EP2357271A1 - Method for controlling electrical actuators of a shedding device - Google Patents

Abstract

The method involves determining remarkable points of a smooth movement profile driven by an actuator or group of actuators according to shed parameters, and determining approximation parameterized function of the profile. Six accelerations are calculated, and a law of smooth movement is determined for a data pick such that the smooth accelerations at beginning and ending of its movement are equal to the fifth and sixth accelerations. A set of setpoint data is generated for the data pick and the actuator or group of actuators, from the determined law of movement.

Description

The invention relates to a method for controlling the electric actuators of a shedding device on a loom.

In the field of the formation of the shed on a loom, it is known to use an electric actuator to move the healds and thus form the shed in which the weft son are inserted during successive picks. The rails can be mounted on frames, in which case the actuator can be of the type described in EP-A-1,489,208 . The rails can also be connected each to an arch which winds on a pulley secured to the rotor of a rotary actuator as known from EP-A-0 926 279 .

These electric actuators must be controlled, for each pick and depending on the or slats they lead, with position instructions that allow them to ensure the smooth movement of a number of parameters, called "parameters of crowd "and which include the armor, the range of motion, an aperture profile and any offset from the business cycle or a median plane of the crowd.

It is known to EP-A-0 774 538 calculating positional readings for a heald frame actuator from generic curves, defined on a pick, from two open shed positions. These generic curves are stored in an accessible library for a control device. This device sends an amplifier associated with each actuator information relating to the type of generic curve to be followed for calculating the set values for each pick. Such a method does not adapt the profile followed by each actuator to optimize the movement of the rails. The number of generic curves is necessarily limited and each of these is defined on a pick, being obtained by the concatenation of a portion of movement and one or two stop portions allowing the connection between two generic curves. The time available for the movement is then necessarily less than the duration of a pick, which increases the acceleration to which the moving parts of the actuators, the rails, the transmission rods and the frames are subjected.

An alternative to this known solution is to calculate the position instructions of the various actuators for the entire armor, before starting the weave, then transmit these instructions to the amplifier associated with each actuator, outside the weaving phases. Relatively complex calculations can be envisaged but this does not allow to take into account changes in crowd parameters during weaving. In addition, high capacity calculation means must be used.

It is these drawbacks that the invention intends to remedy more particularly by proposing a new control method that makes it possible to generate a set of setpoint data, for each pick and each actuator or group of actuators, in an optimized manner. .

1. a) déterminer, en fonction des paramètres de foule, des points remarquables d'un profil de mouvement d'une lisse entraînée par l'actionneur ou le groupe d'actionneurs, au moins pour la duite donnée, la duite précédente et la duite suivante ;
2. b) déterminer une fonction paramétrée d'approximation du profil de mouvement qui passe par certains points remarquables, au moins pour la duite donnée, la duite précédente et la duite suivante ;
3. c) calculer au moins quatre accélérations, à savoir :
   • une première accélération de la lisse au début de son mouvement, à partir de la fonction d'approximation déterminée à l'étape b) pour la duite donnée,
   • une deuxième accélération de la lisse à la fin de son mouvement, à partir de la fonction d'approximation déterminée à l'étape b) pour la duite donnée,
   • une troisième accélération de la lisse à la fin de son mouvement, à partir de la fonction d'approximation déterminée à l'étape b) pour la duite précédente,
   • une quatrième accélération de la lisse au début de son mouvement, à partir de la fonction d'approximation déterminée à l'étape b) pour la duite suivante,
4. d) calculer une cinquième accélération, en fonction des première et troisième accélérations, et une sixième accélération, en fonction des deuxième et quatrième accélérations ;
5. e) déterminer une loi de mouvement de la lisse pour la duite donnée, dont le profil passe par certains points remarquables et telle que l'accélération de la lisse au début de son mouvement est égale à la cinquième accélération et que l'accélération de la lisse à la fin de son mouvement est égale à la sixième accélération ;
6. f) générer l'ensemble de données de consigne, pour la duite donnée et pour l'actionneur ou le groupe d'actionneurs, à partir de la loi de mouvement déterminée à l'étape e).

To this end, the invention relates to a method for controlling the electric actuators of a shedding device on a loom in which a set of setpoint data is generated, for each pick and each actuator or group of actuators, taking into account predetermined mating parameters, characterized in that it comprises, for a given pick and a given actuator or group of actuators, steps constituting:

1. a) determining, according to the shedding parameters, the remarkable points of a movement profile of a rail driven by the actuator or the group of actuators, at least for the given pick, the previous pick and the next pick ;
2. b) determining a parameterized function of approximation of the motion profile passing through certain remarkable points, at least for the given pick, the previous pick and the next pick;
3. c) calculate at least four accelerations, namely:
   • a first acceleration of the bar at the beginning of its movement, from the approximation function determined in step b) for the given pick,
   • a second acceleration of the smooth at the end of its movement, from the approximation function determined in step b) for the given pick,
   • a third acceleration of the smooth at the end of its movement, from the approximation function determined in step b) for the previous pick,
   • a fourth acceleration of the beam at the beginning of its movement, from the approximation function determined in step b) for the next pick,
4. d) calculating a fifth acceleration, based on the first and third acceleration, and a sixth acceleration, based on the second and fourth acceleration;
5. e) determining a law of movement of the rail for the given pick, whose profile passes through certain remarkable points and such that the acceleration of the rail at the beginning of its movement is equal to the fifth acceleration and that the acceleration of the smooth at the end of its movement is equal to the sixth acceleration;
6. f) generating the set of setpoint data, for the given pick and for the actuator or the group of actuators, from the motion law determined in step e).

Thanks to the invention, taking into account the accelerations at the connection between the given pick, the previous pick and the next pick, makes it possible to determine a motion law compatible with a direct connection with the laws of movement provided for the previous pick and the next pick. Since the parameterized approximation function and the calculation operations are performed on the basis of the remarkable points corresponding to the crowd parameters, these computations can take place dynamically and take account of changes made during weaving to the parameters of the parameters. crowd.

• La cinquième accélération est égale à la moyenne des première et troisième accélérations et la sixième accélération est égale à la moyenne des deuxième et quatrième accélérations.
• Il est prévu des étapes antérieures à l'étape e) et consistant, g), à déterminer le type de séquence de mouvement de la lisse sur la duite précédente, la duite donnée et la duite suivante, au sein d'un groupe de huit mouvements type, h), si le type de séquence de mouvement déterminé à l'étape g) est tel que la lisse est déplacée pour la duite donnée, entre deux positions situées respectivement de part et d'autre d'un plan médian de la foule, et que la lisse n'est pas déplacée entre ces deux portions pour la duite précédente et/ou la duite suivante à déterminer au moins un paramètre d'étalement du mouvement de la lisse en tenant compte de certains au moins des points remarquables et, i), lorsque le paramètre d'étalement a été déterminé, déterminer au moins un nouveau point remarquable, en fonction des paramètres de foule et du paramètre d'étalement, alors que, lors de l'étape e), la loi de mouvement est déterminée en tenant compte du nouveau point remarquable et du paramètre d'étalement.

