DE102015100669A1 - ANTI-PENDULUM CONTROL PROCEDURE WITH ADJUSTABLE SUPPORT FOR THE TRANSPORT OF AN ANCHORED LOAD - Google Patents

Abstract

The invention relates to a method for controlling the movement of a suspended load at a lifting point of a tower crane hoist, the lifting point being designed to move by yawing about a first vertical axis (ZZ ') and / or according to a feed movement (T) a second axis (XX ') that intersects said orientation axis (ZZ'), said method comprising a speed setpoint detection step (Vi), then a speed setpoint filtering step by a virtual model (Ẋ = AX + BU) using a state vector (X) having an angular component of pendulum swing and a pendulum angular velocity component, and to which a state control (Ui = ki0 · Vi - Ki · X) is applied whose correction gains (ki1, ki2) in response to a same typing parameter (Tc) freely set by the operator of the hoist lt can be.

Description

The present invention relates to the general field of hoists, such as tower cranes, comprising a mobile lifting point such as a trolley to which a load to be moved can be hung, and the support methods for controlling such hoists.

It is already known to provide a control method which aims to control and, in particular, limit the amplitude of the pendulum oscillations or "pendulum" to which the suspended load is exposed during movement of the crane truck in order to ensure the accuracy and safety of the transport operations of said cranes Load to improve.

For this purpose it is in particular, for example through the document FR-2 704 847 is known to first design a position setpoint for the car by filtering said setpoint through a frequency-controlled "hole" filter designed to eliminate from said setpoint those frequencies suitable to trigger a vibrational mode of the load; and thereafter apply that setpoint to said carriage, and then counteract the commuting by closed loop driving, measuring the actual position and speed of the car as well as the pendulum angle of the load, and then one Correction of the setpoint generated taking into account the detected deviation between the real behavior of the car and the applied setpoint.

Although such a system actually makes it possible to stem commuting, it may nevertheless have certain disadvantages.

First of all, the simple filtering of the setpoint by a "hole" filter, which avoids certain resonance phenomena, is generally insufficient to prevent the occurrence of pendulum oscillations, which are thus to be combated by a complementary control in a closed loop.

However, such a closed-loop control requires the use of numerous sensors, for example, designed to measure the real pendulum angle, thereby reducing the complexity, and therefore the cost, and the risk of disturbing the control system, and in general the Hoist equipped with such a Steursystem increases.

In addition, the complexity of the model used by this control system, as well as the amount of data to be measured and processed, tend to mobilize relatively large and costly amounts of resources in terms of computational power, memory, and power.

In addition, the assistance offered or even imposed on the operator of the hoist (crane operator) may tend to exaggerate the reactions of the hoist to the crane operator's specifications, thus distorting the intuitive sense of the machine's behavior that the crane operator may have. and above all, it gives the latter the unpleasant impression that the machine lacks response and does not faithfully meet the specifications.

The objects of the invention, therefore, seek to remedy the above-mentioned drawbacks and to propose a new control method for moving a suspended load, which ensures efficient control of hunting, which can be implemented easily and inexpensively, and a true feeling for a reactive and relatively intuitive control.

• – man das Pendelverhalten der schwebenden Last gemäß der durch ein virtuelles Modell Ẋ = AX + BU angenommenen Bewegung modelliert, welches einen Zustandsvektor X verwendet, der zumindest eine „Hauptkomponente” genannte Komponente umfasst, die der angesteuerten Größe entspricht, und andere Komponenten, die „Zusatzkomponenten” genannt werden, und repräsentativ für die charakteristischen kinematischen Größen für die Pendelbewegung der schwebenden Last sind, wie der Pendelwinkel oder die Pendel-Winkelgeschwindigkeit gemäß der angenommenen Bewegung,
• – man wendet auf das besagte virtuelle Modell Ẋ = AX + BU eine Zustandsregelung U_i = k_i0·V_i – K_i·X an, die über ein Sollwertglied k_i0·V_i hinaus, das repräsentativ für den Brutto-Steuersollwert ist, der im Erfassungsschritt erfasst worden ist, ein Korrekturglied K_i·X umfasst, das dem Produkt eines Korrekturvektors K_i mal dem Zustandsvektor X entspricht, unter Verwendung eines Korrekturvektors K_i, bei dem zumindest bestimmte Korrekturverstärkungen k_i1, k_i2, die den Zusatzkomponenten zugeordnet sind, in Abhängigkeit von einem selben vorbestimmten Typisierungsparameter Tc ausgedrückt werden, wobei der besagte Typisierungsparameter Tc vom Bediener des Hebezeugs frei eingestellt werden kann, um die besagten Korrekturverstärkungen k_i1, k_i2 einstellen zu können,
• – und man extrahiert aus dem virtuellen Modell einen Steuersollwert, der „gefilterter Steuersollwert” genannt wird, der der Hauptkomponente des Zustandsvektors X entspricht.

The objects associated with the invention are achieved by a method of controlling the movement of a suspended load at an attachment point of a hoist, said abutment point being designed to be referred to by yawing about a first vertical axis (ZZ '), the "orientation axis" , and / or by means of advancing movement along a second axis (XX ') called the "distribution axis" and said vertical axis (ZZ') intersects, the movement of the levitated load being controlled by a control target value comprising a Quantity, which is typically a kinematic variable characteristic of the movement of said load, called "controlled quantity", such as the rotational speed, and / or the advancing speed of the abutment point, said method comprising:
a step (a) of detecting the target value during which a control target value is detected in real time called the "gross control target value" V _i and corresponding to the commanded value that the operator of the hoist is aiming at the presumed time,
then the step (b) of filtering the setpoint, during which:

• Modeling the pendulum behavior of the suspended load according to the motion assumed by a virtual model Ẋ = AX + BU, which uses a state vector X comprising at least one component called "principal component" corresponding to the controlled quantity, and other components which " Are representative of the characteristic kinematic variables for the pendulum movement of the suspended load, such as the pendulum angle or the pendulum angular velocity according to the assumed movement,
• It applies to said virtual model Ẋ = AX + BU a state control U _i = k _i0 · V _i - K _i · X, which is beyond a setpoint _element k _i0 · V _i , which is representative of the gross control setpoint, detected in the detecting step comprises a correction term K _i * X corresponding to the product of a correction vector K _i times the state vector X, using a correction vector K _i at which at least certain correction _gains k _i1 , k _{i2 associated} with the additional components are expressed in response to a same predetermined typing parameter Tc, said typing parameter Tc being freely adjustable by the operator of the hoist to adjust said correction _gains k _i1 , k _i2 ,
• And extracting from the virtual model a control setpoint called the "filtered control setpoint" corresponding to the main component of the state vector X.

