The specific embodiment
Below in conjunction with accompanying drawing preferred implementation of the present invention is further described.
In below describing, easy for describing, when mentioning " angle of jib " or " angle ", be meant the jib of hoisting crane and the angle between the horizontal surface.When mentioning " angle of suspender swing ", depart from the wherein maximum angle of vertical line position when being meant the swing of operation object and/or suspender.
Find in the practical operation, the operation object is swung degree when just being sling by crane hanger 2 bigger, if can be during this time the swing of suspender 2 and/or operation object be reduced to a certain degree, then after process in, the swing degree of suspender 2 and/or operation object can be very not big.Therefore needs during this time are reduced to the swing of operation object to a certain degree.In addition, when the operation object lands, because jib 1 no longer bears weight (not comprising suspender 2 own wts here), the amount of deflection vanishing of jib 1, because therefore the inertia suspension hook will, be necessary to reduce the swing of suspender 2 to the swing of crane cab direction when the operation object lands.
Therefore, as total inventive concept, as shown in Figure 1, according to an embodiment of the invention, provide a kind of hoisting crane anti-sway system, wherein, this system comprises:
Strained detection device 20 is used to detect the stressed size of jib 1; Controller 10 is used to obtain strained detection device 20 detected stressed sizes, judges that according to this stressed size detecting hoisting crane is in stress state or unloaded state, according to the angle between judged result control luffing mechanism 40 adjusting jibs 1 and the horizontal surface.
This anti-sway system can also comprise luffing mechanism 40, is used to regulate the angle between described jib 1 and the horizontal surface.
Described strained detection device 20 can for example detect stressed size by detecting jib 1 top suspender 2 and operation object (if loading) to the gravity that jib 1 top (generally passing through steel rope) applies, and the signal that reflects this stressed size is sent to controller 10.Strained detection device 20 also can be determined the stressed size of jib 1 by the jib 1 that detection luffing mechanism 40 is born to its applied pressure.The mode that described signal sends can be wire communication or radio communication.
It is the angle that control luffing mechanism 40 increases between jibs 1 and the horizontal surface under the situation of stress state that described controller 10 is used in described judged result, and is that control luffing mechanism 40 reduces angle between jib 1 and the horizontal surface under the situation of unloaded state in judged result.
Described controller 10 judges that according to stressed size the method for stress state can for example be that described controller 10 obtains stressed size in real time, if the stressed size that controller 10 obtains becomes the unmodified state by the state that increases always, then controller 10 is thought and is in stress state from the moment hoisting crane of this transformation.Specifically, controller 10 can obtain described stressed size at regular intervals, and the stressed size that the stressed size that current time is obtained and previous moment are obtained compares, if the stressed size that the stressed size that current time obtains is obtained greater than previous moment, then controller 10 begins to enter monitor state.In general,, can set a compare threshold,, then trigger controller 10 and enter monitor state if the stressed size that the stressed size that above-mentioned current time obtains is obtained greater than previous moment surpasses this threshold value according to actual conditions.After controller 10 entered monitor state, controller 10 continued to obtain stressed size, carved the stressed size of obtaining when controller 10 at a time and compared basic identically with the stressed size that the previous moment obtains, and then controller 10 judges that hoisting cranes are in stress state.
Described controller 10 judges that according to stressed size the method for unloaded state can for example be if the stressed size that described controller (10) obtains is sustained state by the state-transition that reduces always, and then controller (10) judgement is in unloaded state from the moment hoisting crane of this transformation.
Similar with the method for judging stress state, difference is if the stressed size that the stressed size that current time obtains is obtained less than previous moment, and then controller 10 begins to enter monitor state.In this case, also can set a compare threshold,, then trigger controller 10 and enter monitor state if the stressed size that the stressed size that above-mentioned current time obtains is obtained less than previous moment surpasses this threshold value.
Replacedly, controller 10 can calculate the rate of change of stressed size, i.e. the variation of stressed size in the unit time, when this rate of change by on the occasion of becoming 0 (or near 0), then controller 10 judges that hoisting cranes enter stress state; When this rate of change becomes 0 (or near 0) by negative value, then controller 10 judges that hoisting cranes are in unloaded state.
