DE102019214526A1 - Virtual product qualification by aligning virtual component models - Google Patents

Abstract

The invention relates to a method for virtual product qualification, in particular in the case of a component arrangement of a vehicle, with: - receiving at least one virtual component model (1) from a component of the product; - receiving target position information (xt) of the component; - aligning the component model ( 1) and determining actual position information (xa) as a function of the selected orientation; - automatic determination of at least one changed orientation of the component model (1) taking into account a deviation (ΔRPS) of the actual position information (xa) from the target position information (xt The invention also relates to a computer device for virtual product qualification.

Description

The invention relates to a method and a computer device for virtual product qualification, wherein the product can in particular be an add-on component arrangement of a vehicle. In general, the invention is directed to the field of virtual process optimization and in particular the virtual optimization of the manufacturing process and, more precisely, the mounting process of add-on parts to a vehicle.

Vehicles have a large number of so-called add-on parts that are attached to a vehicle structure and in particular vehicle body and shape the external appearance of the vehicle. Examples are a hood, all doors and flaps, fenders, bumpers and the like. As part of the vehicle development, it must be ensured that these add-on parts also meet all mechanical and dimensional requirements, for example, when installed. It is known to use a real so-called master jig for this. This is a structure usually made of a metal profile, the position tolerances of which for connection points are very tight, so that the dimensions and dimensions almost correspond to the CAD model. The add-on parts are then mounted on this structure and aligned in a reference point system. In this way it can be determined, for example, whether the components meet all tolerance requirements, are impermissibly deformed or can be positioned in the desired manner relative to neighboring components. If this is not the case, the manufacturing process and / or the basic design of the component can be adapted.

Such real investigations are time-consuming and can only be carried out in a relatively late phase of development, e.g. when tools for component production already exist. In general, the aim is therefore to carry out a corresponding product qualification virtually in advance as far as possible in order to identify any necessary adjustments to the components as early as possible.

From the DE 103 57 413 A1 it is known, for example, to generate CAD models of components and to simulate their installation situation on the basis of these models. For example, changes in geometry due to installation boundary conditions and / or joining processes carried out during installation can be simulated. A tolerance analysis can also be carried out in which the installation position of a component is varied via permissible tolerances in order to control the risk of undesired collisions with neighboring components. Only then, after the component position has been varied accordingly, are any deformations determined using so-called substitute models, which model the effects of external forces on the components.

The variation in the installation position (which can generally also be referred to as aligning the components) takes place manually by aligning the corresponding virtual models. The subsequent determination of any deformations is used to adapt the basic construction of the components. Another essential thought of the DE 103 57 413 A1 it is also to verify the virtual component models and also substitute models by measuring real manufactured components and to adapt them appropriately.

Overall, this teaching is thus characterized by the high level of manual operating effort that is still required. In order to achieve sufficient accuracy, real component measurements are required, which requires the presence of these components and thus the preparation of tools in advance.

One object of the present invention is therefore to improve product qualifications in particular for vehicle development and also in particular for add-on parts of a vehicle body and body parts themselves, and preferably to accelerate and / or make them cheaper. In particular, compared to physical product qualifications, the present virtual solution offers a higher level of repeatability and more flexible analysis options, since case studies or variation calculations can be carried out significantly faster than the construction of new or modified physical components.

This object is achieved by the subjects of the independent claims. Advantageous further developments are given in the dependent claims. It goes without saying that all of the explanations and features mentioned in the introduction also apply to the present invention or can be provided in it.

According to the invention, it was recognized that, in addition to the disadvantages already explained, there is no methodical procedure available for determining a suitable alignment (ie installation position or also installation position) within the higher-level product. In particular, an operator should be relieved of manual entries as much as possible, which has not been the case up to now. Instead, the invention proposes to automatically determine and calculate an orientation (preferably along or around all six spatial degrees of freedom) of a virtual component model, in particular in consideration of deviations of a current position (i.e. a three-dimensional position and orientation) from a desired position. In other words, a deviation between an actual position and a target position can be considered and the alignment can be suitably adapted based on this in a computer-controlled manner. The operator then does not have to manually insert the component model into a higher-level product model and align it there, while monitoring the tolerance limits to be observed. Instead, a suitable alignment is suggested to him, preferably automatically, and / or this alignment is automatically implemented in a virtual component model.

According to preferred embodiments, the described target / actual deviation can be minimized, e.g. by means of an optimization calculation and / or an optimization algorithm. According to likewise preferred embodiments, a suitable alignment can also be determined taking into account stress states (ie deformation states) of a component model. In particular, not only the described target / actual deviation can then be minimized, but also the corresponding voltage state. In this way, an alignment of the component can be determined which assumes a target alignment as precisely as possible, but at the same time also reduces and preferably minimizes component stresses occurring during installation. Optimization calculations can also be carried out for this. As a boundary condition, the scope for movement of connection points to the vehicle body (e.g. of screw holes) can be taken into account and used. It can also be taken into account and exploited that certain deviations from a desired target alignment can be accepted if the stresses are as low as possible for them. In a manner known per se, tensions in built-in components can result from their own weight or from the assembly of the components to a higher-level product, the connection points of the components having to be moved relative to one another and thus the component structure lying between them being deformed.