• Lors de l'étape h), le paramètre d'étalement est déterminé par tests successifs de l'influence de ce paramètre sur la compatibilité d'une loi de mouvement avec au moins un point remarquable.
• La compatibilité du paramètre d'étalement avec la loi de mouvement est testée avec un ou plusieurs points remarquables représentatifs d'une géométrie d'ouverture de foule, pour la duite et l'actionneur ou le groupe d'actionneurs concernés.
• Le paramètre d'étalement est représentatif de la portion supérieure à 360° de l'amplitude angulaire, par rapport à la rotation de l'arbre métier, sur laquelle a lieu le déplacement de la lisse pour la duite donnée.
• Des données représentatives des paramètres de foule sont stockées dans une mémoire dynamique à laquelle il est accédé lors de l'étape a), alors que ces données sont modifiables en cours de tissage.
• Lors de l'étape b), on détermine également une fonction paramétrée d'approximation du profil de mouvement qui passe par certains points remarquables pour une duite précédant de deux coups la duite donnée et une duite suivant de deux coups la duite donnée, alors que, lors de l'étape c), on calcule également :
• une septième accélération de la lisse au début de son mouvement à partir de la fonction d'approximation déterminée à l'étape b) pour la duite précédente,
• une huitième accélération de la lisse à la fin de son mouvement à partir de la fonction d'approximation déterminée à l'étape b) pour la duite suivante,
• une neuvième accélération de la lisse à la fin de son mouvement à partir de la fonction d'approximation déterminée à l'étape b) pour la duite précédant la duite précédente,
• une dixième accélération de la lisse au début de son mouvement à partir de la fonction d'approximation déterminée à l'étape b) pour la duite suivant la duite suivante,
  alors qu'on calcule une onzième accélération, en fonction des septième et neuvième accélérations et une douzième accélération en fonction des huitième et dixième accélérations, et alors que, lors de l'étape e), on détermine la loi de mouvement pour la lisse donnée de telle sorte que l'accélération de la lisse pour la duite précédente en début de son mouvement est égale à la onzième accélération et que l'accélération de la lisse pour la duite suivante à la fin de son mouvement est égale à la douzième accélération. La onzième d'accélération est de préférence égale à la moyenne des septième et neuvième accélérations, alors que la douzième accélération est de préférence égale à la moyenne des huitième et dixième accélérations. Le paramètre d'étalement et/ou le nouveau point remarquable sont avantageusement déterminés en tenant compte des onzième et douzième accélérations.
• Il est prévu des étapes j) et k) antérieures à l'étape c) et consistant, j), à vérifier la compatibilité de la fonction d'approximation avec des points remarquables représentatifs d'une géométrie d'ouverture de foule pour au moins la duite donnée, la duite précédente et la duite suivante, et, k), en cas d'incompatibilité détectée à l'étape j), à imposer ces points remarquables comme points de passage de la fonction d'approximation.
• La fonction paramétrée et/ou la loi de mouvement s'exprime sous la forme $${\displaystyle \sum _{i=0}^m}{a}_i\mathrm{.}\cos \left(i\mathrm{.}\mathit{\omega \theta }\right)$$
  

avec θ égal à l'angle du métier dans son cycle, m entier supérieur ou égal à 1 et a_i constantes, m étant de préférence inférieur ou égal à 6.According to advantageous but non-mandatory aspects of the invention, such a method may incorporate one or more of the following features, taken in any technically permissible combination:

• The fifth acceleration is equal to the average of the first and third acceleration and the sixth acceleration is equal to the average of the second and fourth acceleration.
• There are steps prior to step e) and consisting, g), to determine the type of motion sequence of the smooth on the previous pick, the given pick and the next pick, within a group of eight type movements, h), if the type of motion sequence determined in step g) is such that the beam is displaced for the given pick, between two positions located respectively on either side of a median plane of the crowd, and that the smooth is not moved between these two portions for the previous pick and / or the next pick to determine at least one spreading parameter of the movement of the rail taking into account at least some of the remarkable points and i), when the spreading parameter has been determined, determining at least one new remarkable point, depending on the crowd parameters and the spreading parameter, whereas, during step e), the motion law is determined taking into account the new remarkable point and the spreading parameter.

• During step h), the spreading parameter is determined by successive tests of the influence of this parameter on the compatibility of a motion law with at least one remarkable point.
• The compatibility of the spreading parameter with the motion law is tested with one or more remarkable points representative of a shedding geometry, for the pick and the actuator or group of actuators concerned.
• The spreading parameter is representative of the portion greater than 360 ° of the angular amplitude, relative to the rotation of the business tree, on which the displacement of the bar for the given pick takes place.
• Data representative of the crowd parameters are stored in a dynamic memory to which it is accessed during step a), while these data are modifiable during weaving.
• During step b), a parameterized function of approximation of the motion profile which passes through certain remarkable points for a pick preceding two shots at a given pick and a pick of two shots at a given pick is also determined, whereas during step c), the following are also calculated:
• a seventh acceleration of the smooth at the beginning of its movement from the approximation function determined in step b) for the previous pick,
• an eighth acceleration of the smooth at the end of its movement from the approximation function determined in step b) for the next pick,
• a ninth acceleration of the smooth at the end of its movement from the approximation function determined in step b) for the pick preceding the previous pick,
• a tenth acceleration of the stringer at the beginning of its movement from the approximation function determined in step b) for the pick following the next pick,
  while calculating an eleventh acceleration, as a function of the seventh and ninth accelerations and a twelfth acceleration as a function of the eighth and tenth accelerations, and whereas, during step e), the motion law for the given smooth is determined so that the acceleration of the smooth for the previous pick at the beginning of its movement is equal to the eleventh acceleration and the acceleration of the smooth for the next pick at the end of its movement is equal to the twelfth acceleration. The eleventh acceleration is preferably equal to the seventh and ninth average accelerations, while the twelfth acceleration is preferably equal to the average of the eighth and tenth acceleration. The spread parameter and / or the new remarkable point are advantageously determined taking into account the eleventh and twelfth accelerations.
• Steps j) and k) prior to step c) and consisting of j) are provided to check the compatibility of the approximation function with remarkable points representative of a crowd opening geometry for at least the given duit, the previous pick and the next pick, and, k), in case of incompatibility detected in step j), to impose these remarkable points as passing points of the approximation function.
• The parameterized function and / or the motion law is expressed in the form $${\displaystyle \underset{i=0}{\overset{m}{\Sigma }}}{\mathrm{at}}_i\mathrm{.}\cos \left(i\mathrm{.}\mathit{\omega \theta }\right)$$
  

with θ equal to the angle of the loom in its cycle, m integer greater than or equal to 1 and _i constant, m is preferably less than or equal to 6.

• la figure 1 est une représentation schématique de principe d'un métier à tisser avec lequel peut être mise en oeuvre l'invention ;
• la figure 2 est une représentation schématique d'un profil de mouvement d'une lisse au cours d'une duite, sur le métier de la figure 1 ;
• la figure 3 est une représentation schématique analogue à la figure 2, pour un autre profil de mouvement ;
• la figure 4 est un exemple de représentation des variations de positon d'une lisse au cours du temps sur plusieurs duites ;
• la figure 5 est une représentation schématique de principe, comparable à la figure 4 montrant un mouvement de lisses sur trois duites successives ;
• la figure 6 est une vue analogue à la figure 5 pour un autre mouvement de lisses ; et
• la figure 7 est un ordinogramme d'un procédé conforme à l'invention.

