EP3271233B1 - Assembly with flexible link and flexible link for such an assembly - Google Patents

Description

Field of the invention

The present invention relates to a flexible connection assembly and a flexible connection for such an assembly. It applies in particular to the fixing between composite panels and metal structures of large dimensions.

Technological background

During the manufacture of large-sized structures such as aircraft or train bodies, it is necessary to assemble structures of several tens of meters. This is the case, for example, with the connection between the front panel and the lower part (stretcher) of a train body.

An example of a train body is described in the document EP 0 891 910 A1 .

These panels, initially in metallic materials, are to be replaced at least partially by panels in composite materials which make it possible to design and produce even lighter structures, which is particularly important for the transport industry.

Whether they are metallic or made of composite materials, large-size panels which have a structural function are often sandwich-type panels, that is to say they consist of two skins joined together by various well-known techniques. It is generally the technology which makes it possible to give the panels stiffness and inertia for the lowest possible optimized mass.

• Des tolérances générales liées à l'empilement des couches élémentaires de chaque panneau à assembler
• Des tolérances de forme des pièces

Like all manufactured parts, these panels have manufacturing tolerances, which are in fact of two kinds:

• General tolerances linked to the stacking of the elementary layers of each panel to be assembled
• Part shape tolerances

However, because of the large dimensions, the parts to be assembled have form tolerances which can induce clearances which can reach several millimeters between the parts to be assembled. These clearances are not acceptable for many reasons, mechanical, tightness held over time, and therefore the definers use different methods to fill these clearances.

We can first use panels whose shape tolerances are very low, significantly lower than those of standard products, either by better controlling the manufacturing process, or by performing machining after this manufacturing: it is understood that this type of solution is particularly expensive. Or we can fill the gaps:
For example, in aeronautics, a “rigid filling” is very often practiced, that is to say that metal shims are introduced into the intervals, which requires the clearances to be measured very precisely on a case-by-case basis.

When the play is small - a few tenths of a mm, it can be satisfied with filling it with a putty.

In other industries, it is possible to seek to take up the clearances, and to use the flexibility of the panels and to deform them until the clearances are eliminated, for example by means of screws or rivets.

• Les technologies aéronautiques sont très performantes mais aussi très coûteuses (mesures et réalisation de comblages spécifiques et précis, ré usinage de grandes pièces, ..)
• Les technologies par déformation sont moins coûteuses, mais elles nécessitent que les matériaux présentent de grandes déformations plastiques, et introduisent des contraintes résiduelles dans les matériaux qui rendent les assemblages mécaniquement moins performants, en statique mais aussi en fatigue.

All of these methods have drawbacks, and generally:

• Aeronautical technologies are very efficient but also very expensive (measurements and production of specific and precise fillings, re-machining of large parts, etc.)
• Deformation technologies are less costly, but they require the materials to exhibit large plastic deformations, and introduce residual stresses in the materials which make the assemblies less efficient mechanically, in statics but also in fatigue.

To fix ideas include the NF EN 755-9 standard in the version in force on ¹ March 2015, which defines acceptable form tolerances during manufacturing of Aluminum profiles, evaluates the games can usually expect an overall straightness tolerance of 1.5 mm / m, i.e. 15 mm for a part 10 m long to which is added a tolerance of local straightness of 0.6 mm / 300 mm, i.e. a set of 30 mm for an assembly of two metal parts 10 m long.

According to figure 1 , to connect such parts, it is possible to provide each part with an end wing and to assemble the parts head to tail so that the end wings are fixed on opposite faces of the parts, in particular by riveting: the assembly sets will be located at the junction between the end wings and the opposite faces, areas where it will be necessary to provide solutions to fill these clearances

We are interested here in solutions that allow an optimal compromise between the technical performance of the assembly and its economic cost: it is a question of using an assembly method that is as simple as possible, which is as least demanding as possible. with respect to the manufacturing tolerances of the parts and which induces only a minimum of residual stresses in the assembled parts, so as to preserve the structural strength over time of the assembled parts.

The aim of the solution envisaged here is to obtain contact between the two parts of the connection which will give correct mechanical strength of the connection.

Brief description of the invention

To do this, the present invention provides for an end connection between two panels according to claim 1.

According to a particular embodiment, said straight edge makes an angle of 90 ° ± 15 ° with respect to said faces of the skins of the second panel.

According to an alternative or complementary embodiment, said beveled edge makes an angle of 45 ° ± 15 ° with respect to the normal to said faces of the composite panel.

Said beveled edge is advantageously produced by folding the first skin of the composite panel according to a first inclination at the start of the bevel then according to a second inclination at the end of the bevel where the first skin joins the second skin of the panel to form a first so-called tongue. support tab at the end of said bevelled edge.

Said straight edge is advantageously a wall connecting the two skins of the second panel and disposed at the end of a first of these skins, said catching tab of the second panel being constituted by an extension of a second of its skins cantilevered by report to said right edge.

According to a particular embodiment, the second panel comprising said straight edge is a panel or a metal structure.

Said take-up tab is made so as to have a flexibility adapted to take up the differences in thickness of the panels by bending its free intermediate part.

In particular, by making a catching tongue and a bevel that are sufficiently long, the free intermediate part can be sufficiently deformable to come into contact with the face of the panel receiving it and allow riveting while minimizing the stresses induced by the assembly.