The invention can in particular be carried out on a virtual master jig in the context of a virtual attachment process of add-on parts or general components that are to be aligned via connection points of the type described here. The process previously carried out in real life can therefore be simulated in advance, in which add-on parts are attached to an underlying vehicle structure (which is to be equated with a higher-level product) in order to ensure, for example, its accuracy of fit and freedom from deformation.

The invention can be carried out by means of a computer device, preferably fully automatically or with only limited user input. For example, all of the procedural measures or procedural steps described herein can be carried out automatically, but if necessary triggered or checked by a user. For example, the user can specify in general which component is currently to be verified with regard to its installation situation. If necessary, the user can also specify in which area of a higher-level product (e.g. a CAD model, a measured master jig or a simulated vehicle body) this component is to be inserted and, if necessary, also define a rough initial alignment. Then, however, a suitable alignment and thus installation position of the component can preferably be determined automatically, with the optimization calculations described here being able to be carried out or generally target / actual position deviations as well as stress states being reduced and possibly also minimized.

The virtual component models that are considered according to the method cannot be ideal or error-free component models. Instead, they can be component models with tolerances and / or component models that have been generated by means of simulated manufacturing processes. In this way, the subsequent real assembly process can be approximated in a particularly realistic manner, in which components that are also subject to tolerances and not ideal are used. In principle, however, it is also possible to generate the component models from the measurement of real components, for example by means of a three-dimensional coordinate measurement of the real components.

• - Erhalten wenigstens eines virtuellen Bauteilmodells (z.B. ein CAD-Modell oder ein FEM) von einem Bauteil des Produkts (beispielsweise einem Anbauteil des Fahrzeugs);
• - Erhalten von Soll-Lageinformationen des Bauteils bzw. Bauteilmodells, vorzugsweise in einem vorbestimmten Referenzkoordinatensystem;
• - (virtuelles) Ausrichten des Bauteilmodells (z.B. in einem virtuellen Produktmodell (z.B. einem virtuellen Fahrzeugmodell) und/oder in dem Referenzkoordinatensystem) und Bestimmen von Ist-Lageinformationen in Abhängigkeit der gewählten Ausrichtung;
• - automatisches Ermitteln oder, mit anderen Worten, Berechnen (insbesondere auf Basis einer Optimierungsberechnung) wenigstens einer geänderten Ausrichtung des Bauteilmodells unter Berücksichtigung einer Abweichung der Ist-Lageinformationen von den Soll-Lageinformationen.

In particular, the invention proposes a method for virtual product qualification, wherein the product can in particular be a vehicle and furthermore in particular an (attached) component arrangement of the vehicle. The procedure may include:

• - Obtaining at least one virtual component model (for example a CAD model or an FEM) from a component of the product (for example an add-on part of the vehicle);
• - Obtaining nominal position information of the component or component model, preferably in a predetermined reference coordinate system;
• - (virtual) alignment of the component model (for example in a virtual product model (for example a virtual vehicle model) and / or in the reference coordinate system) and determining actual position information as a function of the selected alignment;
• - automatic determination or, in other words, calculation (in particular on the basis of an optimization calculation) at least one changed alignment of the component model, taking into account a deviation of the actual position information from the target position information.

All of the positional information mentioned herein can be determined in the reference coordinate system or be defined in relation to it. Position information can generally describe the three-dimensional position of a component model or component, that is, its three-dimensional position and three-dimensional orientation. In other words, all six spatial degrees of freedom of a component or component model can be defined by the position information.

The target position information can be defined, for example, by a desired and preferably error-free ideal alignment of the component in the higher-level product. It can be a position in which the component complies with all permissible deformation limits, all required gap dimensions and / or all other requirements.

In order to obtain the virtual component model, according to the method, such a component model can be loaded, e.g. from a memory. It can also be requested from a corresponding storage device, e.g. if a user has identified a component to be aligned. The same applies to the target position information, which can be read from a memory in particular in accordance with a received component model.

The initial alignment of the component model can be carried out automatically by assuming a predetermined position in the reference coordinate system with, for example, a standard predefined alignment (e.g. of a component coordinate system relative to the reference coordinate system). Alternatively, an operator can align the component model at least roughly initially. The actual position information can then preferably be determined automatically in the reference coordinate system, for example on the basis of predetermined reference points explained herein.