The invention will be better understood and other advantages thereof will appear more clearly in the light of the following description of an embodiment of a control method according to its principle, given solely as a example and with reference to the accompanying drawings in which:

• the figure 1 is a schematic representation of a weaving machine with which the invention can be implemented;
• the figure 2 is a schematic representation of a movement profile of a smooth during a pick, on the craft of the figure 1 ;
• the figure 3 is a schematic representation similar to the figure 2 , for another movement profile;
• the figure 4 is an example of representation of positron variations of a smooth over time over several picks;
• the figure 5 is a schematic representation of principle, comparable to the figure 4 showing a movement of smooth on three successive picks;
• the figure 6 is a view similar to the figure 5 for another movement of smooth; and
• the figure 7 is a flow chart of a method according to the invention.

The loom 2 shown schematically and partially to the figure 1 is equipped with stringer frames, only one of which is visible in this figure, with the reference 4. Each stringer frame is set in motion by an electric actuator 6 of the type electric motor, without brush ("brushless"). Each actuator 6 is connected to a smooth frame 4 by means of a mechanical transmission comprising unrepresented gears and connecting rods including an instrumented pulling rod 8 equipped with a stress measuring cell 10. Each cell 10 is connected by a wire link 12 to a measurement conditioner 14.

Each actuator 6 is controlled by an associated amplifier 16 which manages its current supply, as a function of the position of the rotor, not shown, of this actuator 6 measured by an encoder 18 and as a function of a position command which comes to it in the form of an electronic signal S ₁ from a dobby controller 20 which is common for all the actuators 6.

The dobby controller 20 is connected to the controller 22 of the loom 2 by means of a wire link 24 which allows the exchange of information relating to the operating mode of the loom 2 and the shedding device comprising, inter alia, the actuators 6 and the frames 4.

The dobby controller 20 comprises a computer 26 which is capable of transmitting the signals S ₁ and which is connected by the link 24 to the trade controller 22. The computer 26 is also connected to a sensor 28 which provides it with the position of the trade 2 in his cycle. The sensor 28 is, for example, a resolver coupled to the main shaft of the loom 2. The dobby controller 20 is equipped with a memory 30 in which are stored, in the form of a set of data D ₁ , the crowd parameters to be used, these parameters including in particular the armor, the magnitudes of displacement of the frames, the profiles to be followed and the possible time offsets with respect to a median position.

The controller 20 also comprises a user interface 32 which includes display means 32 ₁ and input means 32 ₂ making it possible to modify, among other things, the crowd parameters stored in the memory 30. The elements 30 and 32 are connected to the computer 26 by wired links 34 and 36.

N-order picking is considered within an armor, which can include a large number of picks. This number of picks is called the armor ratio. The duit preceding this given duce of order N is the duity of order N-1 and the duity preceding the given duity N of two moves is the duity of order N-2. The pick following the given pick of order N is the pick of order N + 1, while the next pick of two moves the given pick of order N is the pick of order N + 2. For each given pick N, the position setpoint for the control of an actuator 6 is transmitted in the form of a vector V _N of thirty-three points which defines a succession, of thirty-two regular intervals, of angular positions of the rotor of this actuator on a business cycle. A vector V _N is transmitted to each amplifier 16 as part of the signal S ₁ transmitted by the computer 26 to this amplifier. Each amplifier 16 is capable of interpolating the intermediate positions of the rotor of the actuator that it supplies, between two positions given by the vector V _N transmitted to it by the computer 26.

A business cycle usually corresponds to a 360 ° rotation of the main tree of trade 2.

The vector V _N transmitted to each amplifier 16 for a pick N thus comprises a set of setpoint data which it is important to determine in the best way to obtain a harmonious weave, with optimized accelerations. The determination by calculation of each vector or set of reference data V _N is carried out by the computer 26.

The N pick extends over an angular range of about 360 ° of rotation of the business tree. Note θ ₁ the position of the business tree at the beginning of the pick N and θ ₂ its position at the end of the pick. In practice, the value of θ ₁ is equal to the value of θ ₂ , approximately 360 °. The figure 2 represents the height y of a part of a rail, which defines the position of a warp passing through this rail, as a function of the working angle θ. In the example shown in solid line at the figure 2 this height varies during the N duit to the extent that the rail passes from an initial position, in which its eyelet is located above a median plane P _M of the crowd, to an end position located in below this plan. We denote by P ₁ a point representative of the position of the smooth 1 at the beginning of the N pick, this point being defined by the abscissa θ ₁ and by an ordinate y _{1 which is} strictly positive. We denote by P ₂ a point representative of the position of the smooth at the end of the N pick, this point being defined by the abscissa θ ₂ and a strictly negative ordinate y ₂ . We denote P ₃ an intermediate point corresponding to the position of the smooth for a value of the angle θ ₃ corresponding to the mid-race on the N duit, that is to say equal to the half sum of the values of θ ₁ and θ ₂ . In the case where θ ₁ is equal to 0 and θ ₂ is equal to 360 °, the intermediate position P ₃ corresponds to the position of the bar at 180 °, this position normally corresponding to the swinging stroke of the loom. Note Δ _i the vertical offset of the point P ₃ relative to the plane P _M , that is to say the absolute value of the abscissa y ₃ of this point which, in the example, is strictly positive.

Note P ₄ a representative point of the necessary opening of the shed for the passage of a weft yarn in the crowd on an angular stroke starting from the initial angular position θ ₁ . We note P ₅ a representative point of the necessary opening of the shed for the passage of a weft yarn on an angular stroke leading to the final angular position θ ₂ . The points P ₄ and P ₅ are defined by their abscissas θ ₄ and θ ₅ and their ordinates y ₄ and y ₅ which are respectively strictly positive and strictly negative. The value of y ₄ lies between the value of Δ ₁ and the value of y ₁ .

The curve C along which a rail moves during a pick N can be characterized by the points P ₁ to P ₅ which are remarkable points corresponding to specific positions of the rail, respectively at the beginning, at the end, at the middle of its displacement as well as in two positions resulting from the desired opening profile for the passage of a weft thread.

The respective abscissae and ordinates of the remarkable points P ₁ to P ₅ come from the data of the set of data D ₁ stored in the memory 30.

When it is necessary to determine the vector V _N corresponding to a pick N for an actuator 6, the computer 26 accesses, in a first step 101 of the process represented in FIG. figure 7 , at the memory 30 to collect the information relating to the points P ₁ to P ₅ for the pick N as well as for the two previous picks N-1 and N-2 and for the two picks N + 1 and N + 2. The computer thus acquires in the memory 30 information relating to the points P _1k , P _2k , P _3k , P _4k and P _5k for the k equal to N-2, N-1, N, N + 1 and N + 2.

In practice, the calculator 26 recovers from the data set D ₁ certain crowd parameters such as the amplitude Amp crowd, the desired vertical offset Δ ₁ at the intermediate point P ₃ and the desired opening for the passage of the son of frame. On the basis of these data, the calculator 26 can calculate the ordinates y ₁ , y ₂ and y ₃ and the abscissae and ordinates θ ₄ , θ ₅ , y ₄ , y ₅ of the points P ₄ and P ₅ . Since, by default, each pick is spread over a business cycle of 360 ° of rotation of the main shaft of the loom, the abscissae of the end points P ₁ , P ₂ and the intermediate point P ₃ are known.

In a step 102, the computer 26 determines a parameterized function f _k for approximating the curve C as a function of the angle of operation θ, for each of the N-2 to N + 2 picks, taking into account the constraints represented by the points P _1k to P _3k , where k is from N-2 to N + 2.