According to a preferred aspect of the invention, the connection comprises a window of substantially triangular rectangle section for which said beveled edge constitutes the hypotenuse of said section, the straight edge and said free intermediate part constituting the other two sides of said section.

The invention further provides an assembly of two panels according to claim 10.

The support wall of the second panel is advantageously an edge set back from the face without a tongue of said panel so as to compensate for the thickness of the support tongue so as to have the flat faces outside the two panels in continuity.

The support wall of the panel comprising said straight edge is advantageously an edge set back from the face without a tongue of said panel and connected to said face by a heel forming a stop for the support tongue.

The take-up tab is preferably adapted to bend under the action of the assembly of the first panel and the second panel either in the direction of the first panel or in the opposite direction.

According to a particular embodiment, the flexibility of said take-up tab is calculated to allow it to flex by a distance at least equal to the sum of the tolerances of the layers which constitute each panel. According to an alternative embodiment, the flexibility of said take-up tab is calculated to allow it to flex by a distance at least equal to the quadratic sum of the tolerances of the layers which constitute each panel.

For the quadratic sum we take the square root of the sum of the squares of the uncertainties of each layer.

The deflection of said take-up tab is measured at the docking point between said tab and the face of the panel on which it is applied.

The invention also applies to a mixed composite panel / metal structure assembly for which the second panel is a metal structure.

The metal structure is advantageously the panel comprising said straight edge.

Brief description of the drawings

• en figure 1: une terminaison de structure métallique et une terminaison de panneau composite de l'invention en vue de côté coupe et en vis à vis;
• aux figures 2A et 2B: des vues de la terminaison de la figure 1 respectivement pour un jeu négatif et un jeu positif;
• en figure 3: un détail de la structure de la figure 1 ;
• en figure 4: la modélisation d'une languette de rattrapage de la structure métallique de la figure 1;
• aux figures 5A et 5B: des modélisations des contraintes dans la structure pour les jeux des figures 2A et 2B;
• en figure 6: un exemple de structure recevant un panneau conformément à l'invention;
• aux figures 7A et 7B: des représentations schématique de l'assemblage de la figure 1 au niveau de la languette de rattrapage en vue de face;
• aux figures 8A à 8C: un exemple de rivet utilisable pour l'invention.
• en figure 9: une vue en perspective d'une voiture ferroviaire comportant des panneaux composites fixés avec une structure inférieure métallique au moyen d'une liaison selon l'invention;
• en figure 10: une vue en perspective de l'assemblage d'un partie de panneau et de structure de la figure 9;
• en figure 11 : une vue en coupe de côté de l'assemblage de la figure 10.

Other characteristics and advantages of the invention will become apparent on reading the following description of a non-limiting example of embodiment of the invention with reference to the drawings which represent:

• in figure 1 : a metal structure termination and a composite panel termination of the invention in sectional side view and facing each other;
• to the figures 2A and 2B : views of the end of the figure 1 respectively for a negative clearance and a positive clearance;
• in figure 3 : a detail of the structure of the figure 1 ;
• in figure 4 : the modeling of a retrofit tab of the metal structure of the figure 1 ;
• to the figures 5A and 5B : modelizations of the stresses in the structure for the games of figures 2A and 2B ;
• in figure 6 : an example of a structure receiving a panel in accordance with the invention;
• to the figures 7A and 7B : schematic representations of the assembly of the figure 1 at the level of the take-up tab in front view;
• to the figures 8A to 8C : an example of a rivet that can be used for the invention.
• in figure 9 : a perspective view of a railway car comprising composite panels fixed with a metallic lower structure by means of a link according to the invention;
• in figure 10 : a perspective view of the assembly of part of the panel and the structure of the figure 9 ;
• in figure 11 : a side sectional view of the assembly of the figure 10 .

Detailed description of embodiments of the invention

The present invention falls within the framework of the assembly of panels of considerable length and in particular within the framework of the study of the replacement. of aluminum or steel panels, in particular of railway car bodies, by composite side panels. In this context, as shown in figure 9 , composite panels 300 are attached to metal frame elements 200 making up the base of the cars in order to save weight on the structure and roof panels 301 are attached to said side panels. This makes it possible to increase the payload of the body, whether in terms of the number of passengers or on-board equipment.

One of the constraints of such an embodiment is to minimize the assembly costs of large-sized parts such as those which make up the train body, in particular the side facades, the metal frame and the roof. For this it is necessary to find the ideal compromise between the manufacturing cost of the primary parts and the pure assembly cost. The latter is directed by the time required for handling and the complexity of the operation.

The riveting process is well suited to making connections between large-sized parts and is suitable for assembling the previously mentioned parts.

In the aeronautics sector, many case-by-case solutions are used during assembly (metrology, re-machining of parts, use of machined physical calluses, etc.). In the rail sector concerned, the price accepted per unit of mass gained on a train body is significantly lower than the price accepted for the same unit of mass gained on an aircraft, for example. In addition, the cost of manufacturing parts such as panels must remain low, which does not allow the acceptable dimensional tolerances of the parts to be reduced as much as in aeronautics.