• – man das Pendelverhalten der schwebenden Last gemäß der durch ein virtuelles Modell Ẋ = AX + BU angenommenen Bewegung modelliert, welches einen Zustandsvektor X verwendet, der zumindest eine Komponente der momentanen Geschwindigkeit umfasst, die repräsentativ für die momentane Geschwindigkeit des Anschlagpunktes gemäß der angenommenen Bewegung ist, eine Komponente des Pendelwinkels, die repräsentativ für den Pendelwinkel der schwebenden Last gemäß der angenommenen Bewegung ist, und eine Komponente der Pendelwinkelgeschwindigkeit, die repräsentativ für die Pendelwinkelgeschwindigkeit gemäß der angenommenen Bewegung ist,
• – man wendet auf das besagte virtuelle Modell Ẋ = AX + BU eine Zustandsregelung U_i = k_i0·V_i – K_i·X an, die über das Sollwertglied k_i0·V_i hinaus, das repräsentativ für den Geschwindigkeitssollwert V_i ist, der im Erfassungsschritt erfasst wurde, ein Korrekturglied K_i·X umfasst, das dem Produkt eines Korrekturvektors K_i mal dem Zustandsvektor X entspricht, unter Verwendung eines Korrekturvektors K_i, bei dem die Korrekturverstärkungen k_i1, k_i2, die jeweils der Komponente der Pendelwinkelgeschwindigkeit und der Komponente des Pendelwinkels des Zustandsvektors X zugeordnet werden, in Abhängigkeit von einem selben vorbestimmten Typisierungsparameter Tc ausgedrückt werden, wobei der besagte Typisierungsparameter Tc vom Bediener des Hebezeugs frei eingestellt werden kann, um die besagten Korrekturverstärkungen k_i1, k_i2 einstellen zu können,
• – und man extrahiert aus dem virtuellen Modell einen gefilterten Geschwindigkeitssollwert, der der Komponente der momentanen Geschwindigkeit des Zustandsvektors X entspricht.

In particular, the objects associated with the invention are achieved by a method of controlling the movement of a levitating load at an attachment point of a hoist, said abutment point being designed to be yawed about a first vertical axis (ZZ ') called the "orientation axis" is, and / or according to a feed movement along a second axis (XX '), which is called "distribution axis", and the said vertical axis (ZZ') intersects to move, said method comprising:
a target value acquiring step (a) during which real-time acquisition of a speed command value V _i representative of the rotational speed or feed rate which the operator of the hoist wishes to transmit to the abutment point at the presumed time is provided;
then the step (b) of filtering the setpoint, during which:

• Modeling the pendulum behavior of the suspended load in accordance with the motion assumed by a virtual model Ẋ = AX + BU using a state vector X comprising at least one instantaneous velocity component representative of the instantaneous velocity of the impact point according to the assumed motion a component of the pendulum angle which is representative of the pendulum angle of the levitated load according to the assumed motion, and a component of the pendulum angular velocity which is representative of the pendulum angular velocity according to the assumed motion,
• It applies to the said virtual model Ẋ = AX + BU a state control U _i = k _i0 · V _i - K _i · X, which is beyond the setpoint member k _i0 · V _i , which is representative of the speed setpoint V _i , detected in the detecting step comprises a correction term K _i × X corresponding to the product of a correction vector K _i times the state vector X, using a correction vector K _i at which the correction gains k _i1 , k _i2 , respectively, of the component of the pendulum angular velocity and the component of the pendulum angle of the state vector X are expressed in dependence on a same predetermined typing parameter Tc, said typing parameter Tc being freely adjustable by the operator of the hoist to be able to adjust said correction _gains k _i1 , k _i2 .
• - And extracted from the virtual model a filtered speed setpoint corresponding to the component of the current speed of the state vector X.

Preferably, the invention thus allows the operator of the hoist to freely select, among a selection of predetermined values or from a range of possible predetermined values, a typing parameter Tc (typing coefficient) which enables him to characterize the filtering applied to his speed setpoints and, more precisely, to adjust the behavior of the driven system by defining the trade-off between response time and stability (damping) that suits it best.

In practice, this very particularly simple and rapid setting option for filtering the speed setpoint, which is preferably the same for the various correction gains to which a same typing parameter is applied, actually allows placement of the poles (complexes) of state control with only a single setting make.

Of course, such a possibility of changing the setting of the setpoint filter may also be used to change both the personal setting from one operator to another, for example, between an experienced crane operator who prefers reactive and poorly damped control, and a crane operator with less experience which prefers muffled and safe, but slower in the reaction, as well as to adapt the behavior of the hoist to the particular working conditions at the construction site, for example, depending on space, frailty or weight of the load to be moved, or but depending on the weather conditions, and above all wind force, which prevail at the assumed time at the site.

Preferably, the use of a virtual shuttle model, coupled with a state control, enables the rapid and relatively precise simulation of the virtual behavior of the load in response to said control, but with relatively modest built-in computing devices, and thus the determination of a controller that provides good load stability , with a minimized pendulum motion, with a relatively fast match of the filtered speed setpoint and the speed setpoint expressed by the crane operator.

In this connection, it should be noted that the filter according to the invention has the main objective of minimizing commuting by using a virtual predictive model of the behavior of the pendulum system formed by the hoist and its hovering load.

In practice, the use of such a virtual model allows the theoretical definition of the control conditions, here the speed conditions, and concretely the conditions of the speed developments with time, and consequently the traveling conditions (of the car), which are such that when the said speed conditions on the real system is applied, the commutation of said real system is attenuated, meaning that it remains within a certain amplitude limit.

Thus, for each "gross" control setpoint established by the crane operator, the filtering immediately converts said "gross" control setpoint to a filtered control setpoint whose time evolution the filter has adjusted based on the virtual model so that, on the one hand, said filtered setpoint develops so far until it reaches the "gross" setpoint desired by the crane operator within a response time that is as short as possible (in view of the typing coefficient selected), that is, more specifically, that the speed of the truck is to be corrected by the crane operator on the other hand, so that this development to match the filtered target value with the gross target value, on the other hand, takes place such that no conditions for the occurrence of commuting are created.

In other words, the control setpoint, and more specifically the speed setpoint filtered by the method of the present invention, preferably corresponds exactly to the control setpoint, and more specifically the speed setpoint to be applied to the real system to match the gross control setpoint desired by the crane operator and at the same time to stem the commuting.

Once the parameters of the virtual control are determined to satisfy the dual criterion of speed and stability according to the desired compromise, that is, once the poles of the transfer function of the virtual model are placed in the closed loop, there is the possibility of said virtual model of state control to be applied as a dynamic filter to filter the gross control setpoint (gross speed setpoint) actually expressed by the crane operator, and then to apply the filtered control setpoint (speed setpoint) to the lifting elements of the hoist (normally controlled by electric motors).

Advantageously, the application of said filtered control setpoint to the real system can be carried out even in an open loop, without the need for, for example, obtaining information on the real oscillation angle. Thus, the structure of the hoist can be facilitated, and the drive algorithm can be simplified.

Finally, it is noteworthy that the method according to the invention enables detection and filtering of the control setpoint in real time, giving this method a high response speed, and capable of promptly adapting the filtered control setpoint to unforeseen fluctuations in the gross control setpoint. which is expressed by the crane operator, especially when the crane operator tries to avoid an unforeseen obstacle, manually balance a pendulum, or counteract the effects of a wind gust.