The changed alignment can be calculated using the optimization algorithms explained below. In general, it can be determined which degree of freedom of the component and in particular which possible reference point is to be changed in which way in order to reduce the target / actual deviation. In particular, the changed alignment can take place in such a way that the target / actual deviation is reduced or even minimized.

In general, the target / actual deviation as a result of the alignment adjustment can also be at least within a specified acceptable tolerance range of e.g. less than 1 mm and e.g. +/- 0.0 mm.

The changed alignment can be carried out automatically, e.g. by calculating it with the aid of a computer.

When determining the changed alignment, mechanical connection points of the component model can also be taken into account, as explained below. In particular, it can be taken into account that these must be brought into agreement with corresponding connection points of the superordinate product, for example in order to achieve a predetermined relative arrangement and, in particular, overlap and / or concentric arrangement with this or with this. For example, it can be boreholes which have to be brought into overlap with corresponding boreholes in the product or the vehicle body. In this case too, however, certain tolerances with regard to the accuracy of the relative arrangements described may be permissible. Depending on the specific alignment selected and / or depending on the deviation from an ideal relative arrangement, the stress states explained below can then occur, for example.

According to a further development, it is provided that a stress state of the component model can be determined on the basis of the actual position information and the changed alignment of the component model is also determined taking into account the stress state.

In this case, the component model can in particular be an FEM (finite element model) and the stress state can be determined with the aid of this FEM. Depending on the chosen alignment, e.g. a dead weight can have a different effect and other stress states can occur. Other stress states can also occur because mechanical connection points are then aligned differently relative to corresponding connection points on the higher-level product (e.g. on the vehicle) and the attachment of the component to the product then requires a certain deformation, e.g. for concentric arrangement of these connection points.

In particular, the alignment can be changed in such a way that the stress state is reduced and preferably minimized. In this case, compliance with desired relative arrangements of connection points and / or the generally desired reduction in the target / actual deviation of the position information can be taken into account as additional boundary conditions.

In general, the alignment can be changed randomly or based on rules (e.g. by means of optimization algorithms) and the deviations and / or the stress state can then be calculated in each case. This can be done until a desired reduction in the target / actual deviation and / or the stress state occurs (e.g. a preferably global minimum has been found). It is preferred to determine the changed alignment by means of optimization calculations explained below. To determine a desired alignment, the position of mechanical connection points (joining points) can be varied, e.g. as part of an optimization calculation.

In summary, it can be provided that the alignment is changed in such a way that the voltage state decreases and the deviation of the actual position information from the target position information is within an acceptable range of values (or also a permissible tolerance range, as explained above). The range of values can be, for example, less than 1 mm and preferably +/- 0.05 mm.

A further development provides that the alignment is changed on the basis of the solution to an optimization problem, with the objective function being to minimize the deviation of the actual position information from the target position information and the voltage state. Examples of such an optimization problem are explained below, in particular with the aid of formula 6. However, it can also be provided that an optimization problem can be implemented without taking this stress state into account and only with the objective function of reducing the target / actual deviations. For example, formula 4 explained below can be used for this purpose. In a manner known per se, the optimization problem can be solved by means of common optimization algorithms.

A further development provides that the location information is based on a reference point system ( RPS ) To be defined. The reference point system can uniquely determine the position of a component or component model in a manner known per se, in particular in relation to a reference coordinate system, which can generally also be a global coordinate system (eg a global vehicle coordinate system). The reference points can be permanently assigned to a component. They can be retained through different simulation levels or simulation applications, for example if a manufacturing process for the component is simulated first and then the installation and alignment of the component model. At least six reference points are preferably provided per component, with at least one reference point being able to be present, for example, on three different sides of the component model.

According to a further embodiment, the changed alignment is also determined taking into account motion boundary conditions of the component (or component model). These motion boundary conditions can result, for example, from the dimensions of neighboring components or generally from the installation space available for this component or component model. Instead of movement boundary conditions, one could also speak of displacement boundary conditions, degrees of freedom of movement or freedom of movement.

According to one variant, the motion boundary conditions can apply to the component model and specifically to reference points of this component model. For each reference point, for example, at least one permissible and / or at least one impermissible spatial degree of freedom can be specified as a boundary condition for movement. For this purpose, at one or at least near a reference point (es) (e.g. in or on the same finite element) Nodes of a finite element model can be defined and corresponding displacement boundary conditions can be specified for these.

The target / actual deviations can be reduced by solving an optimization problem of the type described here, taking these motion boundary conditions into account. In particular, provision can be made to spatially fix mechanical connection points of the type described here (i.e. to block movement thereof e.g. by means of movement boundary conditions there and in particular displacement boundary conditions). On the other hand, coordinates of the reference points can be varied taking into account their motion boundary conditions, for example until an at least local minimum of the target / actual deviations and / or a stress state is found in the context of an optimization problem.