Too restrictive manufacturing tolerances would impose excessively high manufacturing prices. It is known that highly hyperstatic structures made up of large, relatively flexible parts which are subject to geometric defects cannot be assembled without undergoing deformations. This results in a geometrical state and partially mastered residual stresses.

The present invention proposes a new approach which uses the flexibility of a part during the assembly of panels which will allow a greater tolerance on the dimensions of the parts in particular of the thickness. composite panels that will be assembled on metal structures. The present invention also makes it possible to obtain larger form defect tolerances for the manufacture of the parts, therefore at the lowest possible cost, while ensuring the performance of the assembly by using the flexibility of part of the assembly. assembly.

Composite panels are solid products comprising skins or plates made with a binding material called a matrix and a reinforcing material generally in the form of fibers which sandwich a light filling material, foam, honeycomb material, nor material. bee or other called soul.

The matrix preserves the geometrical arrangement of the fibers and transmits to them the stresses to which the part is subjected. The material obtained is thus light, has good resistance to fatigue and has good mechanical characteristics in the direction of the fibers.

In the case of pieces of great length, a defect in thickness or flatness of a panel such as a composite panel will result, for example, in corrugations of the panel along the length of the panel.

The preferred method of assembly for joining two thin large parts is riveting. This process frequently uses so-called blind rivets because they require access to only one side of the assembly. Using a pressure gun, we pull on the shank of the rivet, which allows the head to be crushed on the other side. The rod is sized to break at its base from a certain pressure.

In the case of train bodies, the assembly is made up of two lines of rivets. The first is on the outside of the train, the second is on the inside of the train.

Riveting lines are for example represented under the references 110, 120 on the figure 10 between the panel 300 and the metal structure or frame 200 of a rail car sill that can be used to receive a panel fixed according to the invention.

In an application to the production of train bodies in particular, as a function of the heat and of the hydrometry, the temperature of a train body can rise to temperatures exceeding 80 ° C. important assemblies and in particular composite / metal assemblies which imposes particular sizing constraints.

In the context of the invention and according to figure 1 , the edge of a composite panel 1 to be fixed comprises a return to the skin forming a fixing tab 17 resting on one face of the metal panel 2 and here called a support tab in the sense that it constitutes a surface of Substantially rigid reference support for assembly.

The metal panel 2 for its part comprises a metal tab 24 provided with an end part resting on one face of the composite panel but also provided with a free intermediate part supporting over a length which allows flexibility to be provided for assembly, this tongue being called the take-up tongue. The rivets are positioned at the part of the tabs resting on the faces of the panels and inserted into holes, a series of holes P1 passing through the catching tab and a first skin 10 of the panel and a series of holes P2 passing through the part in support of the take-up tab and a support wall 21 of the frame.

By substantially or essentially rigid reference bearing surface is meant that the bearing tongue is sufficiently rigid so that its possible deflection remains negligible compared to the bending of the take-up tab in the event of a difference in thickness of the panels at the level. of the connection, which corresponds for example to an angular deflection of the support tab 10 to 100 times less than that of the take-up tab at their junction with the edge of the panel carrying them for a difference in thickness.

Likewise, the support tongue rests on substantially its entire length expresses that even if a small part of the support tongue is cantilevered, this small part remains of negligible length, for example up to 10% of the length. length of the tongue and is of insufficient length to provide a significant angular deflection of the support tongue.

Defects in the shape of composite parts are caused in particular by the imperfect shape of the mold. The defects of the mold lead to a first family of defects on the part produced.

A second source of defects comes from deformation stresses after polymerization. The manufacturing process of composite materials leads to internal tensions which are released after polymerization, which leads to an overall deformation of the part by a spring effect called in English "spring back".

These two sources of defects induce shape defects D _f on composite parts.

A third source of defects comes, on the face opposite the mold, from variations in thickness linked to manufacturing dispersions (accumulation of tolerances of the successive layers of materials assembled on the mold): this gives rise to stacking defects.

The composite panel 1 comprises a first skin 10 and a second skin 11 arranged on either side of a core 12 made of foam, honeycomb or other light material. It comprises, at the level of at least one edge of the composite panel, an assembly termination comprising the border 17 returning to the skin produced by joining together the two skins 10, 11 and forming the support tab.

To improve the junction between the skins and the rigidity of the edge of the panel at the level of the termination, the edge 13 of the core is bevelled to produce a zone of gradual transition from the first face 14 to the second face 15 of the panel. , the first skin 10 covering the transition zone from the first face to the second face of the panel to come to rest on the second skin 11 which extends this second face and produce the support tab 17 by returning to the skin of the panel composite.

Due to its constitution by securing the two skins, the support tongue is relatively rigid and not flexible.

The metal structure 2 for its part comprises an end 20 of junction with the composite panel provided with a support wall 21 facing the support tab 17.

This structure further comprises a connecting wall 22 between said bearing wall and a second face 23 of the structure, this connecting wall forming a substantially straight edge.

By substantially or essentially straight edge is meant an edge perpendicular or as perpendicular as possible to the skins or plates 23, 24, 21, 26 taking into account the tolerances of cutting, machining or molding. In the context of the present invention, it is considered as edge substantially or substantially straight with an edge 90 ° ± 15 ° from the surface of the skins or plates.

The tongue 24 extends beyond the right edge on the face 23 of the structure opposite to that carrying the support wall 17. The tongue 24 is cantilevered and comes to rest on the first face 14 of the composite panel. .