By way of example, the parameterized function f _k can take the form of a cosine decomposition such that $$\begin{array}{l}\mathrm{there}\mathrm{=}{\mathrm{f}}_{\mathrm{k}}\left(\mathrm{\theta }\right)\\ ={\mathrm{at}}_{\mathrm{0}}\mathrm{+}{\mathrm{at}}_{\mathrm{1}}\mathrm{.}\cos \left(\mathrm{\omega \theta }\right)\mathrm{+}{\mathrm{at}}_{\mathrm{2}}\mathrm{.}\mathrm{<figure-callout\; id="2"\; label="cos"\; filenames="imgf0001.png,imgf0003.png"\; state="\{\{state\}\}">cos</figure-callout>}\left(\mathrm{2}\mathrm{\omega \theta }\right)\mathrm{+}{\mathrm{at}}_{\mathrm{3}}\mathrm{.}\mathrm{<figure-callout\; id="3"\; label="cos"\; filenames="imgf0003.png"\; state="\{\{state\}\}">cos</figure-callout>}\left(\mathrm{3}\mathrm{\omega \theta }\right)\mathrm{+}...\mathrm{+}{\mathrm{at}}_{\mathrm{m}}\mathrm{.}\cos \left(\mathit{m}\mathrm{\omega \theta }\right)\\ \mathrm{=}{\displaystyle \underset{\mathit{i}\mathrm{=}\mathrm{0}}{\overset{\mathit{m}}{\mathrm{\Sigma }}}}{\mathrm{at}}_{\mathrm{i}}\mathrm{.}\cos \left(\mathrm{i}\mathrm{.}\mathrm{\omega \theta }\right)\end{array}$$

where m is an integer which is adapted to the number of constraints to respect according to the number of remarkable points, ω is the pulsation whose value is equal to Π / 360 ° when θ is expressed in degrees.

The determination of the parameterized function f _k consists in calculating the coefficients a ₀ , a ₁ , ..., a _m by solving a system whose number of equations is equal to the number of constraints to be respected, that is to say say equal to m + 1. In the present case, there are three constraints each corresponding to one of the points P _1k to P _3k .

In the above hypothesis, where the parameterized function f _k is a periodic function whose half-period extends between the two extreme points P ₁ and P ₂ , the Fourier theory provides tools for solving the equation system. which translates the realization of the constraints at the different points P ₁ to P ₃ .

The determination of this parameterized function y = f _k (θ) takes place for the pick N, as well as for each of the two previous picks, N-2 and N-1, and each of the two picks N + 1 and N + 2.

où m est égal au nombre de contraintes à respecter moins une.As a variant, the function f _k can be a polynomial function of the type $$\mathrm{there}\mathrm{=}{\mathrm{f}}_{\mathrm{k}}\left(\mathrm{\theta }\right)\mathrm{=}{\mathrm{<figure-callout\; id="0"\; label="b"\; filenames="imgf0006.png"\; state="\{\{state\}\}">b</figure-callout>}}_{\mathrm{0}}\mathrm{+}{\mathrm{<figure-callout\; id="1"\; label="b"\; filenames="imgf0001.png,imgf0003.png"\; state="\{\{state\}\}">b</figure-callout>}}_{\mathrm{1}}\mathrm{\theta }\mathrm{+}{\mathrm{<figure-callout\; id="2"\; label="b"\; filenames="imgf0001.png,imgf0003.png"\; state="\{\{state\}\}">b</figure-callout>}}_{\mathrm{2}}{\mathrm{<figure-callout\; id="2"\; label="\theta "\; filenames="imgf0001.png,imgf0003.png"\; state="\{\{state\}\}">\theta </figure-callout>}}^{\mathrm{2}}\mathrm{+}\mathrm{b}\mathrm{...}\mathrm{+}{\mathrm{b}}_{\mathit{m}}{\mathrm{\theta }}^{\mathit{m}}$$

where m is equal to the number of constraints to respect minus one.

Step 102 is also applicable in the case where, during the N pick, the rail remains on the same side of the median plane of the shed P _M , as shown in FIG. figure 3 . In this case, it is possible to define a first remarkable point P ₁ as being representative of the starting position of the beam, a second remarkable point P ₂ being representative of the end position of the beam, each of these points being defined. by an abscissa θ ₁ or θ ₂ and an ordinate y ₁ or y ₂ . It is also possible to define an intermediate remarkable point P ₃ whose abscissa θ ₃ is equal to 180 °, in the classical case where θ ₁ is 0 ° while θ ₂ is equal to 360 °, and whose ordinate y ₃ corresponds to a vertical offset Δ ₂ with respect to the plane P _M. The variation of the ordinates y of the points of the curve C, between the remarkable points P ₁ and P ₂ , corresponds to a movement of bringing the string of warp yarns closer to the line which is known as the "closure of the step". . In this case, the ordinates y ₁ and y ₂ are known, while the ordinate of the intermediate point P ₃ is known and its derivative is zero.

As in the case of figure 2 , a parameterized function of approximation f _N can be used in step 102 to represent the curve C of the figure 3 , this function parameterized that can be decomposed in cosine as envisaged above. The functions defined f _N-2 , f _N-1 , f _{N + 1} and f _{N + 2} are determined as explained above.

In step 102, only the points P ₁ , P ₂ and P ₃ are considered as imposed for the determination of the parameterized functions of approximation f _k . When these functions f _k have been determined for each of the picks N-2, N-1, N, N + 1 and N + 2, it is checked whether the curve constructed from each of these parameterized functions is compatible with the positions of the points P _4k and P _5k as derived from the crowd parameters for each of these picks. In other words, it is verified, for each of the values of k between N-2 and N + 2, that at the abscissa θ ₄ or θ _s of each opening point P ₄ or P ₅ , the ordinate y ₄ or y ₅ calculated using the parameterized approximation function f _k is greater in absolute value than that of the corresponding opening point P _4k or P _5k . This takes place during a step 103. For all the picks from N-2 to N + 2 for which the result of this comparison is negative, we go to a step 104 during which the points P _4k and P _5k are imposed as constraint values for the determination of the approximated parameter functions f _k during step 102.

The number of available equations that each translate a constraint is then five. The parameterized function is of the type $$\mathrm{there}\mathrm{=}{\mathrm{f}}_{\mathrm{k}}\left(\mathrm{\theta }\right)\mathrm{=}{\mathrm{at}}_{\mathrm{0}}\mathrm{+}{\mathrm{at}}_{\mathrm{1}}\mathrm{.}\cos \left(\mathrm{\omega \theta }\right)\mathrm{+}{\mathrm{at}}_{\mathrm{2}}\mathrm{.}\mathrm{<figure-callout\; id="2"\; label="cos"\; filenames="imgf0001.png,imgf0003.png"\; state="\{\{state\}\}">cos</figure-callout>}\left(\mathrm{2}\mathrm{\omega \theta }\right)\mathrm{+}{\mathrm{at}}_{\mathrm{3}}\mathrm{.}\mathrm{<figure-callout\; id="3"\; label="cos"\; filenames="imgf0003.png"\; state="\{\{state\}\}">cos</figure-callout>}\left(\mathrm{3}\mathrm{\omega \theta }\right)\mathrm{+}{\mathrm{at}}_{\mathrm{4}}\mathrm{.}\mathrm{<figure-callout\; id="4"\; label="cos"\; filenames="imgf0001.png"\; state="\{\{state\}\}">cos</figure-callout>}\left(\mathrm{4}\mathrm{\omega \theta }\right)$$

Fourier's theory provides the tools to determine the parameters a ₀ to a ₄ .