It has been shown that with this approach, in particular, the installation situation of dimensionally stable components can be reliably detected, for example since deformations of the component over the reference points are relatively small. Alternatively, it can also be provided that the movement boundary conditions are determined based on degrees of freedom of movement of mechanical connection points of the component model relative to the product model. The mechanical connection points can be points of the component model at which mechanical connections (in particular connections produced by means of tools and / or joining processes) to the superordinate product are to be formed or which are predetermined for the formation of these mechanical connections. A simple application without further developing a force fit and / or form fit cannot count as a mechanical connection. It can be taken into account that these connection points with corresponding connection points of the superordinate product must be brought into predetermined relative arrangements already discussed above (e.g. in order to be able to form screw connections). In this respect, the degrees of freedom of movement of the mechanical connection points of the component model can be limited and the component as a whole cannot be aligned as desired (i.e. not removed as desired from the corresponding connection points of the product model).

A further development provides that the stress state is determined on the basis of a finite element model (FEM) of the component and, in particular, a node of the finite element model is assigned to the mechanical connection points and the degrees of freedom of movement are determined via the displacement boundary conditions of a respective node (i.e. a respective mechanical connection point assigned node) can be defined.

In general, the finite element model can be the component model, i.e. also the component model initially obtained. Alternatively, a CAD model can be obtained and the finite element model can be generated on the basis of this CAD model. In a manner known per se, finite element models typically consist of finite elements (which, in contrast to infinitesimal calculus, are finitely small elements) that are connected to one another via nodes. The nodes can be assigned displacement boundary conditions, which result in corresponding possibilities of movement, deformation and displacement. Providing separate nodes for the mechanical connection points and defining their displacement boundary conditions creates an efficient way of exploring and, in particular, automatically determining possibilities for aligning the component (especially if the connection with corresponding connection points of the product model is still to be maintained). In particular, a calculation time for finding suitable changed alignments can be reduced as a result.

The method also provides that the influence of the degrees of freedom of movement of the connection points on the deviations of the actual position information from the target position information is determined (for example by means of a sensitivity analysis based on so-called Latin hypercube sampling and the calculation of correlation coefficients with the following formula 8 according to Pearson or equally with the rank correlation according to Spearman). In particular, an influence of changes in the coordinates of the connection points, which are limited accordingly by the available degrees of freedom of movement, on the target / actual deviation can be determined.

In general, only those degrees of freedom of movement can then be taken into account when determining a changed orientation (that is, changes to the orientations can only take place along or by means of these corresponding degrees of freedom of movement) whose influence is above a minimum limit. For example, only the 10 most influential connection points or degrees of freedom of movement can be taken into account. This reduces the number of optimization parameters (or more precisely input parameters of an optimization problem), so that finding a solution to the optimization problem described above can be simplified. Also regardless of whether a In this way, an improved alignment can be found quickly in the optimization calculation, since fewer parameters have to be varied.

• - wenigstens ein virtuelles Bauteilmodell von einem Bauteil des Produkts zu erhalten (und/oder aus einem Speicher der Computervorrichtung auszulesen);
• - Soll-Lageinformationen des Bauteils zu erhalten oder zu ermitteln (beispielsweise in einem vorbestimmten Referenzkoordinatensystem);
• - das Bauteilmodell auszurichten und Ist-Lageinformationen in Abhängigkeit der gewählten Ausrichtung zu bestimmen;
• - wenigstens eine geänderte Ausrichtung des Bauteilmodells unter Berücksichtigung (bevorzugt des Spannungszustands und) einer Abweichung der Ist-Lageinformation von den Soll-Lageinformationen automatisch zu ermitteln.

The invention also relates to a computer device (e.g. comprising at least one processor device) for performing a virtual product qualification, in particular an add-on component arrangement of a vehicle, the computer device being set up to

• to obtain at least one virtual component model of a component of the product (and / or to read it from a memory of the computer device);
• To obtain or determine target position information of the component (for example in a predetermined reference coordinate system);
• - align the component model and determine actual position information depending on the selected alignment;
• to automatically determine at least one changed alignment of the component model, taking into account (preferably the stress state and) a deviation of the actual position information from the target position information.

The computing device can generally be set up to carry out a method in accordance with any aspect described herein. In particular, it can have any further developments and features in order to be able to provide all procedural measures or procedural steps. In general, the computer device can be set up by executing program instructions and / or algorithms for executing the measures described here. In particular, the changed alignment can be determined on the basis of an optimization algorithm executed by the computer device.