The tongue 24 has a free part of length "a" and a resting part of length "b" + "c".

The free part is located above the bevelled edge 13 of the panel 1 and allows the tongue 24 to flex which gives it a function of taking up play and faults as will be seen below.

As regards the bevelled edge, it is considered that a range of 30 ° to 60 °, ie 45 ° ± 15 ° with respect to the normal to the plates, makes it possible to have a free length of the tongue suitable for the invention.

The part of the tongue 24 resting on the first face 14 of the panel and the edge 17 on the support wall 21 constitute zones for receiving fixing means such as rivets.

The defects which the take-up tab 24 makes it possible to catch are in particular the stacking defects of the composite panel and of the structure.

These stacking defects D _e are formed by the sum of the tolerances of all the layers which constitute each panel (skin, souls of the sandwich, etc.). It can be a classic addition, or a quadratic addition (we take the square root of the sum of the squares of the uncertainties of each layer) more realistic from the risk of occurrence point of view, and therefore preferred by the inventors. Its value is of the order of ± 1 millimeter for a stack such as that of the figure 1 , when the panels have a thickness of the order of 50 mm, the skins of the composite sandwich having a thickness of approximately 3 mm.

The tongue is adapted to bend under the action of the assembly of the panel and the structure either in the direction of the panel as shown in figure 2A with a displacement -y corresponding to an under thickness of the panel compared to the structure is in the opposite direction in figure 2B with a displacement + y corresponding to an excess thickness of the panel compared to the structure.

The figure 3 shows an example of a structure 200 comprising a first face 26 terminated at the level of the assembly part with a composite panel by the support wall 21 and which comprises a second face 23 which is extended by the tongue 24. The enlargement in figure 3 makes it possible to distinguish a heel 25 connecting the support wall 21 to the first face 26 of the structure. This heel here makes it possible to obtain a flat outer face between the panel and the structure, the difference in height between the support part 21 and the plate 26 of the structure corresponding to the thickness of the support tab 17 of the panel. composite. The heel can also serve as a stop for the edge 17 of the assembly end of the panel.

As seen above, the tongue 24 is intended by its flexibility to take up the stacking clearances, but it alone cannot take up the clearances due to form tolerances: in fact it is the flexibility of the panel and of the structure. , flexibility due to their large dimension in the direction of the length which makes it possible to catch up this play.

It should be noted that it is well known that this flexibility depends on the shape of the panel and on the structure, and for example a flat panel will be very flexible whereas a curved panel like that of the figure 3 will be much less flexible.

As explained in the previous part, to perform the assembly without having any play to fill, it is necessary to increase the flexibility of the metal part in order to reduce the internal stress at the assembly level. It is the role of the catching tab 24 to provide this flexibility in the assembly, by a suitable design of this tab.

For this adapted design, it is necessary to set a certain number of objectives, linked to the proper functioning of the assembled structure.

• On se donne un critère de contrainte de flambage maximale admissible dans la languette;
• On se donne un critère de contrainte résiduelle maximale admissible dans la languette;
• On se donne un jeux maximum de liaison admissible, susceptible d'un comblage par mastic;
• On définit l'épaisseur du linguet de façon qu'il puisse supporter les efforts en matage dus à la liaison;
• Sur la base de cette épaisseur et du critère de flambage, on en déduit la longueur maximale possible du linguet;
• Sur la base de cette longueur, de l'épaisseur et du critère de contrainte résiduelle, on calcule l'effort maximum de serrage de la liaison, dont on déduit la déformée maximale;
• On compare cette déformée à la tolérance maximale possible due aux empilements des constituants des panneaux, en s'assurant que l'écart maximal entre la déformée et la tolérance maximale est inférieure au jeux maximum de liaison admissible;
• On s'assure qu'avec les efforts ainsi définis, on pourra rattraper les jeux du aux tolérances de forme pour la liaison réelle à réaliser.

In general, for the design of this link:

• One gives oneself a criterion of maximum admissible buckling stress in the tongue;
• One gives oneself a criterion of maximum admissible residual stress in the tongue;
• We give ourselves a maximum permissible bonding clearance, capable of filling with mastic;
• The thickness of the pawl is defined so that it can withstand the matting forces due to the connection;
• On the basis of this thickness and the buckling criterion, the maximum possible length of the latch is deduced therefrom;
• On the basis of this length, the thickness and the residual stress criterion, the maximum clamping force of the connection is calculated, from which the maximum deformation is deduced;
• This deformation is compared with the maximum possible tolerance due to the stacking of the constituents of the panels, making sure that the maximum difference between the deformation and the maximum tolerance is less than the maximum admissible connection clearances;
• It is ensured that with the forces thus defined, it will be possible to catch up the clearances due to the form tolerances for the real connection to be made.

On the assumption that the stacking or shape tolerances cannot be made up respectively, the analysis is repeated by reducing these tolerances respectively.

In practice, starting the analysis with standard tolerances which make it possible to have a minimum cost of the panels, it may be necessary to reduce these tolerances and therefore to increase the cost of the panels, but this increase will be as low as possible.

avec :

• e = épaisseur du linguet (mm);
• a_matage = contrainte de matage admissible sur (MPa);
• D = diamètre du rivet (mm);
• F_cisaillement = effort externe de cisaillement appliqué sur un rivet (N).