In other words, the control obtained by the method according to the invention advantageously enables the hoist to follow the setpoint expressed by the crane operator permanently and without triggering unwanted oscillation, even if this desired value expressed by the crane operator can change at any time.

The said control method is thus particularly versatile, since it adapts to a variety of different control scenarios, since it does not depend on the trajectory traveled by the load, nor on the existence or non-existence of imposed passages, and because it is reliable and reacts accurately to the numerous and random changes in the gross control setpoint, and in particular to the changes in the speed setpoint given by the crane operator.

Other objects, features and advantages of the invention will become more apparent upon reading the following specification, and from the accompanying drawings, which are given by way of illustration and not of limitation, wherein:

The 1 the parameterization of a model of the pendulum system in a schematic representation shows that is formed by the hovering load on the hoist.

The 2 FIG. 2 illustrates a block diagram of a dynamic filter based on a virtual model associated with state control.

The 3 in the complex plane represents the principle of placing the poles of a filter according to the invention by adjusting the typing parameter.

The 4 and 5 in time-oriented diagrams, each for a distribution and an orientation movement, represent an example of a filtered setpoint obtained in response to a speed setpoint (gross control setpoint), such as a skip function, for a first typing parameter value, a "reactive" mode set-up equivalent.

The 6 and 7 in time-oriented diagrams, and for each of a distribution and an orientation movement, is an example of a filtered setpoint obtained in response to a speed setpoint (gross control setpoint), such as a skip function, for a second typing parameter value smaller than that the 4 and 5 , and which corresponds to an operating setting in the "damped" mode.

The present invention relates to a control method for moving a load 1 at an anchor point H of a hoist 2 is suspended, wherein said abutment point H is designed to be referred to by yawing R about a first vertical axis (ZZ '), the "orientation axis", and / or by means of feed movement T along a second axis (XX'), the " Distribution axis "is called, and the said vertical axis (ZZ ') intersects, to move, as in 1 is shown.

The attachment point H is preferably formed by a carriage which is guided along a distribution axis (XX ') in the feed direction, said carriage can be driven by any suitable motorized device.

Weight 1 is by means of a hanger 3 , suspended as a support member on the anchor point H , and brought in connection with such a rope as follows.

Preferably, you can the length L of the support cable 3 vary by the distance of the suspended load 1 to the stop point H to change, and the lifting height of the floating load 1 relative to the ground (by lifting the load 1 by shortening the length L or discharging the said load 1 by being able to vary the length L ) according to a third movement called "lifting movement".

The hoist 2 Above all, it can form a tower crane whose mast 4 the orientation axis (ZZ ') forms, and its boom 5 the distribution axis (XX ') forms.

To simplify the description, it is assumed that such a configuration of a tower crane will follow, it being understood, however, that it is also conceivable to use the same Apply principle of the invention to other hoists, and especially on mobile cranes or cranes with luffing jibs, simply by the respective adaptation of the model.

It can be seen O the intersection of the distribution axis (XX ') and the alignment axis (ZZ').

The distribution axis (XX '), which incidentally also passes through the attachment point H , preferably runs approximately horizontally, and is assumed to be such as to simplify the description, as described below.

m [kg]

ist die Masse der schwebenden Last 1,

M [kg]

ist die Masse des Anschlagpunktes H und genauer des Wagens, der den besagten Anschlagpunkt bildet,

lγ [kg·m²]

ist das Trägheitsmoment des Hebezeugs 2, welches den Anschlagpunkt H trägt, im Verhältnis zur Orientierungsachse (ZZ'),

γ [rad]

ist die Winkelstellung des Anschlagpunkts H um die Orientierungsachse (ZZ'),

x [m]

ist der Abstand vom Anschlagpunkt H zur Orientierungsachse (ZZ'), vorzugsweise gleich der Länge des Segments [OH],

L [m]

ist die Länge des Tragseils 3, das die schwebende Last 1 mit dem Anschlagpunkt H verbindet,

Φ [rad]

ist die radiale Komponente des Pendelwinkels der schwebenden Last 1 in der vertikalen Ebene, die Verteilungsachse (XX') umfassend,

θ [rad]

ist die orthoradiale Komponente des Pendelwinkels der schwebenden Last 1 in der vertikalen Ebene tangential zur Rotationsbewegung des Anschlagpunkts H, das heißt in der Ebene, die senkrecht zur vorherigen verläuft, und vorzugsweise normal zur Verteilungsachse (XX') liegt,

Tγ [N·m]

ist das Motormoment, das angelegt wird, um den Anschlagpunkt H in Rotation um die Orientierungsachse (ZZ') zu versetzen,

Fx [N]

ist die Kraft, die auf den Anschlagpunkt H ausgeübt wird, um diesen entlang der Verteilungsachse (XX') voranzubringen,

To simplify the description, and especially in terms of 1 , the following terms and conventions are also adopted:

m [kg]

is the mass of the floating load 1 .

M [kg]

is the mass of the attachment point H and, more precisely, of the carriage which forms the said attachment point,

lγ [kg · m ² ]

is the moment of inertia of the hoist 2 , which carries the attachment point H , in relation to the orientation axis (ZZ '),

γ [rad]

is the angular position of the attachment point H about the orientation axis (ZZ '),

x [m]

is the distance from the attachment point H to the orientation axis (ZZ '), preferably equal to the length of the segment [OH],

L [m]

is the length of the suspension rope 3 that the floating load 1 connects with the stop point H ,

Φ [rad]

is the radial component of the pendulum angle of the suspended load 1 in the vertical plane, comprising the distribution axis (XX '),

θ [rad]

is the orthoradial component of the pendulum angle of the suspended load 1 in the vertical plane tangential to the rotational movement of the attachment point H , that is to say in the plane which is perpendicular to the previous one, and preferably normal to the distribution axis (XX '),

Tγ [N · m]

is the motor torque which is applied in order to set the abutment point H in rotation about the orientation axis (ZZ '),

Fx [N]

is the force exerted on the attachment point H to advance it along the distribution axis (XX '),

It will be noted that the reference point of projection used to define the "radial" and "orthoradial" components of the pendulum (pendulum movement of the load), and / or that for the movement of the attachment point H, here preferably corresponds to the Frénet reference point which is stated Anvil H is suspended, and its normal (or "radial") vector is preferably carried by the distribution axis (XX ') at all times.

In a manner known per se, the movement of the suspended load 1 preferably controlled by a control target value related to a quantity called "commanded magnitude" and characteristic of the movement of said load 1 is, and which is preferably a kinematic quantity (in the nature of a velocity or acceleration).

This controlled variable can be, for example, the rotational speed V _o and / or the feed rate V _{d of} the attachment point H.

According to the invention, the control method comprises a step (a) of acquiring the set point during which a control target value called the "gross control target value" V _i which corresponds to the value of the commanded variable which the operator of the hoist is in real time is acquired 2 is aimed at the assumed time t .

The gross control setpoint value V _i thus reflects the behavior (normally the vehicle speed) that the crane operator wishes to transmit to the anchor point (trolley) at the presumed time.

