be found in the

In terms of injury and lifetime technically

be judged, certain stiffnesses used in the process according to the invention for the time being, may be detrimental to the life and therefore be excluded. Thus, a component can be developed, which represents a good compromise between the requirements of the running properties and the life. So far, a component has been tailored and developed primarily to the running requirements. The effect on the lifetime was analyzed at a later stage of development and the

Lifespan could then only be more heavily influenced.

Furthermore, with the inventive method a

Damage estimation for identical parts in other applications and thus for other component loads: The

Strive for the use of the same components

A variety of requirements sets a complete

Knowledge of the component properties (component potentials) ahead. So far, this consideration was primarily focused on the geometry and rigidity of the component. The achieved lifetime of the component was known only at a later date. By means of a damage analysis and thus with knowledge of the stiffness gradients and the component layer thicknesses (Elastomerschichtdicken) can be a same part selection also in terms of lifetime optimization.

Also, with the inventive method a

Damage analysis for the creation of lifetime tests to be realized: Component tests are still necessary, on the one hand there is a duty of proof of

Manufacturer of the component for the safety of the component, on the other hand, the manufacturer must internally prove the function of the component. The manufacturer strives to keep the cost of component testing as low as possible. This can be achieved with a test signal having the same damage content as occurs during operation of the component. The test signal can be applied to the

Damage analogy of the component to be optimized.

BRIEF DESCRIPTION OF THE FIGURES For further explanation of the invention, reference is made in the following part of the description to the FIGURE, from which further advantageous embodiments, details and

Developments of the invention can be seen. Showing:

1 model of a component of networked rods,

2 is a Haigh diagram based on measurements,

Fig. 3 Haigh diagram based on the

inventive method.

Embodiment of the invention First, the data of the elastomer component to be modeled are determined, that is, the elastomer layer thickness and the stiffnesses of the component in all three

Spatial directions, where the stiffness is in the form of scalars and / or polynomials. Subsequently, the component is mathematically reduced to a two-dimensional substitute model, ie to a two-dimensional strain state, for

Example of a surface model with edge length 1. All constraints, boundary conditions, loads and

Operating signals (voltages) of the three-dimensional component are reduced to the replacement model.

Then, the voltage acting on the component voltage

fixed, either as a time signal or as medium voltage (medium load) and superimposed oscillation, then only the amplitude and the phase of the superimposed oscillation must be specified.

Since elastomeric materials have a tensile-to-pressure anomaly, i. they experience higher damage in the tensile area than in the pressure area, even if this anomaly should be modeled. This anomaly, for example, here by the change in the set as input parameter layer thickness of the component, depending on the degree of Zuganteiles, adjusted.

The following steps will then be for each pair

Mean voltage and amplitude go through and then give each a so-called partial or single damage: It becomes a motion vector for the 2D replacement model

formed, and accordingly a normally distributed

Main network (first network) with many tie rods, which are normally distributed as random junctions

connect two-dimensional replacement model together. Subsequently, an additional networking (second network) of the replacement model by means of tension rods, depending on the

Medium voltage amplitude ratio and time. Then, the strain can be determined from the composite mesh (first and second meshes) and the motion vector. From the stretching, namely as a function of the elongation, the

Partial damage for the investigated load case can be calculated. FIG. 1 shows an elastomer specimen, with the left half representing the real component and the right half schematically representing the two-dimensional substitute model of rods, in this case tension rods 2 anchored in nodal points 3. The component 1 is here between two rigid,

enclosed parallel cover plates 4, which are normal to a longitudinal axis 5, which simultaneously forms the axis of symmetry of the component 1. Normal to the longitudinal axis of the component 1 is not limited. The distance between the two cover plates 4 corresponds to the layer thickness of the component. 1

If then several partial damages have been calculated, then these can by Palmgren Miner rule to a

Total damage can be summed up.

The individual partial damages can also be represented as a haigh diagram - so you do not just have one

selective result, but a whole field of results. In a Haigh diagram, the dependence of the

Fatigue strength represented by the load type. On the horizontal axis the center load is plotted in% as a function of the layer thickness, on the vertical axis the amplitude of the additional cyclic additional load in%. The load left of 0% medium voltage corresponds to one Compressive load, the load on the right of 0% mean stress corresponds to a tensile load of the component.