It is first necessary that the catch tab modeled at the figure 4 , also called hereafter pawl, and the link are able to withstand the forces of the structure in operation. These supposedly known forces make it possible to determine the number of connecting elements / rivets, their diameter D, the pitch P between two rivets as a function in particular of their characteristics. The operation by matting of these rivets makes it possible to determine the thickness of the pawl: $$\mathrm{e}\ge \mathrm{F}\_\mathrm{shear}/\left(\mathrm{D}*\mathrm{\sigma }\_\mathrm{matting}\right)$$

with:

• e = thickness of the latch (mm);
• a_matage = permissible matting stress on (MPa);
• D = rivet diameter (mm);
• F_cisaillement = external shear force applied to a rivet (N).

Furthermore, the link must be able to withstand all the forces to which it will be subjected in operation. The most dimensioning force is the buckling resistance of the latch during a bending force of the connection to do this, the buckling stress σ _f must be greater than the elastic limit σ _e of the material used, in fact with a safety factor of 1.5.

The buckling stress can be calculated in various ways, it is estimated here by assimilating the latch to an intermediate between an embedded - embedded beam and a embedded - guided beam: $${\mathrm{F}}_{\_\mathrm{critical}\mathrm{buckling}}={\mathrm{<figure-callout\; id="2"\; label="\pi "\; filenames="imgf0001.png"\; state="\{\{state\}\}">\pi </figure-callout>}}^2*\mathrm{E}*{\mathrm{I}}_{\mathrm{f}}/{{\mathrm{<figure-callout\; id="2"\; label="I\; k"\; filenames="imgf0001.png"\; state="\{\{state\}\}">I</figure-callout>}}_{\mathrm{k}}}^2$$

• E = module élastique du matériau du linguet (MPa);
• I_f = Inertie de flexion du linguet = b* e³ /12;
• I_k = 0,6;

With:

• E = elastic modulus of the latch material (MPa);
• I _f = Inertia bending the pawl = b * e ^3/12;
• I _k = 0.6;

avec :

• L: longueur du linguet jusqu'au point d'introduction du serrage (mm);
• e = épaisseur du linguet (mm);
• σ_e = limite élastique du matériau utilisé (MPa);

From where : $${\mathrm{\sigma }}_{\_\mathrm{critical}\mathrm{buckling}}=\left({\mathrm{<figure-callout\; id="2"\; label="\pi "\; filenames="imgf0001.png"\; state="\{\{state\}\}">\pi </figure-callout>}}^2*\mathrm{E}*{\mathrm{<figure-callout\; id="2"\; label="e"\; filenames="imgf0001.png"\; state="\{\{state\}\}">e</figure-callout>}}^2\right)/\left(4.32*{\mathrm{<figure-callout\; id="2"\; label="L"\; filenames="imgf0001.png"\; state="\{\{state\}\}">L</figure-callout>}}^2\right)>1.5{\mathrm{\sigma }}_{\mathrm{e}}$$

with:

• L: length of the latch up to the point of entry of the clamping (mm);
• e = thickness of the latch (mm);
• σ _e = elastic limit of the material used (MPa);

From where : $$\mathrm{L}<\mathrm{\pi }*\mathrm{e}*{\mathrm{E}}^{1/2}*1/{\left(4.32*1.5*{\mathrm{\sigma }}_{\mathrm{e}}\right)}^{1/2}$$

In order for the pawl 24 to remain usable in its function, it is also necessary that the bending stress induced by the clamping does not reach a fraction of the elastic limit of the material; the maximum _{admissible value σ m} for this residual stress depends on the assembly and its function, but as a first estimate we can choose that this maximum stress does not exceed half of the stress potential at elastic limit of the material, i.e. σ _m = 1 / (2 * 1.15) * σ _e . This constraint makes it possible to calculate the maximum deflection attainable by the pawl, and therefore the maximum play which can be taken up.

If y _m is this deflection, it is related to the maximum acceptable stress in the latch by the relation (approximation of the latch by a fixed beam): $${\mathrm{y}}_{\mathrm{m}}={\mathrm{\sigma }}_{\mathrm{m}}*\left(2*{\mathrm{<figure-callout\; id="2"\; label="L"\; filenames="imgf0001.png"\; state="\{\{state\}\}">L</figure-callout>}}^2\right)/\left(3*\mathrm{E}*\mathrm{e}\right)$$

• Soit y_m > D_e - m, où m est le jeux résiduel qui peut être comblé par un mastic ; m vaut au maximum 0,7 mm de façon classique. Alors, l'assemblage est réalisable.
• Soit y_m < D_e - m alors l'assemblage n'est pas réalisable ; il faut revoir la chaine de côte qui doit être plus contraignante ; et les pièces seront plus chère à fabriquer.

The comparison of this maximum attainable deflection with the stacking defect D _e makes it possible to check whether assembly is possible:

• Let y _m > D _e - m, where m is the residual clearance which can be filled by a putty; m is a maximum of 0.7 mm in a conventional manner. Then, the assembly is feasible.
• Let y _m <D _e - m then the assembly is not feasible; it is necessary to review the chain of coast which must be more restrictive; and the parts will be more expensive to manufacture.