• 1 und 2 zeigen ein Referenzpunktesystem nach dem Stand der Technik, wie es bei der hierin geschilderten Lösung vorzugsweise zur Anwendung kommt; die Figuren basieren auf folgender Fundstelle: Rai, B.; Shenglan, L.: RPS Alignment of Automotive Body Parts in Virtual Assembly and Deviation Analyses. International Journal of Scientific & Engineering Research, 2016 ;
• 3 und 4 zeigen als Beispiel eines Bauteilmodells ein Kotflügelmodell samt dessen mechanischer Verbindungspunkte und Referenzpunkte;
• 5 zeigt eine Detailansicht eines mechanischen Verbindungspunktes des Kotflügels aus 3 und 4;
• 6 zeigt ein Ablaufschema eines erfindungsgemäßen Verfahrens.

The invention is explained below with reference to the attached schematic figures.

• 1 and 2 show a reference point system according to the prior art, as it is preferably used in the solution described herein; the figures are based on the following reference: Rai, B .; Shenglan, L .: RPS Alignment of Automotive Body Parts in Virtual Assembly and Deviation Analyzes. International Journal of Scientific & Engineering Research, 2016 ;
• 3rd and 4th show, as an example of a component model, a fender model including its mechanical connection points and reference points;
• 5 FIG. 11 shows a detailed view of a mechanical connection point of the fender from FIG 3rd and 4th ;
• 6th shows a flow chart of a method according to the invention.

In 1 and 2 is a reference coordinate system for an exemplary virtual component model 1 shown. This is designed schematically as a cube, but can also be a fender below. A reference coordinate system is also shown 2 , which can be, for example, a global coordinate system of a superordinate product model and in particular of a vehicle model.

It is shown that a plurality of reference points P on different sides and in the present case on at least three different sides of the component model 1 is distributed. A possible range of values for the degrees of freedom of the component model is then successively displayed 1 limited. In other words, the position of these reference points is used P the spatial location (i.e. the position and orientation) of the model 1 in the reference coordinate system 2 of the component model 1 clearly defined.

In detail are in 1 the individual reference points P shown and based on 2 there are deviations from the nominal and actual coordinates of these points P explained. The figures represent basic principles as they are already used in the prior art.

First referring to 1 it is shown that three points P1-P3 in a common plane or on a common side of the component model 1 are arranged. The points P1-P3 define a base plane which, in the case shown, blocks a translational movement in the y-direction (ie defines the degree of freedom in y) as well as the degree of freedom of rotation about the x- and z-axes. The further reference points P4-P5, which are on a common side and in particular on a plane running at an angle to and preferably orthogonally to the base plane, define the translational degree of freedom in z and the rotation or orientation around the y-axis. The remaining one is not yet limited Degree of freedom (translation in x) is limited by point P6. This lies on a plane which is at an angle to the two preceding planes and preferably runs orthogonally to them. Such a definition of reference points P represents the minimum number of reference points P to define the spatial position of a three-dimensional object, the reference points P are distributed according to the so-called 3-2-1 rule. Significantly more reference points can also be used P be provided as shown below.

Referring to 2 is shown that a target location of the model 1 can differ from an actual position, for example due to a deformation under its own weight or a deformation due to a certain installation situation. A setpoint / actual deviation for reference point P6 is shown as an example, the deviating actual reference point being denoted by P6a. It can be seen that the deviation is present along the x-axis, a target position being denoted by xt and an actual position being denoted by xa. The coordinates of the actual reference point P6a are obtained by projecting the nominal coordinates of P6 onto the surface of the actual component model, to be precise preferably along a normal direction on the plane P6. A target / actual deviation can then be expressed as the difference between the corresponding coordinates xt and xa are formed. It must be taken into account that, according to a generally preferred and common definition, the reference points P are each freely displaceable in at least one specified direction or are defined as displaceable bearings in this direction. The reason is that they are each only based on the location along the above 1 should limit and define the axes explained or around these axes. In this respect, a translational deviation in x is only determined for the reference point P6 provided or defined for this purpose, but not for the other points P that cannot be moved in x.

In general, corresponding target / actual deviations can be determined using Formula 1 below: $$\mathrm{\Delta }\mathrm{R.}P\mathrm{S.}=\{\begin{array}{c}{x}^a-x^t\\ y^a-y^t\\ z^a-z^t\end{array}$$

In 3rd and 4th are component models 1 shown according to a real application. This is a model of a fender and, more precisely, a finite element model (FEM). This is to be attached to a higher-level product model of a vehicle or a vehicle body, which is not shown separately, to which several other add-on parts can also be attached or are still attached as part of the product qualification. The entire process can take place virtually and / or automatically by means of a symbolically entered computer device 100 in 1 . Positions of selected mechanical connection points are shown 10 that are to be connected to corresponding connection points of the body structure (or its virtual model). To do this, these connection points must 10 For example, they can be brought into overlap with these corresponding connection points within permissible tolerances. Then, for example, screw connections can also be realized under real conditions. At the connection points 10 it can be holes, locking projections or the like.