Those skilled in the art are envisioned that and can also adopt other to judge the method for stress state and unloaded state.For example, can adopt simpler mode,, just can judge to be in stress state as long as detected stressed size increases (for example, having increased a certain value) suddenly; As long as detected stressed size reduces (for example, having reduced a certain value) suddenly, just can judge to be in unloaded state.
Here and hereinafter described stress state be meant state when hoisting crane is just lifted from initial position with the operation object, described unloaded state is meant the state when hoisting crane just is discharged into destination locations with the operation object.
The contriver finds, when hoisting crane is in stress state, when the angle of the principal arm of jib 1 and horizontal surface when 60 degree are above, the angle that increases between jib 1 (principal arm) and the horizontal surface can effectively reduce the degree that suspender 2 is swung.Therefore, according to an embodiment of the invention, it is the angle that control luffing mechanism 40 increases between jibs 1 and the horizontal surface under the situation of stress state that controller 10 is used in described judged result.
The contriver also finds, when hoisting crane is in unloaded state, when the angle of the principal arm of jib 1 and horizontal surface when 60 degree are above, reduce the degree that angle between jib 1 (principal arm) and the horizontal surface can effectively reduce suspender 2 swings.Therefore, according to an embodiment of the invention, it is that control luffing mechanism 40 reduces angle between jib 1 and the horizontal surface under the situation of unloaded state that described controller 10 is used in judged result.
Above-mentioned controller 10 control luffing mechanisms 40 increase or reduce the mode that angle between jib 1 and the horizontal surface can adopt open loop.In open loop control mode, can set up mapping between the adjusting angle of the angle of suspender 2 and/or the swing of operation object and jib 1 by limited number of time experiment.For example, when stress state, can detect the angle of suspender 2 and/or operation object swing, control the angle that luffing mechanism 40 increases jib 1 then.Under the identical situation of pendulum angle, the increase angle that repeated attempt is different can determine that finally pendulum angle reduces degree cooresponding increase angle when maximum, so just sets up the mapping between this pendulum angle and the increase angle.By under different pendulum angles, carrying out said method, just can determine the mapping between the adjusting angle of the angle of when stress state, swinging and jib 1.When unloaded state, also can determine described mapping according to this method.
Therefore, according to an embodiment of the invention, described system also comprises angle detector, is used to detect the angle of suspender 2 swings of hoisting crane.
Controller 10 is used to obtain the angle that described angle detector detects, and controls the angle that described luffing mechanism 40 is regulated between described jib 1 and the horizontal surface according to this angle.
Generally speaking, need repeatedly regulate that pendulum angle is in the tolerance to jib 1, controller 10 can periodically be controlled the angle that luffing mechanism 40 is regulated between jib 1 and the horizontal surface.Therefore, can set a threshold value, this threshold range can for example be-0.5 ° to 0.5 °.When detected angle during less than this threshold value, controller 10 stops to control the angle that luffing mechanism 40 is regulated between jibs 1 and the horizontal surface.Here the scope in the cycle of controller 10 control luffing mechanisms 40 can for example be 200ms-600ms.
In replaceable embodiment of the present invention, can set up the mapping between the adjusting angle of the cireular frequency of suspender 2 swing and jib 1.The method for building up of the mapping between the method for building up of this mapping and pendulum angle and the adjusting angle is similar, repeats no more.
In this case, described system can comprise angular velocity detector, be used to detect the cireular frequency of suspender 2 swings, controller 10 is used to obtain the cireular frequency that angular velocity detector detects, according to the angle between described luffing mechanism 40 described jibs 1 of adjusting of this cireular frequency control control and the horizontal surface.Similarly, also can set a threshold value in this case, when detected cireular frequency during less than this threshold value, controller 10 is no longer controlled the angle that described luffing mechanism 40 is regulated between described jibs 1 and the horizontal surface.