It remains to verify that the effort to feed maximum landings F _applied to latch end will be sufficient for this distortion. To carry out the assembly operation, pins are usually used, which are temporary docking devices that are placed in the holes provided for the rivets of the final assembly; knowing the maximum docking force of a pin (T _pin ) and considering a pin in each hole, we calculate K _a = T _pin / P and that this value is greater than the force K necessary to bend the latch of the maximum fault value D _e : $$\mathrm{K}=3*{\mathrm{D}}_{\mathrm{e}}*\mathrm{E}*{\mathrm{e}}^3/\left({\mathrm{L}}^3*12\right)$$

At the end of this first phase of predefining the latch, the analysis continues with a calculation which makes it possible to verify the assembly requirements by taking into account the shape defects of the panels to be assembled in their long lengths.

This requires a finite element calculation, in which all of the two parts to be assembled are modeled, including the catch or catch tab defined previously, with the following assumptions;

The deformation is ensured by assembly pins using the places provided by the rivets; assembly pins are standard products, the clamping force of which is known.

The shape defect is modeled either by tools (such as 3DCS assembly simulation software from the HITEX company), or with finite element computation software in the form of a sinusoidal defect of variable wavelength, including the amplitude is the expected shape defect.

• le jeux résiduels qui doivent être au maximum mentionnés ci-dessus (m)
• les contraintes résiduelles qui ne doivent pas dépasser σ_m

The calculation ultimately gives:

• the residual clearances which must be at most mentioned above (m)
• the residual stresses which must not exceed σ _m

If these objectives are not achieved, it is necessary to specify smaller form defects, which complicates the manufacturing processes of the parts and therefore their cost.

It will be noted that the approach mentioned above relates to the worst case, which indicates the feasibility of the assembly, and the class of manufacturing tolerances necessary for the success of the assembly. When this feasibility is demonstrated, it is always possible to optimize the design in terms of latch length, choice of material, residual stresses, residual clearances and cost.

Increasing the length of the latch therefore makes it possible to greatly reduce the maximum internal stress on the frame or metal structure. However, indefinitely increasing the length of the latch is not possible. There is a risk of buckling of the latch.

Also, a compromise between the length of the latch and therefore of the termination of the composite panel and of its part returning to the skin and its flexibility is possible depending on the parts to be connected, their dimensions and the forces which they have to pass.

The figures 5A and 5B represent the location of the maximum stresses σ, σ _M1 in the case of a negative clearance - y for the figure 5A where the maximum stress σ _M1 is located at the junction between the catch or catch tab and the connecting wall 22 and σ _M2 in the case of a positive clearance + y in the case of the figure 5B where the maximum stress σ _M2 is located in the docking zone of the latch on the first face 14 of the panel.

To better describe the implementation of the invention, we will consider the case presented in figure 6 . This figure shows the aluminum structure 200 of the figure 3 in perspective, this structure to be assembled with a panel of composite sandwich material. The length of the two pieces and the connection is 10 m and the rivets are aligned along two lines 110 and 120 on both sides of the structure.

The figure 7A represents a form defect to be corrected D _f between the composite panel and the metal structure The figure 7B represents in the left part the deviation Δy and in the right part the result after tightening by means of suitable tools (pins for example) or rivets 30.

By way of example, we will consider a panel such that the thickness of the foam core of the sandwich panel is 38mm, the thickness of the skins is 3 mm, The return to the skin of the composite panel is carried out with a 45 ° angle, and a play of 10 mm is provided between the catch or catching tab 24 and the first face 14 of the panel when the edge 17 is placed on the support wall 21 has been chosen in order to avoid double contact.

The structure of the example is in 6005 T6 aluminum 3 to 5 mm thick, its modulus of elasticity is E = 70GPa, and its elastic limit is σ _e> 220MPa

In addition, the stresses in use made it possible to define the connection which will be made using passivated stainless steel blind rivets with a diameter of 6.58mm, distributed at a pitch of P = 60mm, to ensure the holding under a shear stress flow of 70 N / mm

The pins used to carry out the assembly allow to pass a force of 2000N.

Concerning the stacking defect (chain of dimensions), it is considered equal to ± 0.74mm (quadratic approach adopted).

• Epaisseur > 2,1 mm soit 3mm retenu;
• Longueur maximale L < 69 mm avec cette épaisseur de 3mm;

We therefore obtain the following result for the latch:

• Thickness> 2.1 mm or 3mm retained;
• Maximum length L <69mm with this thickness of 3mm;

Maximum deflection admissible by the latch in bending = 1.5mm, therefore greater than the fault to be remedied

Docking force flux required to deform the latch of the default value = 1.1 N / mm, therefore less than the maximum docking force flux provided by the pins of 33N / mm

It can therefore be seen that the criteria which have been set to define the latch are therefore verified. The analysis will continue by taking into account the defects in the shape of the parts to be assembled, neglected at this stage.

• Pour la structure métallique: une sinusoïde d'amplitude 15mm crête à crête, et une longueur d'onde de 2,5m, ce défaut correspondant à une tolérance de forme globale de 15mm associé à une tolérance de forme locale de 0,6mm sur 300mm, conforme aux défauts susceptible d'être rencontré sur un profilé en aluminium filé;
• Pour le panneau composite: une sinusoïde d'amplitude 15 mm crête à crête, et une longueur d'onde du défaut identique à celui du panneau métallique, soit 2,5 m, mais en opposition de phase par rapport à ce dernier.