In 4th are also reference points P registered. These are each identified by their own number (from 001 to 105). In addition, the meaning is indicated, for example whether it is a main reference point designated with uppercase letters or a so-called auxiliary reference point (lowercase letters). The latter can be used in a known manner to detect local deformations. The associated x, y, and z axes are also designated, in which the model is defined. The definition of these reference points P can be done in advance depending on the component model and is known in principle in the prior art. The reference points P can be saved as part of a model data set and then read out as required.

In 5 is a detailed view of the component model 1 out 3rd shown. More precisely, a section is shown which has a connection point 10 comprises in the form of a recess. This recess can be used, for example, to form a screw connection. The networking of the finite element model is also indicated. Optionally, these screw connections can be modeled more realistically with the help of replacement elements, for example by integrating tightening torques.

A geometric center of the recess and thus the connection point 10 is defined as an additional node of the network or the superordinate finite element model. At this Point is also a reference coordinate system 3rd entered and can be defined in a known manner displacement boundary conditions, which allowable movements around and along the axes of this reference coordinate system 3rd define. The extent of these permissible movements around and along the axes can also be specified. This extent can be achieved, for example, by permissible tolerances in the relative arrangement of the connection point 10 be defined to a corresponding connection point, not shown, on the vehicle body.

Based on 6th and the following formulas are now a possible calculation method for determining a changed alignment and in particular an optimal alignment, with the target-actual deviations, but also a stress state of a component model 1 are reduced in a desired manner, explained. In a measure S1, the following vector is first set up according to formula 2. The translational degrees of freedom of each point (which can also have a certain two-dimensional extension (e.g., a borehole)) are designated by x, y, z and the three rotational degrees of freedom by α, β, γ. The number of connection points 10 is denoted by n and relates, for example, to nine (see also 3rd ). This results in 54 entries or degrees of freedom within the vector according to formula 2. $${\phi }^T=\left(x_1^{\text{'}},{\text{y}}_1^{\text{'}},{\text{z}}_1^{\text{'}},{\alpha }_1^{\text{'}},{\beta }_1^{\text{'}},{\gamma }_1^{\text{'}},...,x_n^{\text{'}},{\text{y}}_n^{\text{'}},{\text{z}}_n^{\text{'}},{\alpha }_n^{\text{'}},{\beta }_n^{\text{'}},{\gamma }_n^{\text{'}}\right)$$

In the case of a measure S2, the component model is aligned 1 in the reference coordinate system 2 determined (e.g. based on an analog state, as in 3rd and 4th shown). The actual position is based on the coordinates of the reference points P out 4th Are defined. The target position (or the target position information) is read out from a memory of the computer device 100 and is stored there, for example, in a database and / or can be determined from an ideal virtual vehicle model. There will then be for each of the reference points P out 4th which are generally based on 2 explained deviations ΔRPS as defined in Formula 1. For all reference points P (m = 11), therefore, a vector δ can be determined according to Formula 3 below. $${\mathrm{\delta }}^{\text{T}}=\left(\mathrm{\Delta }\mathrm{R.}P{\mathrm{S.}}_1,...,\mathrm{\Delta }\mathrm{R.}P{\mathrm{S.}}_{\mathrm{M.}}\right)$$

This can be a general output vector of an installation simulation carried out for the component model 1 act in a superordinate vehicle model.

$$\mathrm{\Phi }=\left\{{\phi }_i\in ℝ\mid {\phi }_i^l\le {\phi }_i\le {\phi }_i^{u}i=\mathrm{1,}\dots \mathrm{,54}\right\}$$

A procedure for changing the orientation is now explained below. For this purpose, an optimization problem according to the following formula 6 is defined and solved in a measure S3. This formula is derived from formula 4 shown below, into which the value Φ according to formula 5 below is to be inserted. $${\mathrm{<figure-callout\; id="1"\; label="\omega "\; filenames="DE102019214526A1\_0009.png,DE102019214526A1\_0010.png"\; state="\{\{state\}\}">\omega </figure-callout>}}_1^{*}\left({\mathrm{<figure-callout\; id="1"\; label="\phi "\; filenames="DE102019214526A1\_0009.png,DE102019214526A1\_0010.png"\; state="\{\{state\}\}">\phi </figure-callout>}}_1^{*}\right)=\underset{\phi }{{\displaystyle min}}\left\{{\omega }_1\left(\phi \right)\mid \phi \in \mathrm{\Phi }\right\}=\underset{\phi }{{\displaystyle min}}\left\{{\displaystyle \sum _{e=1}^m{\omega }_e\frac{\mathrm{\Delta }\mathrm{R.}P{\mathrm{S.}}_e^2\left(\phi \right)}{s^2}|\phi \in \mathrm{\Phi }}\right\}$$

$$\mathrm{\Phi }=\left\{{\phi }_i\in ℝ\mid {\phi }_i^l\le {\phi }_i\le {\phi }_i^{u}i=\mathrm{1,}...\mathrm{,\; 54}\right\}$$