Strained detection device 20 can be can the stressed device of inspected object, for example pressure sensor.In a preferred embodiment of the present invention, this strained detection device can be a limiter of moment 22.Described limiter of moment 22 is for well known to a person skilled in the art device.The function of limiter of moment 22 is a lot, except can detecting stressed size, can also detect hoisting crane automatically and hang the weight of carrying the operation object, the length and the residing angle (being the angle of jib 1 and horizontal surface) of jib 1, and can send the signal of these physical quantitys of reflection.
The factor that influences crane hanger 2 swing degree has a lot, the for example angle of the weight of the angle between the length of jib 1, jib 1 and the horizontal surface, operation object, suspender 2 swings and cireular frequency etc. can be taken all factors into consideration the factor of these aspects when therefore being desirably in the angle that control luffing mechanism 40 regulates between jibs 1 and the horizontal surface.
Fig. 2 shows the block diagram according to the hoisting crane anti-sway system of an embodiment of the invention.As shown in Figure 2, according to an embodiment of the invention, described strained detection device is a limiter of moment 22, is used to detect the length of the jib 1 of hoisting crane, the weight that detects the angle between this jib 1 and the horizontal surface and detect the operation object that the suspender 2 of hoisting crane slings; And send corresponding length signals, angle signal and weight signal;
Described system also comprises angle and angular velocity detector 30, is used to detect the angle and the cireular frequency of operation object and/or suspender 2 swings, and sends corresponding angle signal and angular velocity signal;
Controller 10 also is used to receive described length signals, angle signal, weight signal, angle signal and angular velocity signal, determines the expectation variable quantity of the angle between jib 1 and the horizontal surface according to these signals;
Luffing mechanism 40 is used at the angle between described jib 1 of adjusting and the horizontal surface under the control of controller 10 so that this angle has changed this expectation variable quantity.
Controller 10 can be determined described expectation variable quantity based on the various control model, for example BP neural network model, fuzzy neural network model, LQR Optimal Control Model, nonlinear PID controller model.Under the preferable case, controller 10 can adopt fuzzy neural network model to control, and this will be discussed in more detail below.
After having determined this expectation variable quantity, controller 10 generates the control signal that is associated with this expectation variable quantity, and this control signal is exported to luffing mechanism 40.The size of this control signal is big or small corresponding with the expectation variable quantity.Control signal transmission manner between described luffing mechanism 40 and the controller 10 can be wire transmission or transmission over radio.
In general, if the swing of suspender 2 can't be reduced to satisfied degree by the angle that once changes between jib 1 and the horizontal surface, another embodiment then according to the present invention, fuzzy neural network model in the described controller 10 can periodically be determined the expectation variable quantity of the value of the angle between jib 1 and the horizontal surface, generates control signal corresponding.Luffing mechanism 40 is under the control of controller 10, and the angle in each cycle between adjusting jib 1 and the horizontal surface is so that this angle has changed the expectation variable quantity in this cycle.Promptly regulate angle between jib and the horizontal surface, to reduce the swing of operation object and/or suspender better through the size of the control signal that repeatedly changes output.When the angle signal that is input to controller 10 was positioned at the tolerance of setting, fuzzy neural network model is the regeneration control signal not.The scope in described cycle can for example be 200ms-600ms.Tolerance can be set flexibly according to hoisting crane size, the anti-effect height etc. of shaking, and in general, tolerance can be set at suspender and wave angle in ± 0.5 degree, and perhaps cireular frequency is lower than 0.2 degree/s.
In loading procedure, if the weight ratio suspender 2 of operation object from great a lot, then can not consider the deadweight of suspender 2.But, also consider the deadweight of suspender 2 for more accurate.In addition, in uninstall process,, at this moment must consider the deadweight of suspender 2 because the operation object is released.Therefore, according to an embodiment of the invention, the weight signal that is input to controller 10 has also added the compensation value of reflection suspender 2 weight.