In Fig. 2 that is from measurements on material samples of

Elastomers and curve fitting identified Haigh diagram shown. The drawn lines indicate states of equal strength, with a tendency to larger ones

Strengths are present under pressure. The grayscale gives information about the amount of strength, with the highest strength, on the scale right with "20"

is not reached. In the low amplitude pressure range (up to about 2%) there is a strength of about "17." As the amplitude increases, the strength decreases according to the scale, with the decrease being in the range of

Tensile load is faster than in the area of

Pressure load.

In Fig. 3 is the method of the invention

shown Haigh diagram shown. The absolute

Strength values differ somewhat from those of FIG. 2. For example, in Fig. 3 for the printing area with lower

Amplitude (up to about 2%) with about "20" reaches a higher strength than in Fig. 2. However, the

Medium voltage dependence, ie the straight horizontal lines in Fig. 2, basically reproduced correctly, so that the medium voltage dependence can be well simulated with the inventive method.

LIST OF REFERENCE NUMBERS

1 component

2 bar (tie rod)

3 node Cover plate longitudinal axis

Claims

Claims:

Method for mathematically modeling the damage of a component (1) made of a nonlinear elastic material, characterized

- That the component (1) in the model by several

Replacing nodes (3) which are interconnected by a plurality of bars (2), each bar (2) being subjected to a mathematical material model and the bars forming a first network,

that at least one cyclic load with a known load value is exerted on the model of the component (1) and the load changes the position and / or length of the rods (2) as a function of the mathematical material model of the rods,

- That depending on the strain state of the rods (2) damage to the rods is determined.

A method according to claim 1, characterized in that from the damage of the individual rods (2) by

Summation damage to the component (1) is determined.

A method according to claim 1 or 2, characterized in that at expansions of individual bars (2) which exceed a predetermined limit, the connection of nodes (3) by these bars (2) is considered dissolved.

Method according to one of claims 1 to 3, characterized in that each rod (2) as a material model, a linear-elastic or nonlinear-elastic

Behavior is imposed.

5. The method according to any one of claims 1 to 4, characterized in that for modeling the

Medium voltage dependence of the physical

Properties of the component more rods in the

Nodes (3) are provided, which form at least a second network, wherein each additional rod a mathematical material model is imposed.

6. The method according to claim 5, characterized in that each further rod as a material model, a linear-elastic or non-linear-elastic behavior is imposed.

7. The method according to any one of claims 1 to 6, characterized

characterized in that the rods (2) and optionally further rods are placed according to a normal distribution at random in the nodes (3).

8. The method according to any one of claims 1 to 7, characterized

characterized in that the three-dimensional component (1) is replaced in the model before the formation of the first network by a two-dimensional replacement model. 9. The method according to any one of claims 1 to 8, characterized

characterized in that the input variables for the model of the component (1) are the stiffnesses or

Stiffness curves of the component and the layer thickness of the component (1) can be used. 10. The method according to claim 9, characterized in that by varying the stiffnesses and / or the

Layer thickness is determined a component (1) with a long service life.

11. The method according to any one of claims 1 to 10, characterized

characterized in that the predetermined as input Layer thickness of the component (1) to account for a train-pressure anomaly of the component (1) is changed.

12. The method according to any one of claims 1 to 11, characterized

in that a path-time signal or force / time signal is used as the input variable for the cyclic loading.

13. The method according to any one of claims 1 to 12, characterized

characterized in that based on different loads, the damage of the component (1) for each of these

Loads is calculated and the total damage is determined by means of linear damage accumulation.

14. The method according to any one of claims 1 to 13, characterized

in that the component (1) is a vehicle component, in particular a spring and coupling element.

15. The method according to any one of claims 1 to 14, characterized

in that the component (1) is a component of a rail vehicle.

A computer program product comprising a program and directly loadable into a memory of a computer, comprising program means for carrying out all the steps of the method according to one of claims 1 to 15, when the

Program is executed by the computer.