Formula 4 initially defines as the output optimization problem that the deviations, as they are based on 3rd should be minimized. For this purpose, the objective function is defined according to the corresponding formula 4. All deviations from the aforementioned Art ΔRPS are raised to the power of 2 and then multiplied by a weighting factor w, which is optional but preferred. The weighting factor w can be the main reference points explained above P towards auxiliary reference points P weight more heavily. The coefficient s can optionally be added as a scaling factor to enable control of the dimension or unit obtained. This is particularly useful in the context of the following supplementary formula 5, in which the dimensions of the deviations ΔRPS (typically in millimeters) and the voltages (typically in pascals) should be unified.

It can be seen from formula 5 that every value φ has a lower limit φ ^l and an upper limit φ ^u is assigned and thus the specific value φi lies between these limits or can maximally correspond to them. This takes into account that a position of the connection points 10 due to the above The required relative arrangement described in relation to corresponding arrangement points on the vehicle cannot be varied at will.

In principle, the alignment can be suitably changed using formula 4, as can the target / actual deviation, taking into account the limited degrees of freedom of movement of the arrangement points 10 be minimized. For this purpose, in S5 below, for example, iteratively φi values are selected within the stated limits, which results in a changed alignment in each case, and the associated deviations are then determined ΔRPS certainly. This can be done until a local or global minimum is found. However, this can lead to the fact that a solution is found in which a stress state of the component model 1 is undesirably increased.

A preferred development therefore provides for stress states of the component model as part of measure S3 1 to be determined and taken into account. Specifically, another optional measure is defined as the target / actual deviations ΔRPS may vary within a range of values of +/- 0.05 mm so that they can still be viewed as acceptable. On the other hand, however, tensions should be as low as possible.

$$\mathrm{\Phi }=\left\{{\phi }_i\in ℝ\mid {\phi }_i^l\le {\phi }_i\le {\phi }_i^{u}i=\mathrm{1,}\dots \mathrm{,54}\right\}$$

In formula 6 below, in which the expression according to formula 7 is to be inserted, the stress state is calculated in a manner known per se _{as a comparison stress according to von Mises, namely as a 0.95 quantile σ Q95 of all elements within the component model} 1 . This limits the influence of singularities. The corresponding objective function according to 6th is reproduced below: $${\mathrm{<figure-callout\; id="2"\; label="\omega "\; filenames="DE102019214526A1\_0009.png,DE102019214526A1\_0010.png"\; state="\{\{state\}\}">\omega </figure-callout>}}_2^{*}\left({\mathrm{<figure-callout\; id="2"\; label="\phi "\; filenames="DE102019214526A1\_0009.png,DE102019214526A1\_0010.png"\; state="\{\{state\}\}">\phi </figure-callout>}}_2^{*}\right)=\underset{\phi }{{\displaystyle min}}\left\{{\omega }_2\left(\phi \right)\mid \phi \in \mathrm{\Phi }\right\}=\underset{\phi }{{\displaystyle min}}\left\{{\displaystyle \sum _{e=1}^m{\omega }_e\frac{\mathrm{\Delta }\mathrm{R.}P{\mathrm{S.}}_e^2\left(\phi \right)}{s^2}+{\omega }_{\sigma }\frac{{\sigma }_{Q95}^2\left(\phi \right)}{s_{\mathrm{<figure-callout\; id="2"\; label="\sigma "\; filenames="DE102019214526A1\_0009.png,DE102019214526A1\_0010.png"\; state="\{\{state\}\}">\sigma </figure-callout>}}^2}|\phi \in \mathrm{\Phi }}\right\}$$

$$\mathrm{\Phi }=\left\{{\phi }_i\in ℝ\mid {\phi }_i^l\le {\phi }_i\le {\phi }_i^{u}i=\mathrm{1,}...\mathrm{,\; 54}\right\}$$

In this way, the voltage can thus be minimized, while at the same time the target / actual deviations can also be reduced or at least lie within the permissible range of values explained above.