Angle and angular velocity detector 30 can be the combinations of angular transducer and angular velocity sensor, or have the separate part of detection angles and angular acceleration function.For example, in an embodiment of the invention, this angle and angular velocity detector 30 can comprise light-emitting device, light reflecting device and optical pickup apparatus.This light-emitting device can for example be the infrared light feedway, is positioned at the top end of jib 1; Light reflecting device can be positioned at suspender 2 (for example suspension hook) to be located, optical pickup apparatus can be positioned at light-emitting device near.The light-emitting device vertical-horizontal is towards the direction emission light beam (for example infrared light) of suspender 2, and to optical pickup apparatus, optical pickup apparatus receives this light beam to light reflecting device with beam reflection.When suspender 2 swings, light reflecting device also can be along with suspender 2 be swung together, and like this, the light beam that reflexes to optical pickup apparatus can move on optical pickup apparatus.Can obtain angle and the cireular frequency that the operation object is swung by miles of relative movement and the speed of calculating the folded light beam on the optical pickup apparatus.For instance, the light sensation unit at interval can be arranged and have each other to optical pickup apparatus for a plurality of linear patterns, and the light sensation unit of midway location can be positioned on the midperpendicalar of suspender 2.Interval between the light sensation unit can equate also can be unequal.When reflected light shone this light sensation unit, this light sensation unit can produce signal.Angle and angular velocity detector 30 can also comprise treater, this treater is used to receive that the light sensation unit sends about the miles of relative movement of light beam and the signal of moving velocity, according to miles of relative movement and the moving velocity of this calculated signals light beam on optical pickup apparatus, determine the angle and the cireular frequency of suspender 2 swings thus respectively.Specifically, by calculating the angle of calculating suspender 2 swings in the light sensation unit that sends signal from midway location distance farthest; Calculate the cireular frequency that suspender 2 is swung by calculating the time difference that sends signal between two light sensation unit.More specifically, treater can be encoded to the light sensation unit, because adjacent two light sensation unit is predetermined at interval, so each light sensation unit is determined far from the distance of midway location.Corresponding distance of each coding like this.When one of them light sensation unit by light beam irradiates and when controller sent signal, treater can be known the distance of this light sensation unit far from midway location.The light sensation unit that treater can for example calculate midway location sends in signal and the light sensation unit that sends signal to treater the time difference that sends from midway location light sensation unit farthest between the signal, can determine cireular frequency that suspender 2 swing far from the distance of midway location with this time difference according to this light sensation unit farthest again.Certainly, treater can select any two light sensation unit to come computing time poor, or repeatedly select and calculate respectively after ask average.Replacedly, the function of calculating angle and cireular frequency can be carried out by controller 10.
Replacedly, angle and angular velocity detector 30 can also for example be GPS difference testers.
In replaceable embodiment of the present invention, can not adopt limiter of moment 22, and replace limiter of moment 22 with linear transducer, angular transducer, three independent devices of pressure sensor.Described linear transducer, angular transducer and pressure sensor can be to well known to a person skilled in the art device, no longer describe in detail here.
In replaceable embodiment of the present invention, can not adopt controller 10, and realize the function of controller 10 with limiter of moment 22.
In the preferred implementation that adopts fuzzy neural network model, described controller 10 is used to receive above-mentioned length signals, angle signal, weight signal, angle signal and angular velocity signal, these signals are input to fuzzy neural network model in this controller 10, this fuzzy neural network model is determined the expectation variable quantity of jib 1 and the angle of horizontal surface to generate the control signal that is associated with this expectation variable quantity according to these signals.Control signal is generally voltage signal or current signal, for example pwm signal.If desired, can to this voltage signal carry out D/A conversion (in controller 10, do not have D/A converter, its output be digital signal, and luffing mechanism 40 can only receive the situation of analog signal).Here the control signal of Sheng Chenging can only determine the size that the 40 pairs of jibs of luffing mechanism 1 and the angle of horizontal surface are regulated, and does not determine to increase angle or reduce angle, and this control signal can become the scalar control signal.Replacedly, control signal can determine that the size of angular adjustment can determine again to increase angle or reduce angle, and this control signal becomes control signal vector.In the situation of scalar control signal, be that to decide luffing mechanism 40 by stress state or unloaded state be to increase or reduce angle.For example, when controller 10 judgements were in stress state, then the control signal of Sheng Chenging made luffing mechanism 40 increase the angle of jibs 1 and horizontal surface.When controller 10 judgements were in unloaded state, the control signal of generation made luffing mechanism 40 reduce the angle of jib 1 and horizontal surface.For the situation of control signal vector, the control signal of these two kinds different state decision generations has different directions, for example the control signal of positively charged pressure and the control signal of negative voltage.Below relating in the description of control signal mainly is at the scalar control signal.