The shape defects of the parts in the great length of the 10m panel of the example, as an initialization at the start of the analysis, will be represented by:

• For the metallic structure: a sinusoid of amplitude 15mm peak to peak, and a wavelength of 2.5m, this defect corresponding to an overall shape tolerance of 15mm associated with a local shape tolerance of 0.6mm over 300mm , conforms to the defects likely to be encountered on a extruded aluminum profile;
• For the composite panel: a sinusoid of amplitude 15 mm peak to peak, and a wavelength of the fault identical to that of the metal panel, ie 2.5 m, but in phase opposition with respect to the latter.

The boundary conditions making it possible to carry out the assembly simulations and the conditions of validity of the connection are specified below and by way of example, the docking forces are set at 2000N every 240mm (1 pin every 5 holes), i.e. a pinning force flux of 8333N / mm.

• En B : m < 0,7mm;
• En A : m < 0,1 mm (plus faible qu'en A car on ne souhaite pas faire de comblage liquide dans cette zone);

In terms of the maximum allowable clearance between parts after docking:

• In B: m <0.7mm;
• In A: m <0.1 mm (lower than in A because we do not want to make liquid filling in this zone);

• Pièce métallique : critère de Von Mises σ_m < 220 * (1 / (2 * 1,15) = 95MPa;
• Pièce composite : critère de Hashin < 0,15 = (1 / (2 * 1,3))²

In terms of maximum admissible stresses in the parts after docking:

• Metal part: Von Mises criterion σ _m <220 * (1 / (2 * 1.15) = 95MPa;
• Composite part: Hashin criterion <0.15 = (1 / (2 * 1.3)) ²

The coefficient 1.3 corresponds to the breaking safety coefficient specified in the dimensioning standard for boxes NF EN 12663-1.

Since the flexibility of the parts depends on the curvatures of the panels, two cases of assemblies corresponding to two geometries of composite panels were studied: the case of a flat composite panel (such as one face), and the case of a composite panel curved (such as a roof panel).

The figure 11 shows the view of a curved panel case 300 'assembled with the body structure 200 or other panel.

In the analyzes which follow, the length of the catch or catching tab 24 'is geometrically fixed at 62mm.

• une sinusoïde d'amplitude 4,5mm crête à crête et une longueur d'onde de 2,5m, ce défaut correspondant à une tolérance de forme globale de 4,5mm associé à une tolérance de forme locale de 0,2mm sur 300mm;
• une sinusoïde d'amplitude 7,5mm crête à crête et une longueur d'onde de 5m, ce défaut correspondant à une tolérance de forme globale de 7,5mm associé à une tolérance de forme locale de 0,1mm sur 300mm.

The results of the two finite element analyzes are presented below:
For the connection between a flat composite panel as in figure 10 and a metal structure or panel, the bond validity requirements are met for the following tolerances on the composite panel and the metal panel:

• a sinusoid of 4.5mm peak-to-peak amplitude and a wavelength of 2.5m, this defect corresponding to an overall shape tolerance of 4.5mm associated with a local shape tolerance of 0.2mm over 300mm;
• a sinusoid of 7.5mm peak-to-peak amplitude and a wavelength of 5m, this defect corresponding to an overall shape tolerance of 7.5mm associated with a local shape tolerance of 0.1mm over 300mm.

For the curved composite panel of the figure 11 , the validity requirements of the connection are ensured for more restrictive tolerances on the panels to be assembled, since the stiffness of the composite panel limits its ability to deform. The maximum justified fault corresponds to a sinusoid of amplitude 4.5mm peak to peak and a wavelength of 5m. This defect corresponds to an overall shape tolerance of 4.5mm associated with a local shape tolerance of 0.1mm over 300mm.

It can thus be seen that, as expected, the assembly in the case of the curved composite panel requires that its shape tolerance be almost twice lower than that required in the case of the assembly of the flat panel; and therefore that its cost is higher

In addition, to obtain compliance with the docking criteria under these conditions, the length of the latch had to be extended to its maximum value, ie 69mm in order to maximize the flexibility of the latter.

The figures 8A to 8C show a blind rivet that can be used to make the assembly of the invention.

In figure 8A the rivet 30 is shown before it is crushed in order to clamp an external part A on an internal part B, for example the bearing wall and the catch or catching tab and the first skin of the panel.

The rivet according to the example comprises a rod 31 terminated by a bulge 36, a head 32, here a countersunk head, a hollow body 33 receiving the rod and comprising a conical part 33a, a helical washer 34 and a deformable sleeve 35.

The helical washer is disposed between the sleeve 35 and the conical part 33a of the hollow body 33 and the deformable sleeve 35 is retained by the bulge 36 of the rod 31 against the washer.

In figure 8B , a pull on the rod 31 pulls on the bulge 36 which pushes the washer 34 and the sleeve 35 on the body 33. The washer is expanded and the sleeve collapses longitudinally while widening while rising on the conical part of the body .

In figure 8C , the washer 34 is in abutment against the internal part B and is held by means of the crushing of the sleeve 35. Once the calibrated clamping force corresponding to the rupture of the rod 31 is reached, the latter breaks to leave an external surface without roughness.