More specifically, said method may comprise a step (a) of detecting a set point during which a value for the speed command value V _i (that is, a control target value expressed in the form of a speed command value V _i ) is detected in real time, which is representative for the rotational speed V _o and / or for the feed speed V _d , which is the operator of the hoist 2 wants to transfer the anchor point at the assumed time t .

Hereinafter, it is preferably and for ease of description assumed that the control over the velocity of the impact point H (car) takes place, and one equates for this reason "control setpoint" and "speed setpoint", without this being a limitation of the invention.

By convention, the index i is assigned the value "o" to make reference to the orientation movement (rotation R ) and the value "d" to make reference to the distribution movement (feed T ).

Advantageously, the method according to the invention forms an iterative method which first enables monitoring of the control setpoint, and in particular of the fluctuating and unpredictable speed setpoint in approximately real time, fixed at all times by the operator of the hoist, and thereafter permanently refreshing the calculations and consequently the corresponding filtered setpoint, regardless of the total duration of the necessary path for the transport of the floating load 1 , from their starting point to the destination point.

For this purpose, the method preferably has a relatively short sampling time, which is clearly smaller than the total duration of the path.

The said sampling time is thus preferably less than 100 ms, and for example of the order of 40 ms.

Of course, the speed setpoint value V _i can be set by the operator of the hoist 2 on the basis of all suitable control devices 6 , as fixed to a joystick, which can preferably simultaneously define the target value for the feed rate V _d and the target value for the rotational speed V _o that the operator wishes to convey to the stop point H.

The control method also includes, after the step (a) for the set point detection, a step (b) for target value filtering.

In the course of this step (b) for the setpoint filtering one models, as this is especially true in 2 is shown, the pendulum behavior of the floating load 1 according to the considered motion R , T by a virtual model Ẋ = AX + BU, which works with a state vector X.

According to the invention, this state vector X comprises at least one component "x ₁ " called the "main component", which corresponds to the controlled variable, and other components "x ₃ , x ₄ " called "additional components" which are representative of the kinematic variables characteristic of the pendulum motion of the suspended load are, such as the pendulum angle θ, φ or the angular velocity of oscillation θ, φ. depending on the assumed movement R , T.

More specifically, said state vector X preferably comprises at least one component of the instantaneous velocity x ₁ which is representative of the instantaneous velocity of the impact point H according to the assumed motion (here the rotational velocity component about the orientation axis (ZZ ') γ. = ∂γ / ∂t according to the rotation speed R , or respectively the linear distribution component ẋ = = ∂x / ∂t according to the advancing movement T ), an angular component of the pendulum x ₄ which is representative of the pendulum angle of the suspended load for the assumed motion (here, θ is the orthoradial pendulum component for the rotational motion R , and φ is the radial pendulum component for the advancing movement T ), and an angular velocity component x3 for the pendulum, representative of the angular velocity of the pendulum for the assumed motion (here the orthoradial component θ. = ∂θ / ∂t for the rotational movement and in each case the radial component φ. = ∂φ / ∂t for the feed movement).

In the expression of the model shown above, "A" represents the state matrix (evolution matrix of the system), "X" the state vector, "B" the application matrix of the controller, "U" the vector of the controls (inputs).

The vector "Ẋ" corresponds to the first derivative with respect to the time of the state vector X.

In 2 the letter "p" corresponds to the complex variable used by the Laplace transforms; thus one receives clearly Ẋ = pX

Preferably, the application matrix of the controller B is a column vector and the vector of the controllers U is limited to a matrix of dimension 1 × 1.

In practice, the virtual model Ẋ = AX + BU preferably corresponds to the matrix expression of a system of equations from Newtonian mechanics and allows the description, here as a projection in the vertical radial plane, comprising the distribution axis (XX ') for the feed movement T , and / or as a projection in the vertical orthoradial plane for the rotational motion R , the behavior, and more particularly the motion components of a virtual pendulum system, the characteristics of the suspended load 1 has at the stop point H.

To simplify this modeling, it is preferred to take the "small angles" by taking into account, as a first approximation, that the pendulum amplitude and thus the pendulum angle components are relatively low, thereby making it possible, in particular, to simplify the trigonometric expressions by developments that occur restrict the first order.

In the course of step (b) for filtering one preferably employs a state control U _i = k _i0 · V _i - to K _i · X on the virtual model X = AX + BU, which said next one (for the gross control target value, and more specifically velocity setpoint) V _i representative setpoint _element k _i0 * V _i , which was detected during the detection _step , comprises a correction _element K _i * X corresponding to the product of a correction vector K _i (here a row vector) times the state vector X (here a column vector) equivalent.

The use of a (virtual) state control, which in this case applies proportional feedbacks in the nature of "gains" k _i1 , k _i2, advantageously provides a simulation of the theoretical behavior of the floating mass 1 in response to said regulation, according to the equivalent of a closed-loop (virtual) system in which one can study and select the dynamics, and above all the stability and responsiveness, by choosing the appropriate ones Reinforcement performs a placement of the poles of the transfer function, which corresponds to the development matrix A - B · K of the closed-loop system.

For the definition of the state control one advantageously uses a correction vector K _i , in which at least certain correction gains among the correction _gains k _i1 , k _i2 , which are assigned to the additional components x ₃ , x ₄ , are expressed in dependence on a same predetermined typing parameter Tc. wherein said typing parameter Tc can be freely set by the operator of the hoist to adjust said correction _gains k _i1 , k _i2 .

More preferably, a correction vector K _{i is used} in which the correction _gains k _i1 , k _{i2 associated} respectively with the angular velocity component x ₃ and the pendulum angle component x _{4 of} the state vector X are expressed in response to a same predetermined typing parameter Tc said typing parameter Tc from the operator of the hoist 2 can be adjusted freely to adjust the said correction _gains k _i1 , k _i2 can.

Advantageously, the state feedback, and hence the behavior of the closed-loop system, and more particularly the placement of the poles of the development matrix A-B * K, depends on the selection of said typing parameter Tc by formally defining a parameter (coefficient) for typing Tc into the definition of the gains of the correction _vector k _i1 , k _i2 , which are respectively assigned to the angular velocity component x ₃ and the pendulum angle component x ₄ .

In other words, the method according to the invention generally makes possible a speed setpoint value V _i in accordance with a feature which in itself could already be an invention a state control applied in anticipation of a typing parameter Tc to a virtual model Ẋ = AX + BU, which can be freely set by the operator of the hoist, and an arbitrary change in the placement of the poles of the corresponding development matrix A - B · K, and consequently the dynamics of the filtered system.

The operator of the hoist can thus as part of an adjustment step, the maneuvering of the suspended load 1 preceded, or can be performed in the course of said maneuver, let the typing parameter Tc vary according to his will.

Thus, the operator of the hoist can by a single setting, which can be easily and quickly implemented, the degree of reaction and stability in maneuvering the suspended load 1 which is available to him by the filtering of his speed setpoints V _i according to the inventive method, change and adapt.