After an additional check 103, then a step 105 is reached in which the computer 26 calculates the acceleration of a rail at the beginning of the pick N, considering that this rail follows the path determined by the parameterized function f _N . Indeed, on the basis of equation 1) above, it is possible to calculate the speed of movement of the beam at a point as being equal to $$\mathrm{V}\left(\mathrm{\theta }\right)=\frac{\mathit{dy}\left(\mathit{\theta }\right)}{\mathit{dt}}=-{\displaystyle \underset{i=0}{\overset{m}{\Sigma }}}\mathrm{i}\mathrm{.}{\mathrm{at}}_{\mathrm{i}}\mathrm{\omega sin}\left(\mathrm{i}\mathrm{.}\mathrm{\omega \theta }\right)$$

On the basis of equation 3, it is possible to calculate the acceleration at one point of the N duite as being equal to $$\mathrm{\gamma }\left(\mathrm{\theta }\right)\mathrm{=}{\mathrm{<figure-callout\; id="2"\; label="d"\; filenames="imgf0001.png,imgf0003.png"\; state="\{\{state\}\}">d</figure-callout>}}^{\mathrm{2}}\mathrm{there}\left(\mathrm{\theta }\right)\mathrm{/}{\mathrm{<figure-callout\; id="2"\; label="dt"\; filenames="imgf0001.png,imgf0003.png"\; state="\{\{state\}\}">dt</figure-callout>}}^{\mathrm{2}}=-{\displaystyle \underset{i=0}{\overset{m}{\Sigma }}}{\mathrm{<figure-callout\; id="2"\; label="i"\; filenames="imgf0001.png,imgf0003.png"\; state="\{\{state\}\}">i</figure-callout>}}^2\mathrm{.}{\mathrm{at}}_{\mathrm{i}}{\mathrm{\omega }}^2\mathrm{.}\cos \left(\mathrm{i}\mathrm{.}\mathrm{\omega \theta }\right)$$

This operation is performed for each of the functions f _k corresponding to one of the picks N-2, N-1, N, N + 1 and N + 2.

This makes it possible to consider the acceleration γ _{(N-2) F} at the end of the N-2 pick which is equal to the acceleration for the abscissa θ ₂ corresponding to the parameterized function f _N - ₂ used for the pick N -2.

In the same way, it is possible to determine the acceleration γ _{(N-1) D} at the beginning of the N-1 pick corresponding to the acceleration for the abscissa θ ₁ for the function f _N-1 used for the _N-1 pick. 1, the acceleration γ _{(N-1) F} at the end of the pick N-1, the acceleration γ _ND at the beginning of the pick N from the function f _N , the acceleration γ _NF at the end of the pick N, the acceleration γ _{(N + 1) D} at the beginning of the N + 1 pick from the function f _{N + 1} , the acceleration γ _{(N + 1) F} at the end of the N + pick 1 and the acceleration γ _{(N + 2) D} at the beginning of the N + 2 pick, starting from the function f _{N + 2} .

If the acceleration at the end of a pick defined above is equal to the acceleration at the beginning of the next pick, then there is junction smoothly between the trajectories obtained by the parameterized approximation functions f _k and f _{k + 1} , which is quite positive. In the opposite case, we can consider the average between the acceleration at the end of a pick and the acceleration at the beginning of the next pick, this average being equal to the half-sum of these accelerations.

Thus, a mean acceleration at the end of the N-2 pick and at the beginning of the N-1 pick is defined as being equal to γ _{N-2 / N-1} = γ _(N - _{2) F} + γ _{(N-1) ) D} ) / 2.

• accélération moyenne à la fin de la duite N-1 et au début de la duite N : $${\mathrm{\gamma }}_{\mathrm{N}\mathrm{-}\mathrm{1}\mathrm{/}\mathrm{N}}\mathrm{=}\left({\mathrm{\gamma }}_{\left(\mathrm{N}\mathrm{-}\mathrm{1}\right)\mathrm{F}}\mathrm{+}{\mathrm{\gamma }}_{\mathrm{ND}}\right)\mathrm{/}\mathrm{2}$$
  
• accélération moyenne à la fin de la duite N et au début de la duite N+1 : $${\mathrm{\gamma }}_{\mathrm{N}+\mathrm{1}}\mathrm{=}\left({\mathrm{\gamma }}_{\mathrm{NF}}\mathrm{+}{\mathrm{\gamma }}_{\left(\mathrm{N}+\mathrm{1}\right)\mathrm{D}}\right)\mathrm{/}\mathrm{2}$$
  
• accélération moyenne à la fin de duite N+1 et au début de la duite N+2 : $${\mathrm{\gamma }}_{\mathrm{N}\mathrm{+}\mathrm{1}\mathrm{/}\mathrm{N}\mathrm{+}\mathrm{2}}\mathrm{=}\left({\mathrm{\gamma }}_{\left(\mathrm{N}\mathrm{+}\mathrm{1}\right)\mathrm{F}}\mathrm{+}{\mathrm{\gamma }}_{\left(\mathrm{N}\mathrm{+}\mathrm{2}\right)\mathrm{D}}\right)\mathrm{/}\mathrm{2}$$
  

The following accelerations are defined in the same way:

• average acceleration at the end of the N-1 pick and at the beginning of the N pick: $${\mathrm{\gamma }}_{\mathrm{NOT}\mathrm{-}\mathrm{1}\mathrm{/}\mathrm{NOT}}\mathrm{=}\left({\mathrm{\gamma }}_{\left(\mathrm{NOT}\mathrm{-}\mathrm{1}\right)\mathrm{F}}\mathrm{+}{\mathrm{\gamma }}_{\mathrm{ND}}\right)\mathrm{/}\mathrm{2}$$
  
• average acceleration at the end of the N pick and at the beginning of the N + 1 pick: $${\mathrm{\gamma }}_{\mathrm{NOT}+\mathrm{1}}\mathrm{=}\left({\mathrm{\gamma }}_{\mathrm{NF}}\mathrm{+}{\mathrm{\gamma }}_{\left(\mathrm{NOT}+\mathrm{1}\right)\mathrm{D}}\right)\mathrm{/}\mathrm{2}$$
  
• average acceleration at the end of the N + 1 pick and at the beginning of the N + 2 pick: $${\mathrm{\gamma }}_{\mathrm{NOT}\mathrm{+}\mathrm{1}\mathrm{/}\mathrm{NOT}\mathrm{+}\mathrm{2}}\mathrm{=}\left({\mathrm{\gamma }}_{\left(\mathrm{NOT}\mathrm{+}\mathrm{1}\right)\mathrm{F}}\mathrm{+}{\mathrm{\gamma }}_{\left(\mathrm{NOT}\mathrm{+}\mathrm{2}\right)\mathrm{D}}\right)\mathrm{/}\mathrm{2}$$
  

If the accelerations averaged are zero, then their average is zero. As averages, the accelerations γ _{N-2 / N-1} , γ _{N-1 / N} , γ _{N / N + 1} and γ _{N + 1 / N + 2} depend on the accelerations γ _{(N-2) F} , γ _{(N-1) D} ··· γ _{(N + 1) D.} As a variant, the accelerations γ _{N-2 / N-1} , γ _{N-1 / N} , γ _{N / N + 1} and γ _{N + 1 / N + 2} can depend on the accelerations γ _(N - _{2) F} to γ _{( N + 2) D} in another way.