For the rail car of the figure 9 comprising composite panels fixed with a metal lower structure, the side faces 300 of the car are riveted in accordance with the invention on the body 200, a structure located in the lower part of the car.

The invention is not limited to the examples shown and in particular a tubular reinforcing element or according to another profile such as an omega can be provided in the panel under the first skin 10 to the right of the drilling zones P1 receiving the rivets 30 for strengthen the panel. Likewise, the metal structure can be of varied height H, be straight or have a radius of curvature different from that shown.

Claims (18)

1. An end link between a composite panel (1, 300, 300') including a first skin (10) and a second skin (11), defining outer faces of said composite panel, connected by a solid or honeycomb core (12) extending between the first skin (10) and second skin (11), and a second panel (2, 200) provided with two outer faces defined by two skins separated by a solid or honeycomb core or a box building, for which one of the skins of each of said panels is continued with a tab (17, 24), the tab of a first of said panels being applied, upon linking the panels, to the external face of the skin (10, 26) devoid of tab of a second of said panels, said second panel (2, 200) comprising a straight edge (22) substantially perpendicular to said faces of the second panel, the tab of the second panel, referred to as take-up tab (24), including in its width a bearing terminal part (b, c), characterised in that the core (12) of the composite panel includes a termination edge (13), made bevelled relative to said faces of the composite panel (10, 11), the bevelled edge starting from the first skin to extend to the second skin, in that the first skin snugly fits the slope of the bevelled edge to join the second skin and make the tab (17) of the composite panel, referred to as bearing tab, so that the bearing tab bears substantially over its entire width against the bearing part (21) of the face of the second panel receiving the same and defines a substantially rigid reference surface for linking, the tab of the straight-edge panel including a free intermediate part (a) likely to bend to take-up thickness differences between the panels.
2. The end link according to claim 1, for which said straight edge (22) forms an angle of 90° ± 15° relative to said faces of the skins (23, 26) of the second panel.
3. The end link according to claim 1 or 2, for which said bevelled edge (13) forms an angle of 45° ± 15° relative to the normal to said faces (10, 11) of the composite panel.
4. The end link according to any of the previous claims, for which said bevelled edge (13) is made by folding the first skin (10) of the composite panel at a first tilt at the beginning of the bevel and then at a second tilt at the end of the bevel where the first skin (10) joins the second skin (11) of the panel to form the first tab referred to as bearing tab (17) at the end of said bevelled edge.
5. The end link according to any of the previous claims, for which said straight edge (22) is a wall connecting both skins (23, 26) of the second panel and disposed at the end of a first one of these skins (26), said take-up tab (24) of the second panel (22) consisting of a continuation of a second one of its skins (23) cantilevered relative to said straight edge (22).
6. The end link according to any of the previous claims, for which the second panel (2, 200) including said straight edge is a metal panel or structure.
7. The end link according to any of the previous claims, for which said take-up tab (24) is made so as it has a flexibility adapted to take up thickness differences of panels by bending its free intermediate part (a).
8. The end link according to any of the previous claims, for which, by making a take-up tab (24) and a bevel (13) long enough, the free intermediate part (a) is deformable enough to contact with the face of the panel receiving the same and enable riveting while minimising stresses induced by the assembly.
9. The end link according to any of the previous claims, for which the link includes an openwork with a substantially rectangle triangular cross-section, said bevelled edge (13) making up the hypotenuse of said cross-section, the straight edge (22) and said free intermediate part (a) making up both other sides of said cross-section.
10. An assembly of two panels by means of a link according to any of the previous claims which comprises between both panels a riveted link comprising at least one first row of rivets (P2, 120) for fastening the bearing tab (17) of the composite panel to a bearing face (26, 21) of the second panel and at least one second row of rivets (P1, 110) for fastening said terminal part of said take-up tab (24) to a bearing face (10) of the composite panel.
11. The assembly according to claim 10, for which the bearing wall (21) of the second panel is advantageously a recessed border of the face (26) devoid of tab of said second panel so as to compensate for the thickness of the bearing tab to have the planar faces (26, 11) external to both panels in continuity.
12. The assembly according to claim 10 or 11, for which the bearing wall of the panel including said straight edge is advantageously a recessed border (21) of the face devoid of tab of said panel and connected to said face by a land (25) forming a stop for the bearing tab (17).
13. The assembly according to claim 10, 11 or 12, for which the take-up tab (24) is preferably adapted to curve under the action of the assembly of the first panel and second panel either in the direction of the first panel (-y) or in the opposite direction (+y).
14. The assembly according to claim 13, for which the flexibility of said take-up tab is calculated to enable the same to bend by a distance at least equal to the sum of tolerances of the layers that make up each panel.
15. The assembly according to claim 13, for which the flexibility of said take-up tab is calculated to enable the same to bend by a distance at least equal to the quadratic sum of tolerances of layers that make up each panel.
16. The assembly according to claim 15, for which for the quadratic sum, the square root of the sum of the squares of uncertainties of each layer is taken.
17. The assembly according to claim 14, 15 or 16, for which bending of said take-up tab is measured at the landing point between said tab (24) and the face (10) of the panel to which it is applied.
18. The assembly according to any of claims 10 to 17, for which the second of said panels is a metal structure.

Images