More specifically, the operator can choose either the "muted" support mode, which graphically (see 3 ) represents the poles which are relatively far away from the oscillating mode of the imaginary axis, and according to which the filtered speed setpoint moves relatively slowly but particularly stably and without exceeding the speed setpoint value V _i , as shown in FIGS 6 and 7 or, if the muted mode appears too "soft" to it, opt for the "reactive" assist mode, where the poles are closer to the imaginary axis, and the damping coefficient is lower than in the damped mode, so that the filtered Speed setpoint moves faster to the speed setpoint V _i (for example, according to a 5% reaction time, which is lower than that of the damped mode, and as in 4 and 5 each compared to 6 and 7 the case is) but may tolerate a slight overshoot, and / or some damped oscillations of the filtered speed setpoint (see, for example, FIG 5 ).

Of course, it is also conceivable to provide more than two or three setting values, and above all a continuous adjustment range of the typing parameter Tc between a first damped assist mode and a second reactive assist mode, preferably with increasing Tc values, with various intervening support modes ,

It should be noted that the (adjustment) selection of the typing parameter Tc advantageously takes place freely, that is to say arbitrarily, and that the fixation of the said typing parameter Tc depends on the (sole) will of the operator of the hoist, and therefore the said typing parameter Tc, which preferably remains constant between two successive changes by the operator, exerts its own influence on the setting of the correction _gains k _i1 , k _i2 , irrespective of the configuration of the hoist 2 or the speed command values V _i applied by the operator of the hoist.

In particular, the fixation of the typing parameter Tc is preferably separated from the speed setpoint V _i .

The said typing parameter Tc is also separate and independent of the mass m of the suspended load 1 or from that M of the attachment point H , from the moment of inertia of the hoist 2 around the orientation axis (ZZ ') and the length L of the suspension rope 3 ,

Of course, the definition (or selection) of the value of the typing parameter Tc may be made by the operator of the hoist 2 via any suitable selector switch, potentiometer or any suitable electronic or electromechanical programming device.

According to one possibility, the method may be implemented by a computer having a non-volatile memory, preferably reprogrammable by the operator of the hoist or by a service technician, intended to store a plurality of predefined settings of the typing parameter Tc, for example to different drivers and / or different working conditions (especially weather conditions).

In addition, it should be noted that the invention preferably allows the operator of the hoist to perform the typing parameter Tc as many times as necessary, and more specifically several times in succession before and / or during the maneuver of the suspended load 1 to change and adapt.

This availability and access to the setting of the typing parameter Tc, which in this case takes place in real time, offers a high degree of versatility and flexibility in the use of the hoist thus equipped 2 ,

Finally, in the course of step (b) for filtering, a control setpoint is extracted from the virtual model, the "filtered control setpoint". Y, γ, ẋ is called, and the main component x _{1 of} the state vector X, that is the controlled size corresponds.

More specifically, one extracts from the virtual model a filtered speed setpoint Y, γ, ẋ, the component x ₁ corresponds to the instantaneous velocity of the state vector X.

To simplify the description, we will subsequently set the filtered control setpoint Y with a filtered speed setpoint γ., ẋ but this would not be a limitation on the present invention.

This filtered setpoint Y, γ, ẋ (here preferably corresponds to the single coefficient x _{1 of} the vector of the outputs Y), which is represented by the curves with the solid line in 4 to 7 It is now shown that it is applied to the (not shown) drive devices of the anchor point H , which are designed to drive said abutment point by means of assumed movement R , T.

Typically, the filtered control setpoint may correspond to the setpoint feed rate ẋ applied to a first frequency controller that drives a first electric motor designed to position the carriage on the boom 5 to move in the feed, or the target value of the rotational speed γ., which is applied to a second frequency controller which drives a second electric motor designed to operate the boom 5 around the mast 4 to turn.

Preferably, the method uses to model the pendulum behavior of the suspended load 1 a dynamic (virtual) model showing the mass M of the hitch H (and more precisely that of the corresponding wagon) and the mass m of the suspended load 1 brings to the approach.

Advantageously, such a model allows for the approximate but relatively reliable and precise description of the pendulum behavior of the load by means of relatively simple equations for rapid calculation, for which even few physical and electrical resources are required.

The mass M of the attachment point H can advantageously from the manufacturer of the hoist 2 to provide.

The mass m of the suspended load 1 can be measured or estimated by any suitable means and, for example, by measuring the torque to be applied by the lift motor to vertically move said load.

In this specific case, a first approximation may be taken as a fixed average value representative of an "average" floating load 1 or, on the contrary, it is possible to measure the said mass m on a case by case basis at each charge in order to set the model of the filter as accurately as possible relative to this parameter.

mit:

In a particularly preferred case, the dynamic model is defined as follows by a model called the "no coupling" model, and can be expressed as follows: Ẋ = AX + BU Y = CX where "Y" represents the vector of the outputs and "C" represents the observation matrix,
and

With:

As stated above, the equations have been simplified here by the assumption of small pendulum angles.

It is preferably found that the general expression for the model here remains the same for all movements, whether the said model is now applied to the orientation control (rotation R ) or to the distribution (feed T ), where only the coefficients of the state matrix A and of the State vector X of the movement to which the said model is applied.

It should also be noted that it is quite conceivable to apply the method to the control of a single movement (for example, only the distribution), especially if the hoist 2 a loading portal is in the manner of a crane, which has only a linear feed movement of the floating load, and no rotational movement.

In contrast, it is also advantageous to simultaneously control both the distribution movement (feed T ) and the orientation movement (rotation R ) when the hoist offers these two movements, using a same model, and thus with a relatively limited built-in computing power.

For this purpose, it can be stated that the typing parameter Tc can advantageously be the same, that is to say can have an identical value for the application of the distribution model and for the application of the orientation model, which further facilitates the setting of the filtering for the crane operator.

Moreover, it is noted that in the equations of the above model, in a first approximation, it was decided not to consider phenomena of coupling between the axes, that is, the centrifugal acceleration.

In view of this fact, it is quite conceivable to add a link to the coupling that would involve centrifugal acceleration, while maintaining the fiction principle of typification.

mit:

Thus, as an alternative to the previous "uncoupled" model, it is possible to analogously use a model called "coupled," which is expressed as follows: Ẋ = AX + B ₁ u ₁ + B ₂ u ₂ Y = CX where Y represents the vector of the outputs and "C" the observation matrix,
and

With:

Advantageously, irrespective of whether the model with coupling or the model without coupling is considered, the state control is applied to the above-mentioned model: U _i = k _i0 · V _i - K _i · X with Ki = [ki0,0, ki1, ki2] as correction vector (multiplier),
and with the determination i = d for the distribution control, and i = o for the orientation control.

Here too, regardless of whether one accepts the distribution movement (feed) or the orientation movement (rotation), one can mutatis mutandis use the same virtual model type and the same principle of state control with the placement of the poles.

Preferably and to simplify the calculation one takes as in 2 indicates that the gain k _{i0 is the same} for the velocity setpoint V _i and for the state feedback (in the correction vector K _i ).

In a particularly preferred case, and also to simplify the calculation, k _{i0 is assumed} to be 1, that is, the controller has state feedback with a uniform gain for the instantaneous velocity component x ₁ .