The computations of these accelerations γ _{N-2 / N-1} at γ _{N + 1 / N + 2} , at the transition between two successive picks are carried out during a step 106 by the computer 26.

In a subsequent step 107, the type of motion sequence corresponding to the succession of the three picks centered on the N pick, namely the N-1, N and N + 1 picks, is determined. Indeed, it can be considered that there is a movement M during a pick if a rail passes from a position above the median plane P _M to a position below, or vice versa. If the value 1 corresponds to a high position of the arm and the value 0 corresponds to a low position, we consider that there is movement if the armor goes from 1 to 0 or from 0 to 1. On the contrary, we consider there is a stop A during this duit if the armor remains at 1 or at 0, this being able to correspond to a movement of closing of the pitch, as explained above with reference to the figure 3 . For example, for a type 0010 armor, the movement sequence is of the AMM type since, at the N-1 pick, the frame remains stationary at the bottom (position 0), at the pick N, the frame moves from its low position. (0) at its high position (1) and, at the N + 1 pick, the frame returns to the lower position from the high position (1) to the low position (0).

• AAA correspondant aux armures 0000 et 1111,
• AAM correspondant aux armures 0001 et 1110,
• AMA correspondant aux armures 0011 et 1100,
• AMM correspondant aux armures 0010 et 1101,
• MAA correspondant aux armures 0111 et 1000,
• MAM correspondant aux armures 0110 et 1001,
• MMA correspondant aux armures 0100 et 1011, et
• MMM correspondant aux armures 0101 et 1010.

There are eight types of sequence of movements for three successive picks, namely:

• AAA corresponding to armor 0000 and 1111,
• AAM corresponding to the armor 0001 and 1110,
• AMA corresponding to the armor 0011 and 1100,
• AMM corresponding to the armor 0010 and 1101,
• MAA corresponding to 0111 and 1000 armor,
• MAM corresponding to armor 0110 and 1001,
• MMA corresponding to the armor 0100 and 1011, and
• MMM corresponding to 0101 and 1010 armor.

The types of motion sequences AMA, AMM and MMA have the particularity that the movement M which takes place at the N pick is contiguous with at least one stop, during one of the picks N-1 or N + 1.

When the type of the motion sequence has been determined in step 107, two cases are considered: either this motion sequence is of the AMA, AMM or MMA type, or it is of a different type.

If the motion sequence determined in step 107 is of a type other than AMA, AMM or MMA, the computer 26 proceeds to a step 108 during which it determines a parametric motion law L which gives the abscissa y a smooth as a function of the angle θ during the N duit, through the remarkable points P ₁ to P ₃ , respecting the opening constraints P ₄ and P ₅ and the transition accelerations determined at step 106.

The calculator 26 proceeds according to an approach similar to that adopted for the step 102, with additional constraints. A first calculation is performed to determine the parameters of an approximation law which passes through the remarkable points P _1N , P _2N and P _3N and presents the accelerations calculated in step 106. Next, the ordinates calculated on the abscissa of the points P _4N and P _5N are compared with the values y ₄ and y ₅ . If these computed coordinates are greater in absolute value than the values y ₄ and y ₅ , then the parameterized approximation law y = L (θ) is suitable and is retained. In the opposite case, a new calculation of the parameters of the law of approximation is carried out by imposing the passage by the points P _1N to P _5N and the respect of the accelerations calculated in the step 106.

The parametric motion law L (θ) can be of the form $$\mathrm{there}\mathrm{=}\mathrm{The}\left(\mathrm{\theta }\right)\mathrm{=}{\mathrm{at}}_{\mathrm{0}}\mathrm{+}{\mathrm{at}}_{\mathrm{1}}\mathrm{.}\cos \left(\mathrm{\omega \theta }\right)\mathrm{+}{\mathrm{at}}_{\mathrm{2}}\mathrm{.}\mathrm{<figure-callout\; id="2"\; label="cos"\; filenames="imgf0001.png,imgf0003.png"\; state="\{\{state\}\}">cos</figure-callout>}\left(\mathrm{2}\mathrm{\omega \theta }\right)\mathrm{+}{\mathrm{at}}_{\mathrm{3}}\mathrm{.}\mathrm{<figure-callout\; id="3"\; label="cos"\; filenames="imgf0003.png"\; state="\{\{state\}\}">cos</figure-callout>}\left(\mathrm{3}\mathrm{\omega \theta }\right)\mathrm{+}{\mathrm{at}}_{\mathrm{4}}\mathrm{.}\mathrm{<figure-callout\; id="4"\; label="cos"\; filenames="imgf0001.png"\; state="\{\{state\}\}">cos</figure-callout>}\left(\mathrm{4}\mathrm{\omega \theta }\right)\mathrm{+}{\mathrm{at}}_{\mathrm{5}}\mathrm{.}\cos \left(\mathrm{5}\mathrm{\omega \theta }\right)\mathrm{+}{\mathrm{at}}_{\mathrm{6}}\mathrm{.}\mathrm{<figure-callout\; id="6"\; label="cos"\; filenames="imgf0001.png"\; state="\{\{state\}\}">cos</figure-callout>}\left(\mathrm{6}\mathrm{\omega \theta }\right)$$

The coefficients a ₀ to a ₆ can be determined since they are the solutions of a system of seven equations, each of which expresses the realization of a constraint.

This part or process is performed during steps 103 'and 104' which are comparable to steps 103 and 104, except that the checks and calculations are performed only by the N pick.

If the type of the motion sequence determined in step 107 is AMA, AMM or MMA, it can be taken into account that a rail is stationary during N-1 and / or N + 1 picking. to spread the movement of the beam during the N duit "overflowing" on the N-1 or N + 1 pick.

In the case of an AMM type movement represented at figure 5 , the beam is normally stopped during the N-1 pick, while it passes from position 1 to position 0 during the pick N. It is envisaged that the remarkable point P ₁ corresponding to the starting position of movement for the smooth, as defined above with reference to the figure 2 , is moved within the N-1 pick to a point P ' ₁ with the same ordinate y ₁ as the point P ₁ . This is obtained by defining a spreading parameter Δθ of the movement of the bar from the N pick to the N-1 pick. This parameter Δθ is determined iteratively by moving the point P ' ₁ representative of the beginning of displacement with respect to the beginning of the pick N and verifying that this displacement remains compatible with the opening points P ₄ and P ₅ . For this, the computer 26 determines a parameterized approximation function of the same type as the parameterized approximation function f _N determined in the step 108 which satisfies the crossing constraints at the points P ' ₁ , P ₃ and P ₂ and presents zero acceleration at point P ' ₁ and a transition acceleration determined in step 106 at point P ₂ . This approximation function is defined on the interval A ₀ corresponding to the sum of 360 ° and of Δθ. In the example of the figure 5 , the point P ' ₁ is moved to the left as long as the ordinate of the point P ₄ remains compatible with the passage of a weft thread in the crowd. The value of the parameter Δθ can be sought in a dichotomous manner or according to another approach, at the choice of the programmer of the computer 26.