Controller 10 can include but not limited to micro controller system, computing machine, microprocessor, DSP, PLC controller 10, field programmable gate array (FPGA) etc.Fig. 3 shows a kind of controller 10 of example.As shown in Figure 3, described controller 10 can comprise: signal receiving unit 12, signal processing unit 14 and signal output unit 16.Signal receiving unit 12 is used to receive above-mentioned various signal, if needed, carries out the A/D conversion to the received signal.Signal processing unit 14 is embedded with fuzzy neural network model, is used for as above-mentioned embodiment describedly, generates control signal according to the signal of input.The control signal that signal output unit 16 is used for generating outputs to luffing mechanism 40, and if needs are arranged, can before output, carry out D/A and change control signal.
Described luffing mechanism 40 can have multiple way of realization.In general, divide from structure, luffing mechanism 40 can be divided into two kinds, and a kind of is by hydraulic efficiency pressure system, and (or drawing) crane arm support that utilizes the piston rod of hydraulic ram to promote 1 changes the angle of itself and horizontal surface; Another kind is to utilize motor drives reel 52, and scrolling steel rope 54 pulling/release jibs 1 change the angle of itself and horizontal surface.
Fig. 4 is the scheme drawing that adopts the hoisting crane of first kind of luffing mechanism.In this case, according to an embodiment of the invention, controller 10 is directly actuated to the valve in the hydraulic efficiency pressure system that likes this luffing mechanism, for example electromagnetic switch apportioning valve.
Fig. 5 is a kind of example hydraulic efficiency pressure system of the luffing mechanism 40 shown in Fig. 4.As shown in Figure 5, hydraulic efficiency pressure system can comprise hydraulic actuating cylinder 41, balance cock 42, electromagnetic switch apportioning valve 43, by pass valve 44, Hydraulic Pump 45, fuel tank 46 and oil circuit.Oil circuit can comprise oil-feed loop 47 and go out oil return line 48.In hydraulic efficiency pressure system shown in Figure 5, controller 10 can direct control electromagnetic switch apportioning valve 43, come the hydraulic oil that enters/go out in the modulated pressure cylinder 41 by the ON/OFF of control electromagnetic switch apportioning valve 43, thereby the control plunger bar is flexible to reach the purpose of regulating jib 1 and horizontal plane angle, thus, can reduce the swing of operation object.In this case, controller 10 can be connected with the control end of electromagnetic switch apportioning valve 43.When controller 10 judgements are in stress state, the control signal that is generated by fuzzy neural network model that outputs to electromagnetic switch apportioning valve 43 makes 43 actions of electromagnetic switch apportioning valve, hydraulic oil in the fuel tank 46 is injected into the latter half (part that the piston shown in Fig. 5 is following) of hydraulic actuating cylinder 41 by oil-feed loop 47, hydraulic oil in the first half of hydraulic actuating cylinder 41 (part that the piston shown in Fig. 5 is above) flows to fuel tank 46 by going out oil return line 48 simultaneously, piston rod stretches out thus, promotes jib 1 angle of jib 1 is increased.Otherwise when controller 10 judgements were in unloaded state, the control signal that outputs to electromagnetic switch apportioning valve 43 made electromagnetic switch apportioning valve 43 do opposite action, made the piston rod retraction, and the angle of jib 1 reduces thus.Certainly, except control electromagnetic valve, controller 10 can be controlled other valve.In replaceable embodiment of the present invention, controller 10 can control example such as electro-hydraulic proportional valve, electrohydraulic digital valve etc.