Finally, the dynamic filter applied to the speed command value V _i in step (a) of the filtering, and which conforms to the law of control, can be found in 2 is shown for the model without coupling as follows: Ẋ = (A - BK) X + BV _i Y = C _out X with C _out = [1 0 0 0]

For the model with coupling you also get: Ẋ = (A - B ₁ K) X + B ₁ V _i + B ₂ T _γ Y = C _out X with C _out = [1 0 0 0]

The matrix A - BK (or A - B ₁ K) forms the development matrix of the virtual model in a closed loop.

In this property, it should be noted that the application matrices of control "B" of the model without coupling and B _{1 of} the model with coupling are preferably identical (B = B ₁ ), the development matrix A - BK is the same regardless of whether one is the model now considered with or without coupling.

The method according to the invention for the resolution (and for the definition of the filter) can thus be applied without distinction to both models for the sake of advantage.

In the case of the dynamic model proposed above, the said closed-loop virtual model development matrix can be represented as follows:

The correction _gains k _i1 , k _i2 are preferably expressed in the step of the setpoint filtering in dependence on the typing parameter Tc, so that a sub-matrix Ar of dimension 2 × 2 can be extracted from the development matrix A-BK of the closed-loop virtual model said correction _gains k _i1 , k _i2 connect to two additional components of the state vector, these components preferably corresponding respectively to the angular velocity component of oscillation x ₃ and the angular component of oscillation x ₄ , and whose eigenvalues λ on the other hand have a real non-zero portion completely due to the said typing parameter Tc is determined.

With respect to the above model, one can extract here from the closed-loop development matrix the sub-matrix Ar = [(A-BK) i, j], where i = (3, 4) and j = (3, 4) which in practice is sufficient to describe the dynamics of the system:

In fact, the inventors have found that it is possible to adjust at a first approximation in accordance with a feature which in itself may already be a separate invention and whether or not the correction gains are expressed in dependence on a same typing parameter Tc the dynamics of the model with closed-loop state control Control loop, wherein the model of the (complete) development matrix A - BK has, starting from a placement of the poles of a sub-matrix Ar of the said development matrix A - BK make, said sub-matrix Ar is actually a sufficient subsystem to approximately to describe the dynamics of the entire system.

However, such a feature advantageously allows the simplified and accelerated calculation of the eigenvalues, and thus the implementation of the filtering of the velocity setpoint.

In the present case, said sub-system is a non-zero determinant system connecting the correction _gains k _i1 , k _i2 to the two components of the state vector X corresponding respectively to the angular velocity component x ₃ and the pendulum angle component x ₄ .

The validity of the approximation may depend above all on the fact that the above-mentioned correction _gains k _i1 , k _i2 in the complete system corresponding to the development _matrix A - BK have a relatively small or even negligible influence in the determination of the representative component of the instantaneous velocity x ₁ , especially in comparison to the influence of the coefficients of the state matrix A (and more precisely the coefficient a ₁ in the previous example, which in turn depends on the structural features of the hoist 2 and from the mass m of the hovering load 1 depends).

In practice, this approximation can be checked especially as long as the typing parameter Tc remains below a predetermined limit.

In a particularly preferred case, the eigenvalues λ of the sub-matrix Ar have a real section Re (λ) not equal to zero and directly proportional or even equal to the said typing parameter Tc or conversely 1 / Tc to the typing parameter.

Preferably, the real portion Re (λ) of said eigenvalues is itself determined solely by the typing parameter Tc and, for example, exactly equal to the inverse of -1 / Tc of the typing parameter, as in FIG 3 shown.

Thus, one can easily make an immediate positioning of the poles only by adjusting the value of the typing parameter Tc, which is sufficient to characterize the said real section Re (λ).

More specifically, an increasing typing parameter Tc here approximates the imaginary axis poles to each other (by reducing the absolute value of their real portion) and thus gives the filtering a more reactive character.

Conversely, a decreasing typing parameter Tc removes the imaginary axis poles from each other by increasing the absolute value of their real portion, thus giving the filtering a less reactive but more subdued character.

In addition, the sub-matrix Ar and the expression of the correction _gains k _i1 , k _{i2 are} preferably selected such that the imaginary portion of said eigenvalues Im (λ) are independent of the typing parameter Tc.

Thus, for a given material configuration of the hoist, and more specifically for a given constant imaginary section Im (λ), the dynamics of the system may preferably be completely defined by the selection of the value of the typing parameter Tc.

In practice, this imaginary section Im (λ) of the material configuration of the hoist, and especially in the above-mentioned dynamic model of the length L of the suspension rope 3 , from the distance x to the axis of rotation of the attachment point H and from the relationships between the mass M of the attachment point H , the mass m of the suspended load 1 and the moment of inertia lγ of the hoist depend.

wobei i = d für die Verteilungssteuerung, und i = o für die Orientierungssteuerung ist.The selection of the correction _gains (k _i1 , k _i2 ) preferably takes place as follows:

where i = d for the distribution control, and i = o for the orientation control.

sodass sich ihre Eigenwerte wie folgt darstellen:

The sub-matrix Ar can be expressed as follows:

so their eigenvalues are as follows:

One finds here, as indicated above and as in 3 represented a placement of the poles (that is, the eigenvalues that are in 3 are indicated by crosses) whose dynamics and more precisely their attenuation is determined exclusively by the selection of the typing parameters Tc.

hat, vorzugsweise unabhängig vom Typisierungsparameter ist.For the real section Re (λ) of the said poles (eigenvalues) has the value 1 / Tc here, while the imaginary section Im (λ) of said poles (eigenvalues), here the value

has, is preferably independent of Typisierungsparameter.

Advantageously, said imaginary section here is constant for a given material configuration of the hoist, assuming a constant value for the length L of the hoisting rope, in the case of control according to the advancing movement T , or if constant values for the length L of the suspending rope and assumes for the distance x from the stop point H to the axis of rotation, in the case of the control by means of rotational movement R.

It should also be noted that the method, and more particularly the proposed models, advantageously allow the length L of the support rope 3 to take into account that the floating load 1 connects with the stop point H.

In particular, the correction _gains k _i1 , k _i2 preferably also _depend on the length L of the support cable 3 expressed, that the floating load 1 connects with the stop point H.

In this way, the correction _gains k _i1 , k _{i2 of} the state control, and consequently the filtering of the speed setpoint depending on the variations in the length L of the support rope 3 vary in real time.

In the above example, the length L is included in the calculation of the gains by the coefficient b ₂ .

Thus, one can permanently change the virtual model to the actual instantaneous configuration of the hoist 2 and thus obtain a modeling of the behavior of said hoist and the hovering load 1 reliably and accurately reflected at any given moment and regardless of the lifting height of said load.

In this way one increases the precision and reliability of the control by filtering the speed setpoint.

The estimation of the length L of the support rope can be, for example, from the count of the performed revolution of a winch for rolling up / unrolling the suspension rope 3 result, which is driven by the lifting motor.