For example, the value of the parameter Δθ can initially be set at 10 °, so that the movement corresponding to the pick N is provided with an angular amplitude of 370 °. The computer 26 then determines a parameterized approximation function according to an approach similar to that envisaged for step 108, but defined on the interval A ₀ equal to 360 ° + Δθ. The computer then verifies the compatibility of this parameterized approximation function with the opening points P ₄ and P ₅ . If these conditions are met, the value of the offset parameter Δθ is increased, for example by 10 °. In the opposite case, the value of this parameter is reduced by 50%, ie by 5 °. This can be done according to several occurrences, in order to determine, during a step 109, an optimum value of the parameter Δθ for which the conditions of passage to the opening points P ₄ and P ₅ are satisfied, while the value of Δθ is maximum. The motion law y = L (θ) is then calculated, at a subsequent step 110, taking into account the non-zero parameter Δθ determined during step 109.

The displacement of the smooth corresponding to the N pick is thus performed on an angular amplitude A ₀ equal to 360 ° + Δθ. In the example shown in figure 5 , Δθ is 120 ° and A ₀ is 480 °.

In the case of an MMA type movement, the beam is moving at the N-1 and N picks, while it does not change position relative to the P _M plane at the N + 1 pick. In a manner analogous to the case of a movement of the AMM type, the computer 26 determines a parameterized motion law y = L (θ) and the value of an offset parameter Δθ then corresponding to the spread on the pick N + 1, from the movement of the smooth to the N.

In the example of the figure 6 corresponding to an AMA-type motion sequence, a first parameter Δθ is used to spread the N pick to the N-1 pick while a second parameter Δ'θ is used to spread the N pick to the N + 1 pick. For the rest, the operation is comparable to that mentioned with reference to the figure 5 and the displacement of the smooth corresponding to the pick N is performed on an angular amplitude A _θ equal to 360 ° + Δθ + Δ'θ, ie about 515 ° in the example.

In this case, two new remarkable points P ' ₁ and P' ₂ are defined, respectively before the beginning of the angular range of 360 ° normally corresponding to the pick N and after this range. The points P ' ₁ and P' ₂ are taken into account for the determination of the motion law y = L (θ) in step 110.

As represented in figure 4 , the possibility of spreading the pick N on the previous and next picks N-1 and N + 1 makes it possible to shift the movement of the heddles on the picks during which there is no displacement planned, which is the case for the N-1 and N + 2 picks. The solid line corresponding to the curve C of displacement of a bar is offset with respect to the basic movement of the rails represented by the dashed lines, at the respective beginning and end of the N-1 and N + 2 picks. In other words, the displacements at the N-2 and N picks overflow at the N-1 pick, while the displacements at the N + 1 and N + 3 picks overflow the N + 2 pick.

After step 108 or step 110, the vector V _N corresponding to the pick N is generated, in a step 111, in the form of a set of setpoint data which are supplied to the amplifier 16 of each actuator 6.

According to a not represented variant of the invention, the spreading parameter Δθ can be determined taking into account the accelerations γ _{N-2 / N-1} and γ _{N + 1 / N + 2} .

Thanks to the invention, the movement of each frame of smooth is continuous and harmonious because the accelerations are relatively low and continuous. Whatever the movement sequence of the pick, the motion law y = L (θ) has continuous accelerations across the pick. In fact, by allowing the movement phases to be spread on the stop picks, for certain types of movement sequence, the accelerations necessary for the setting in motion or the stopping of each frame decrease. In addition, the calculation method used makes it possible to reduce constraints on the curves used. For example, the constraints at the opening points P ₄ and P ₅ are most often verified in steps 103 and 103 ', and not imposed.

When the setpoint vectors V _N have been calculated for all the actuators 6, for a pick N, it is possible to calculate the same vectors for the pick N + 1. In this case, it may be possible to take into account the calculations already carried out for the analyzes of the N pick. For example, the transition accelerations γ _{N-1 / N} , γ _{N + 1 / N} and γ _{N + 2 / N + 1} , between the N-1 and N, N and N + 1, N + 1 and N + 2 pickovers are already known from the calculation relating to the N pick, as well as the N and N + 1 movement sequences. It may be taken into account in steps 105 and 106. On this occasion, the knowledge of the remarkable points P ₁ , P ₂ , P ₃ , P ' ₁ and P' ₂ and possibly the spreading parameter Δθ of the pick N can be used to determine the law of this movement for the N + 1 pick.

When the calculator 26 has made these calculations for all the picks of an armor ratio, it returns to the first pick of this armor ratio and performs the same calculations again for all picks in the report. In this way, it is possible for him to take into account, during weaving, a modification of the crowd parameters since he accesses the memory 30 during these calculations. For example, if the user wishes to increase the amplitude Amp of the crowd on one of the frames 4, he can modify the set of data D ₁ through the interface 32 and this value can be taken into account by the calculator 26 during the successive steps of determining the motion laws. The effective taking into account of the modification carried out by the user takes place as soon as possible, considering the delay which separates the transmission of the position instructions for the N pick at the beginning of this pick and the number of picks intervening by anticipation in the calculation of position instructions.

The invention also makes it possible to envisage an application of the modification of the crowd parameters which does not result from the intervention of a user, such as the weaver, but from a real-time analysis based on the results. for example, in the event of detection of a large frame stop rate, the craft controller 22 can increase the aperture at points P ₄ and P ₅ , that is to say the absolute value of the abscissae y ₄ and y ₅ , which is taken into account at a later stage, during the next calculation of the motion law y = L (θ) for the pick corresponding, without stopping the trade 2.

It is also possible for the controller 20 to take into account the level of stress in the rods 8, as detected by cells 10 and transmitted to the conditioner 14, to change during weaving the shedding parameters of the data set D ₁ , so as to generate parametric motion laws y = L (θ) that cause a lowering this level of stress. For example, the amplitude Amp of the crowd can be decreased for all the picks of the armor. This decrease in amplitude is generally accompanied by a decrease in the forces in the transmission elements.

The invention has been described in the case where the N-2 pick and the N + 2 pick are taken into account, in addition to the N-1 and N + 1 picks, for the determination of the motion law y = L ( θ) corresponding to the N pick. According to a simplified version of the invention, only the N-1, N and N + 1 picks can be taken into account at this stage.

The invention has been shown in the case of electric actuators 6 controlling smooth frames 4. It is also applicable to the case of actuators individually controlling smooth, as envisaged in EP-A-926,279 .

The invention has been described in the case where an actuator is controlled individually according to a motion profile. Alternatively, several actuators can be controlled according to such a profile, for example if a pattern is reproduced over the width of a fabric. In this case, a set of setpoint data is generated for a group of actuators.

The invention has been described in the case where, for k between N-2 and N + 2, the parameterized functions f _k and L (θ) are in the form of a cosine decomposition. Other types of functions can be envisaged. In addition to the case already mentioned of a parameterized function of polynomial type, a parameterized function of spline type can be used. It also ensures that the amplitude of the crowd never exceeds the sum of the absolute values of the values defined at the terminals of each pick.

The invention has been described in the case where the amplitude of the crowd remains constant from one pick to the other. The invention also allows the determination of positional instructions in the case where the amplitude of the crowd varies from one pick to the other. In this case, during step 101, the computer determines the remarkable points as a function of the crowd parameters among which are the amplitudes. The other calculation steps can then take place as envisaged above.

In the case where computers integrated with the amplifiers 16 are sufficiently efficient, the computer 26 can transmit to the amplifiers 16 the parameters of the parameterized functions L (θ), instead of the vectors V _N , which allows the generation of the instructions at the level of the amplifiers 16 more precisely. In this case also, a set of setpoint data is provided, for each pick and each actuator, by the amplifier 16 associated with each actuator 6.