Fig. 6 is the scheme drawing that adopts the hoisting crane of second kind of luffing mechanism.This luffing mechanism generally can comprise motor, reel 52, steel rope 54 and assembly pulley.Motor is used to drive reel 52 and rotates, and is wound with steel rope 54 on the reel 52, and steel rope 54 is connected with crane arm support 1 by assembly pulley, and when driven by motor reel 52 rotated, reel 52 was rolled or discharged steel rope 54, changed the angle of jib 1 and horizontal surface thus.In this case, controller 10 can be connected with the control end of motor.When controller 10 judgements were in stress state, the control signal of exporting to the control end of motor was just changeing to drive reel 52 motor and rolls steel rope 54, and the angle of jib 1 increases thus.When controller 10 judgements were in unloaded state, the control signal of exporting to the control end of motor made the motor counter-rotating discharge steel ropes 54 to drive reel 52, and the angle of jib 1 reduces thus.Decide luffing mechanism 40 to carry out opposite action according to state and can well known to a person skilled in the art decision circuit for example gating circuit or analogous circuit by setting, for example gating circuit, comparator circuit wait and realize.
It will be understood by those skilled in the art that luffing mechanism 40 also can adopt other form except above-mentioned two kinds of forms.
The foundation of fuzzy neural network model
Fuzzy neural network model determines to export to the control signal of luffing mechanism 40 thus according to weight (containing suspender 2 deadweights), operation object and/or the angle of suspender swing and the expectation variable quantity that angular acceleration is determined the angle between jib 1 and the horizontal surface of the angle of length, jib 1 and the horizontal surface of crane arm support 1, the operation object sling.Introduce fuzzy neural network model below.
Fig. 7 is a kind of scheme drawing of fuzzy neural network model.As shown in Figure 7, neural network generally can be divided input layer, a plurality of hidden layer and output layer.In embodiments of the present invention, can comprise three hidden layers.Owing to need to consider above 5 parameters, be the length, jib 1 of crane arm support 1 and the angle and the angular acceleration of weight (containing suspender 2 deadweights), operation object and/or the suspender swing of the angle of horizontal surface, the operation object sling, therefore, the neuron number of input layer can be 5.The neuron number of output layer can be one.
First hidden layer is the obfuscation layer, is used for the input obfuscation with input layer, can adopt bell membership function to calculate degree of membership according to fuzzy rule.
For example can adopt Takagi-Sugeno (TS) fuzzy logic system, this system is the very strong fuzzy system of a kind of adaptive ability, and this model can not only upgrade automatically, and can constantly revise fuzzy subset's subordinate function.The TS fuzzy system defines with following " if-then " rule format, is that fuzzy reasoning is as follows under the situation of Rj in rule:
R
j:If?x
1?is
x
2?is
...,x
k?is
then?y
j=p
j0+p
j1x
1+...+p
jnx
k
Wherein,
Be x
iJ linguistic variable value, p
Ji(j=1,2 ... m) be the fuzzy system parameter; y
jBe the output result who obtains according to j bar fuzzy rule.
Suppose for input x=[x
1, x
1..., x
n], at first calculate each input variable x according to fuzzy rule
iDegree of membership, subordinate function adopts Gaussian function, promptly
J=1,2 ..., m; I=1,2 ..., n formula (1)
In the formula,
Be respectively the center and the width of membership function; N is the input parameter number; M is fuzzy subset's number;
Be corresponding subordinate function.
Second hidden layer is used for the fuzzy rule computation layer, and each degree of membership is carried out Fuzzy Calculation, adopts fuzzy operator to take advantage of operator for connecting.
J=1,2 ..., m formula (2)
The 3rd hidden layer is used to carry out normalization method and calculates.For example can carry out normalization method according to formula (3) calculates:
Formula (3)
Output layer is used to realize that sharpening calculates.For example can carry out sharpening according to formula (4) calculates:
Formula
Formula (4)
Thus, the pass of the input of this fuzzy neural network model and output is:
Formula (5)
In above formula,
Be respectively the center and the width of membership function; N is the input parameter number; M is fuzzy subset's number; x
iBe i input variable; y
jBe the output result who obtains according to j bar fuzzy rule.