Incidentally, it is noted that the intentional manual adjustment of the dynamics of the filtering by selecting the value of the typing parameter Tc advantageously differs from and is uncorrelated with the automatic setting which consists of the model, and more specifically one or the other Coefficients of the state matrix A to the current material configuration of the hoist 2 , and more precisely to the length L of the suspension rope 3 adapt.

In other words, the typing parameter Tc thus makes it possible to set the type of desired dynamic behavior independent of the material configuration of the hoist, and more specifically, given a given material configuration (weights, inertia, lifting height), the dynamics of the system freely from a plurality of select available modes.

It should also be noted that the correction _gains k _i1 , k _i2, as well as the poles characterizing the dynamics of the modeled system, are advantageously defined in explicit and deterministic form by formulas which depend exclusively on the one hand on the typing parameter Tc and on the other hand on input data in Connection with the material configuration of the system (mass m of the suspended load, mass M of the wagon, length L of the suspension rope, inertia modulus lγ of the hoist, distance x to the axis of rotation of the attachment point).

In this way, the correction _gains k _i1 , k _i2 as well as the poles can be computed directly and promptly based on these data without being forced to predetermine, store in memory and periodically interrogate the cartographies (abaki or databases) assign different correction gain values, which are adapted to each loading situation and / or each spatial configuration of the hoist different predictive operating situations of the hoist, for example in the form of point clouds.

Again, the method used to facilitate the data storage capacities that are necessary to control the hoist.

As an indication, the typing parameter Tc may be selected from a range between 0.2 and 2, preferably between 0.3 ( 6 and 7 ) and 1.8 ( 4 and 5 ) lies.

In the present case, the low Tc values correspond to a "damped" assist mode, relatively slow and stable, and the high Tc values to a "reactive" assist mode, faster and somewhat less stable than the damped mode.

Preferably, the method comprises an open-loop control step consisting in passing the filtered control setpoint (filtered speed setpoint) Y, ie x1 = γ. for the rotation R, and x1 = x. for the feed T , to a control system of the hoist 2 to apply in the open loop, that is to say to a control system which controls the said filtered set point (speed) Y, γ, x. to the drive devices of the attachment point H , which are designed to move the attachment point H in an assumed movement R , T , without the measured or calculated return for the angle, or the angular velocity, of the effective oscillation of the real levitated load, more preferably Use the measured or calculated feedback for the effective velocity of the real impact point movement.

In other words, the invention enables the virtual definition of a speed setpoint x ₁ in the filtering that allows the virtual system to satisfy the speed and stability criteria set by the selection of the typing coefficient Tc, and that speed setpoint x ₁ from a virtual modeling to the real hoist 2 as an effective, filtered speed reference, and this in open-loop, ie blind, no control, which aims to combat thereafter the real commutation that may result from the application of this filtered setpoint, or that from external ones Disorders emerged.

For this purpose, it is noted that any real commutation applied to the real hoist as a consequence of the application of the filtered set point would in any case be intrinsically reduced, if only for the reason that the said filtered set point is worked out in the same way, that the occurrence of such commuting is minimized or even prevented.

In any case, the invention so advantageously allows the simplification of the structure of the hoist 2 in particular, since it is not necessary to provide sensors (still corresponding wirings) intended for measuring and monitoring the real pendulum values.

Moreover, this limits the number of measurements and information processing to be performed, thereby facilitating the calculations, and consequently the size and power consumption of the built-in electronic control device on the hoist 2 can be reduced.

Of course, the reliability of such an open-loop application assumes that the real system is the hovering load on the hoist 2 has a behavior close to that of the model, which is the case here.

As such, the invention also relates to a computer or a data carrier that can be read by a computer that receives or contains code elements of a computer program that allows the implementation of a driving method according to the invention, when said code elements are read by said computer.

Finally, the invention relates as such to a hoist 2 like a tower crane, a mast 4 comprising, extending approximately over a first vertical axis, the orientation axis (ZZ ') is called, a boom 5 which intersects said mast which extends over a second axis, called the distribution axis (XX '), and which carries an abutment point H against which a hovering load 1 can be hung, as well as driving devices with which said abutment point H in a rotational movement R about the orientation axis (ZZ ') and in a feed movement T along the distribution axis (XX') can be added.

According to the invention, the drive device (s) of said hoist are 2 which are associated with at least one of the two movements for rotation R and feed T , and preferably each of these two movements, are controlled by a control device provided with calculating and programming means which are designed to operate in real time by a method according to the invention Gross control setpoint (speed setpoint) V _i to be filtered, defined by the operator of the hoist, and to apply the resulting filtered control setpoint (filtered speed setpoint) Y, γ, x. on the corresponding drive device.

Of course, the invention can in no way be limited only to the variants described, since above all the skilled person is able to isolate one or the other features mentioned above or combine freely with each other, or to replace them with equivalent.

In particular, it would be quite possible to use any other suitable model and / or spatial coordinate system, even though the number of components of the state vector X and / or the formulation of the correction gains must be adjusted accordingly depending on the typing parameter Tc.

Preferably, the invention enables the commutation of a suspended load 1 when moving the latter in a reliable and relatively simple manner, according to one or two separate or combined movements R , T , thereby imparting to the hoist an adjustable, safe and predictable behavior, which depends on the personal typing Tc of the operator himself defined, provides a faithful and comfortable sense of control.

For the sake of convenience, the simplicity of the method and model used allows one to obtain a sufficiently precise estimate of the pendulum behavior of the suspended load, on the one hand, to somewhat reduce commuting, and, on the other, to maintain the desired responsiveness to the velocity setpoint fluctuations, without resorting to heavy computations involving complex algorithms which would aim to obtain an optimized speed and / or path.

QUOTES INCLUDE IN THE DESCRIPTION

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

Cited patent literature

• FR 2704847 [0003]

Claims (11)