Claims (10)

A method of controlling the electric actuators (6) of a shedding device on a loom (2) in which a set of setpoint data (V _N ) is generated for each pick (N) and each actuator or group of actuators, taking into account the predetermined shedding parameters (D ₁ ), characterized in that it comprises, for a pick (N) and a given actuator (6) or group of actuators, steps constituting : a) determining (101), as a function of the shedding parameters (D ₁ ), the remarkable points (P ₁ -P ₅ ) of a movement profile of a rail driven by the actuator (6) or the group d actuators, at least for the given pick (N), the previous pick (N-1) and the next pick (N + 1); b) determining (102) a parameterized function (f _k ) for approximating the motion profile passing through certain remarkable points (P ₁ , P ₂ , P ₃ ), at least for the given pick (N), the previous pick (N-1) and the next pick (N + 1); c) calculating (105) at least four accelerations, namely: a first acceleration (γ _ND ) of the smooth at the beginning of its movement (θ ₁ ) from the approximation function (f _N ) determined in step b) for the given pick (N), a second acceleration (γ _NF ) of the smooth at the end of its movement (θ ₂ ) from the approximation function (f _N ) determined in step b) for the given pick (N), a third acceleration (γ _{(N-1) F} ) of the smooth at the end of its movement (θ ₂ ) from the approximation function (f _N-1 ) determined in step b) for the pick previous (N-1), a fourth acceleration (γ _{(N + 1) D} ) of the smooth at the beginning of its movement (θ ₁ ) from the approximation function (f _{N + 1} ) determined in step b) for the next pick (N + 1), d) calculating (106) a fifth acceleration (γ _{N-1 / N} ) as a function of the first and third accelerations, and a sixth acceleration (γ _{N / N + 1} ), as a function of the second and fourth accelerations; e) determining (108, 110) a motion law (y = L (θ)) of the smooth for the given pick (N), whose profile passes through certain remarkable points (P ₁ -P ₃ ) and such that the acceleration of the beam at the beginning of its movement is equal to the fifth acceleration (γ _{N-1 / N} ) and that the acceleration of the beam at the end of its movement is equal to the sixth acceleration (γ _{N / N + 1} ); f) generating (111) the set of setpoint data (V _N ), for the given pick (N) and for the actuator (6) or the group of actuators, from the motion law (y = L (θ)) determined in step e). Method according to Claim 1, characterized in that the fifth acceleration (γ _{N-1 / N} ) is equal to the average of the first and third accelerations (γ _ND , γ _{(N-1) F} ) and the sixth acceleration (γ _{N / N + 1} ) is equal to the average of the second and fourth accelerations (γ _NF , γ _{(N + 1) D} ). Method according to one of claims 1 or 2, characterized in that it comprises steps prior to step e) and consisting of: g) determining (107) the type of motion sequence of the smooth on the previous pick (N-1), the given pick (N) and the next pick (N + 1), within a group of eight motions type (AAA, AAM, AMA, MA, MAA, MAM, MMA, MMM); h) if the type of motion sequence (AMA, AMM, MMA) determined in step g) is such that the arm is moved for the given pick, between two positions (0, 1) respectively located on the side and another of a median plane (P _M ) of the crowd, and that the smooth is not moved between these two portions for the previous pick (N-1) and / or the next pick (N + 1), determine ( 109) at least one parameter (Δθ, Δ'θ) for spreading the movement of the arm taking into account at least some of the remarkable points (P _4, P ₅ ); i) when the spreading parameter has been determined, determining (109) at least one new remarkable point (P ' ₁ , P' ₂ ), as a function of the shedding parameters (D ₁ ) and the spreading parameter (Δθ) ) and in that , in step e), the motion law (y = L (θ)) is determined (110) taking into account the new remarkable point (P ' ₁ , P' ₂ ) and the parameter d spreading (Δθ). Method according to claim 3, characterized in that during step h), the spreading parameter (Δθ) is determined by successive tests of the influence of this parameter on the compatibility of a motion law with at least a remarkable point (P _4, P ₅ ). Method according to Claim 4, characterized in that the compatibility of the spreading parameter (Δθ) with the motion law is tested with one or more remarkable points (P _4, P ₅ ) representative of a crowd opening geometry , for the pick (N) and the actuator (6) or the group of actuators concerned. Method according to one of Claims 3 to 5, characterized in that the spreading parameter (Δθ) is representative of the portion greater than 360 ° of the angular amplitude (Aθ), with respect to the rotation of the shaft occupation, on which takes place the displacement of the smooth for the given pick (N). Method according to one of the preceding claims, characterized in that data (D ₁ ) representative of the crowd parameters are stored in a dynamic memory (30) to which it is accessed during step a) and that Data is editable during weaving. Method according to one of the preceding claims, characterized in that during step b), a parameterized function (f _k ) for approximating the motion profile passing through certain remarkable points for a pick (N- 2) preceding by two strokes the given pick (N) and a pick (N + 2) following by two strokes the given pick, in that , in step c), also calculating: a seventh acceleration (γ _{(N-1) D} ) of the smooth at the beginning of its movement (θ ₁ ) from the approximation function (f _N-1 ) determined in step b) for the previous pick , an eighth acceleration (γ _{(N + 1) F} ) of the smooth at the end of its movement (θ ₂ ) from the approximation function (f _{N + 1} ) determined in step b) for the pick next, a ninth acceleration (γ _{(N-2) F} ) of the smooth at the end of its movement (θ _d from the approximation function (f _N-2 ) determined in step b) for the preceding pick the previous pick (N-2), a tenth acceleration (γ _{(N + 2) D} ) of the smooth at the beginning of its movement (θ ₁ ) from the approximation function (f _{N + 2} ) determined in step b) for the next pick the next pick (N + 2), in that an eleventh acceleration (γ _{N-2 / N-1} ) is calculated as a function of the seventh and ninth accelerations and a twelfth acceleration (γ _{N + 1 / N + 2} ) as a function of the eighth and tenth accelerations and in that , in step e), determining (108, 110) the law of movement (y = L (θ)) of the bar for the given pick so that the acceleration of the bar for the previous pick (N-1) at the beginning of the movement is equal to the eleventh acceleration (γ _{N-2 / N-1} ) and the acceleration of the smooth for the next pick (N + 1) at the end of its movement is equal to the twelfth acceleration (γ _{N + 1 / N + 2} ). Method according to one of the preceding claims, characterized in that it comprises steps prior to step c) and consisting of: j) checking (103) the compatibility of the approximation function (f _k ) with remarkable points (P _4, P ₅ ) representative of a shedding geometry for at least the given pick (N), the preceding pick (N-1) and the next pick (N + 1), and k) in case of incompatibility detected in step j), impose (104) these remarkable points (P _4, P ₅ ) as passing points of the approximation function (f _k ). Method according to one of the preceding claims, characterized in that the parameterized function (f _k ) and / or the motion law (L) is expressed in the form $${\displaystyle \underset{i=0}{\overset{m}{\Sigma }}}{\mathrm{at}}_i\mathrm{.}\cos \left(i\mathrm{.}\mathit{\omega \theta }\right)$$

with θ equal to the angle of the loom in its cycle, m integer greater than or equal to 1 and _i constant, m is preferably less than or equal to 6.

Images