In an optional measure S4, which can also be carried out before S3 or at least its solution, finding a solution to the described optimization problem is simplified by means of preparatory measures. More precisely, it is taken into account that the input vector φ comprises a large number of variables (namely 54), which initially make the optimization problem significantly more difficult. A sensitivity study is therefore carried out by means of what is known as Latin hypercube sampling (LHS) in order to determine what influence these variables have on the target / actual deviations, for example according to formula 3 above. This procedure is known in principle from a mathematical point of view and is therefore not explained in detail. In this way, however, it can be determined which changes in one of these variables correlate particularly strongly with which changes in the target / actual deviations. Figuratively speaking, those connection points can 10 whose changed position has the greatest influence on the target / actual deviations or offers the greatest potential for improvement. The correlation can be determined in a manner known per se according to Pearson according to the following formula 8, where I is the number of samples. For the sake of simplicity, correlations can be assumed to be linear in this case. In addition, rank correlations according to Spearman can be calculated, which in the present example deliver almost identical results, although this method does not assume any linear relationships. $${\mathrm{C.}}_P\left({\phi }_i,{\delta }_j\right)=\frac{{\displaystyle {\sum }_{k=1}^l\left({\phi }_{i,k}-{\overline{\phi }}_i\right)\left({\delta }_{j,k}-{\overline{\delta }}_j\right)}}{\sqrt{{{\displaystyle {\sum }_{k=1}^l\left({\phi }_{i,k}-{\overline{\phi }}_i\right)}}^2}\sqrt{{\displaystyle {\sum }_{k=1}^l{\left({\delta }_{j,k}-{\overline{\delta }}_j\right)}^2}}}$$

As a result, only those variables can be further considered in step S4 which have the greatest influence on the target / actual deviations or whose influence is above a specified minimum limit. For example, only the 10 variables with the greatest influences can be considered.

In a subsequent step S5, the optimization problem (that is to say in particular the above formula 6) can then be completely solved. In particular, a so-called Sequential Quadratic Programming (SQP) algorithm, which is based on the quasi-Newton method, can be used for this purpose. Since the number of (influential) variables to be taken into account was reduced in step S4, the solution of the optimization problem is simplified accordingly.

List of reference symbols

1

Component model

2.3

Reference coordinate system

10

Connection point

P

Reference point

xt

Target position information

xa

Current situation information

ΔRPS

Target / actual deviation

RPS

Reference point system

w1 *, w2 *

Objective function

φ ^l , φ ^u

Motion boundary conditions

QUOTES INCLUDED IN THE DESCRIPTION

This list of the documents listed by the applicant was generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Patent literature cited

• DE 10357413 A1 [0004, 0005]

Non-patent literature cited

• Rai, B .; Shenglan, L .: RPS Alignment of Automotive Body Parts in Virtual Assembly and Deviation Analyzes. International Journal of Scientific & Engineering Research, 2016 [0040]

Claims (10)

Method for virtual product qualification, in particular for a component arrangement of a vehicle, with: - Obtaining at least one virtual component model (1) from a component of the product; - Obtaining nominal position information (xt) of the component; - Alignment of the component model (1) and determination of actual position information (xa) as a function of the selected alignment; - Automatic determination of at least one changed alignment of the component model (1), taking into account a deviation (ΔRPS) of the actual position information (xa) from the target position information (xt). Procedure according to Claim 1 , further characterized by : - determining a stress state of the component model (1) on the basis of the actual position information (xa); and - determining the changed alignment of the component model (1) also taking into account the stress state. Procedure according to Claim 2 , characterized in that the alignment is changed in such a way that the voltage state decreases and the deviation of the actual position information (xa) from the target position information (xt) lies within an acceptable range of values. Method according to one of the preceding claims, characterized in that the alignment is changed on the basis of the solution to an optimization problem in which the target function (w1 *, w2 *) is the deviation of the actual position information (xa) from the target position information (xt) and also the state of tension should be minimized. Method according to one of the preceding claims, characterized in that the target and actual position information (xt, xa) are defined on the basis of a reference point system (RPS). Method according to one of the preceding claims, characterized in that the changed alignment is also determined taking into account motion boundary conditions (φ ^l , (φ ^u ) of the component. Procedure according to Claim 6 , characterized in that the movement ^{boundary conditions (φ l} , (φ ^u ) are determined based on degrees of freedom of movement of mechanical connection points (10) of the component model (1) relative to a product model. Procedure according to Claim 7 , characterized in that the stress state is determined on the basis of a finite element model of the component, and in particular with the mechanical connection points (10) each being assigned at least one node of the finite element model and defining the degrees of freedom of movement via displacement boundary conditions of a respective node will. Procedure according to Claim 7 or 8th , characterized in that the influence of the ^{degrees of freedom of movement (φ l} , (φ ^u ) of the connection points (10) on the deviations (ΔRPS) of the actual position information (xa) from the target position information (xt) is determined, and only that straight line of freedom of movement be taken into account when determining a changed alignment, the influence of which is above a minimum limit. Computer device (100) for performing a virtual product qualification, in particular in the case of a component arrangement of a vehicle, which is set up to - To obtain at least one virtual component model (1) of a component of the product; - To obtain target position information (xt) of the component; - to align the component model (1) and to determine it from actual position information (xa) as a function of the selected alignment; - to automatically determine at least one changed alignment of the component model (1) taking into account and a deviation (ΔRPS) of the actual position information (xa) from the target position information (xt).

Images