Method for controlling the movement of a suspended load ( 1 ) at an attachment point (H) of a hoist ( 2 ), wherein said abutment point (H) is designed to be referred to by yawing (R) about a first vertical axis (ZZ '), the "orientation axis", and / or according to a feed movement (T) along a second axis ( XX '), called the "distribution axis", and said vertical axis (ZZ') intersects to move, the movement of the suspended load ( 1 ) is controlled by a control setpoint, which is based on a characteristic quantity for the movement of the said load ( 1 ), which is called "controlled quantity", such as the rotational speed (V _o ) and / or the feed rate (V _d ) of the impact point (H), said method comprising: a step (a) for detecting the Command value during which a control target value is recorded in real time, called the "gross control target value" (V _i ), which corresponds to the value of the commanded variable which the operator of the hoist ( 2 ) at the assumed point in time (t), then a step (b) for filtering the setpoint, in the course of which: - the pendulum behavior of the suspended load ( 1 ) is modeled according to the assumed motion (R, T) by a virtual model (Ẋ = AX + BU) using a state vector (X) comprising at least one component called "principal component" (x ₁ ) corresponding to the commanded magnitude , as well as other "accessory components" (x ₃ , x ₄ ) components representative of kinematic variables characteristic of the pendulum motion of the suspended load, such as the pendulum angle (θ, φ) or the pendulum angular velocity (θ., φ.) depending on the assumed movement, - one _applies a state control (U _i = k _i0 · V _{i -K i} · X) to the virtual model (Ẋ = AX + BU), which in addition to a representative for the gross control setpoint value _element (k _i0 · V _i), which has been detected within the detecting step comprises a correction element (K _i · X) corresponding to (the product of a correction vector K _i) times (the state vector X), (using a correction vector K _i), wherein the at least determined correction _gains (k _i1 , k _i2 ) associated with the additional components are expressed in dependence on a same predetermined typing parameter (Tc), said typing parameter (Tc) being freely adjustable by the operator of the hoist to be able to adjust said correction _gains (k _i1 , k _i2 ), and to extract from the virtual model a control setpoint, the "filtered control setpoint" (Y, γ., X.) which corresponds to the main component (x ₁ ) of the state vector (X). A method according to claim 1, characterized in that: in the course of step (a) for detecting the target value in real time a speed command value (V _i ) is detected, representative of the rotational speed (V _o ) and / or for the feed rate (V _d ) the operator of the hoist ( 2 ) at the point in time (t) to be transmitted to the attachment point (H), and then during step (b) of the target value filtering: - the pendulum behavior of the suspended load ( 1 ) is modeled according to the assumed motion (R, T) by a virtual model (Ẋ = AX + BU) which uses a state vector (X) comprising at least one component of the instantaneous velocity (x ₁ ) representative of the instantaneous velocity (γ., x.) of the impact point (H) according to the assumed motion, a pendulum angle component (x ₄ ) representative of the pendulum angle (θ, φ) of the suspended load ( 1 ) according to the assumed motion, and an angular velocity component (x ₃ ) for the oscillation representative of the angular velocity for the oscillation (θ., φ.) according to the assumed movement - to a state control (U _i = k _i0 · V _i - K _i · X) applying to said virtual model (X = AX + BU), the representative beside a for the speed setpoint (V _i) Setpoint member (k _i0 · V _i ), which was detected in the detection _step , a correction term (K _i · X) corresponding to the product of a correction vector (K _i ) times the state vector (X), using a correction vector (K _i ) in which the correction _gains (k _i1 , k _i2 ) respectively associated with the angular velocity component (x ₃ ) and the pendulum angle component (x ₄ ) of the state vector (X) are expressed in response to a same predetermined typing parameter (Tc) and said typing parameter (Tc) can be freely set by the hoist operator to set said correction _gains (k _i1 , k _i2 ). - From the virtual model, a filtered speed setpoint (Y, γ., X.) which corresponds to the component (x ₁ ) of the instantaneous velocity of the state vector (X). Method according to claim 1 or 2, characterized in that the correction _gains (k _i1 , k _i2 ) are expressed in the step of the setpoint filtering as a function of the typing parameter (Tc), such that one derives from the development matrix (A - BK) of the closed-loop virtual model can extract a sub-matrix (Ar) of dimension 2 × 2, which on the one hand, the said correction _gains (k _i1 , k _i2 ) with two additional components of the state vector, additional components which preferably respectively correspond to the pendulum angular velocity component (x ₃ ) and the pendulum angle component (x ₄ ), and whose eigenvalues (λ) on the other hand have a real non-zero portion completely due to the said typing parameter (Tc) is determined. Method according to Claim 3, characterized in that the eigenvalues (λ) of the sub-matrix (Ar) have a real section (Re (λ)) not equal to zero and directly proportional to or even equal to said typing parameter (Tc) or vice versa (1 / Tc ) to the typing parameter. Method according to one of the preceding claims, characterized in that the correction _gains (k _i1 , k _i2 ) also _depend on the length (L) of the supporting cable ( 3 ) expressing the suspended load ( 1 ) connects to the stop point (H). Method according to one of the preceding claims, characterized in that it uses a dynamic model which measures the mass (M) of the impact point (H) and that (m) of the suspended load (M). 1 ), wherein said dynamic model is defined as follows: either by a model called "uncoupled", expressed by:

With:

or by a model called 'coupled', expressed by:

With:

Models in which: m [kg] is the mass of the suspended load ( 1 ), M [kg] is the mass of the lifting point (H), lγ [kg · m ² ] the moment of inertia of the hoist ( 2 ), which carries the attachment point (H), in relation to the orientation axis (ZZ '), γ [rad] is the angular position of the attachment point (H) about the orientation axis (ZZ'), x [m] is the distance from the attachment point (H ) to the orientation axis (ZZ '), L [m] the length of the support cable ( 3 ), which is the floating load ( 1 ) with the attachment point (H), Φ [rad] the radial component of the pendulum angle of the suspended load ( 1 ) in the vertical plane comprising the distribution axis (XX '), θ [rad] is the orthoradial component of the pendulum angle of the levitated load ( 1 ) in the vertical plane is tangential to the rotational movement of the abutment point (H), Tγ [N · m] is the motor torque which is applied to set the abutment point (H) in rotation about the orientation axis (ZZ '), Fx [N ] is the force exerted on the attachment point (H) in order to advance it along the distribution axis (XX '),

and by applying the state control to the assumed model: U _i = k _i0 · V _i - K _i X with K _i = [k _i0 , 0, k _i1 , k _i2 ], with the determination i = d for the distribution control, and i = o for the orientation control. Method according to Claim 6, characterized in that the selection of the correction _gains (k _i1 , k _i2 ) is effected as follows:

where i = d for the distribution control, and i = o for the orientation control. Method according to one of the preceding claims, characterized in that the typing parameter (Tc) is selected from a range which is between 0.2 and 2, preferably between 0.3 and 1.8. Method according to one of the preceding claims, characterized in that it comprises an open-loop control step consisting in passing the filtered control setpoint (Y, γ., X.) to apply to a control system of the hoist in the open loop, that is to a control system, the said control setpoint (Y, γ., X.) applies to abutment point (H) propulsion devices designed to move the abutment point in the assumed motion (R, T), without the measured or calculated return for the angle, or the angular velocity, of the effective oscillation of the real levitated load; it is still preferable to use the measured or calculated feedback for the effective velocity of the real impact point movement. Computer or data carrier that can be read by a computer, characterized in that it contains code elements of a computer program that allows the implementation of a driving method according to the invention, when said code elements are read by said computer. Hoist ( 2 ), in the manner of a tower crane, a mast ( 4 ), which extends approximately over a first vertical axis, the orientation axis (ZZ ') is called, a boom ( 5 ) which intersects said mast which extends over a second axis, called the distribution axis (XX '), and which carries an abutment point (H) to which a suspended load ( 1 ) and drive means for translating said abutment point (H) into a rotational movement (R) about the orientation axis (ZZ ') and an advancing movement (T) along the distribution axis (XX') Hoist characterized in that the drive device (s) of said hoist, which are assigned to at least one of the two movements for rotation (R) and feed (T), and preferably each of these two movements, are controlled by a control device with rake and programming means designed to filter, in real time by the method of any one of claims 1 to 9, a gross control setpoint (V _i ) defined by the operator of the hoist and to apply the resulting filtered control setpoint (Y, γ., X.) on the corresponding drive device.

Images