The learning algorithm of fuzzy neural network model is as follows:
(1) error is calculated
Formula (6)
Wherein, y
dIt is the network desired output; y
cBe the network real output; E is the error of desired output and real output.
(2) coefficient correction
Formula (7)
In the formula, p
JiBe the neural network coefficient; η is the e-learning rate, and the learning rate that adopts among the design is 0.2; x
iBe the network input parameter; a
jBe input parameter degree of membership continued product; K is the study number of times, and the study number of times is when 600-800 among the design, and effect is better; y
d, y
c, e is the same.
(3) parameter correction
Formula (8)
Formula (9)
In the formula,
Be respectively the center and the width of membership function, all the other are the same.
Learning sample
The study of fuzzy neural network model needs learning sample to determine parameter p
Ji,
With
The accuracy of the accuracy of learning sample itself and the quantity of sample decision fuzzy neural network model.Therefore, the acquisition of learning sample is very important.
Can under different brachiums, different angle, the different situation of hanging load carrying ability, obtain the angle and the cireular frequency of operation object (or suspender 2) swing, obtain many group input samples { brachium, angle, weight, angle, cireular frequency } thus, wherein brachium is meant the length of jib 1, angle is meant that angle, the weight between jib 1 and the horizontal surface is meant the weight (can comprise suspender 2 deadweights) of the operation object of slinging, and angle and cireular frequency refer to the angle and the cireular frequency of suspender 2 and/or the swing of operation object respectively.After determining the input sample, also need to determine and the cooresponding desired output of input sample.When obtaining this group input sample, by repeatedly experiment, control luffing mechanism 40 changes the angle of jib 1 to reduce the swing of operation object (suspender 2).When the amplitude of swing in certain experiment is positioned at when setting tolerance, think that then the variable quantity of angle of this jib 1 imports the desired output of sample for this group.(if having repeatedly in the experiment amplitude of fluctuation all be positioned at and set tolerance, the variable quantity of angle that then can the minimum pairing jib 1 of amplitude of fluctuation is a desired output).In addition, the variable quantity of the angle between jib 1 and the horizontal surface is carried out by luffing mechanism 40.Therefore be input to exist one to one between the size of size and expectation variable quantity of control signal of luffing mechanism 40 and concern.Therefore, determine the expectation variable quantity, just can determine to be input to the control signal of luffing mechanism 40.
According to above method, can obtain many group input samples and cooresponding desired output, i.e. learning sample.Fuzzy neural network model can utilize these learning samples to regulate parameter p
Ji,
With
After learning process finished, the foundation of fuzzy neural network model had just been finished.
Above-mentioned fuzzy neural network model can realize that this software can be to be embedded in the controller 10 by the form of software, maybe can be stored in the internal storage (in-chip FLASH, ram in slice) or external memory storage of controller 10.The software that uses for example has C language, MATLAB etc.
Hoisting crane anti-sway system according to the embodiment of the present invention, it utilizes fuzzy neural network model according to the crane arm support length that influences operation object and/or suspender 2 swings, the angle of jib 1 and horizontal surface, the weight of operation object, the angle of swing and cireular frequency are regulated the angle of jib 1 and horizontal surface, especially the operation object when initial position has just been sling stress state and/or regulate the angle of jib 1 at the unloaded state that the operation object is placed to destination locations, thereby effectively reduce the swing of operation object and/or suspender 2, play the well anti-effect of shaking.This embodiment has been considered the angle and the angular acceleration of the length, jib 1 of the jib 1 of hoisting crane and angle, operation weight of object, operation object and/or suspender 2 swings of horizontal surface, utilizes based on the controller 10 of fuzzy neural network and regulates angle between jib 1 and the horizontal surface to reduce the swing of operation object.Therefore, this anti-sway system can adapt to different operation object load, brachium and different angles, and its controller performance has good robustness.
Below described the preferred implementation of this aspect in conjunction with the accompanying drawings, but these embodiments are exemplary and nonrestrictive.Ability office technical personnel can carry out various combinations to these embodiments with reasonable manner, and can carry out various modifications, distortion and replacement without departing from the scope of the invention.