CN113620024B - Data-driven multi-drive conveyor torque control method and device - Google Patents

Abstract

The invention discloses a data-driven multi-drive conveyor torque control method and a device, wherein the control method comprises the following steps: the method comprises the steps of dividing a bearing section of a conveying belt into n equal parts with fixed length, and obtaining the material quantity on each equal part of the conveying belt by combining an online image loading quantity measuring means. And establishing a data driving learning model of the deviation of the main driving torque and the middle driving torque and the deviation of the middle driving torque and the tail driving torque by taking the belt speed, the distribution vector, the main driving unit torque, the middle driving unit torque and the tail driving unit torque as input quantities. The control performance of the unloading type multi-point driving conveyor is improved, and the safety of the conveyor is improved. The LSSVM is adopted to complete modeling training, the requirements on storage space and on-line force calculation are low, the LSSVM is suitable for industrial controllers such as PLC and the like, and the problem of the tension of the rubber belt at the unloading point of the unloading type multi-point driving belt conveyor is solved.

Description

A data-driven multi-drive conveyor torque control method and device

technical field

The invention relates to the field of conveying devices, in particular to a data-driven torque control method and device for a multi-drive conveyor.

Background technique

As a continuous conveying equipment for bulk materials, the belt conveyor system is widely used in various industries of the national economy such as steel, coal, port terminals, electric power, building materials, etc., with a large amount of use and a wide range of applications. Belt conveyors are developing in the direction of large-scale, long-distance, high-speed, large-capacity and intelligent. When designing and developing a long-distance belt conveyor system, due to the limited driving force provided by a single drive motor and the limited maximum tension that the conveyor belt can withstand, large-scale belt conveyors usually use multiple motors. In this way, on the one hand, the capacity of a single motor can be reduced, and on the other hand, the tension of the tape can also be reduced. The multi-point drive not only reduces the requirements for the strength of the tape, but also facilitates the selection of equipment and realizes the miniaturization of the equipment. In essence, multi-point drive is a way to improve cost performance. There are many ways of multi-point drive, including linear friction type, wire rope traction type and middle drum unloading type. Among them, the roller unloading multi-point drive is to add one or several groups of driving devices in the middle of the ordinary belt conveyor, and distribute the driving force usually set at the head to several parts. The unloading intermediate drive mode has a simple structure and is easy to arrange, which can effectively reduce the cost of manufacturing, transportation, installation and maintenance management, and is widely used in multi-drive conveyor systems. However, the middle transfer point will cause the distribution of the entire belt tension to be completely different from the centralized drive layout, and the tension on the belt winding out side of the intermediate drive wheel will be significantly reduced, and if the control is not appropriate, the belt will slip; The problem of falling, uneven distribution on the tape, and difficult control of the drive system of the central roller unloading multi-drive conveyor.

For example, a "new type belt conveyor" disclosed in the Chinese patent document, the bulletin number CN103662715, includes a fuselage part, an unloading part, a machine head, a storage belt tensioning part, a belt retracting device, a driving device, a transition Anti-deviation front idler set, tape, self-moving tail, belt protection and belt conveyor control system, video monitoring system. However, in the above scheme, the belt is driven by the tail drum, and the drive drum is controlled by the frequency converter. Only the tail drum is driven. When the goods on the tape are unevenly distributed during long-distance transportation, the distribution of the entire tape tension will be uneven. Even if the tail is adjusted The roller also cannot balance the tension of the entire tape. If the control is not appropriate, it will cause the tape to slip, and the intermediate transfer point will also cause the material to fall.

SUMMARY OF THE INVENTION

The present invention is to overcome the problem of slippage caused by the tape tension on the winding out side of the driving roller becoming smaller in the prior art of the unloading multi-drive conveyor under changing working conditions, and provides a method for obtaining a master drive unit and a slave drive unit by adopting a data drive method. The torque deviation prediction model between the two can realize the coordinated control between the master and slave drive units, and improve the control performance of the unloaded multi-point drive conveyor and the safety of the conveyor.

To achieve the above object, the present invention adopts the following technical solutions:

A data-driven torque control method for a multi-drive conveyor includes a main drive unit, a middle drive unit and a tail drive unit, and the control method includes the following steps:

Step S1: the head drive unit is set to the speed drive mode, and both the middle drive unit and the tail drive unit are set to the torque control mode;

Step S2: set the main sampling period T _S , collect the material distribution vector x ^k , the main drive unit torque T _ET , the middle drive unit torque T _EZ , and the tail drive unit torque T _EW ;

Step S3: Add T _ET , T _EW , and conveyor belt speed v into x ^k to form the input vector of the middle drive unit

Add T _EZ , T _ET , and v to the material distribution vector x ^k to form the input vector of the tail drive unit

和尾部驱动单元的输入向量

输入到预测模型中进行计算，获得模型输出值即中部转矩偏差ΔT_E(TZ)和尾部转矩偏差ΔT_E(ZW)；Step S4: call the prediction model established and trained offline, read the pre-stored model parameters, and convert the input vector of the central drive unit

and the input vector of the tail drive unit

Input into the prediction model for calculation, and obtain the model output values, namely, the middle torque deviation ΔT _E(TZ) and the tail torque deviation ΔT _E(ZW) ;

Step S6: Set the torque adjustment value of the middle drive unit and the torque adjustment value of the tail drive unit according to the middle torque deviation ΔT _E(TZ) and the tail torque deviation ΔT _E(ZW) .

Preferably, obtaining the material distribution vector in the step S2 includes the following steps:

Step S21: Divide the carrying section of the conveyor belt into n equal parts of the conveyor belt section of fixed length, and obtain the instantaneous load q of the conveyor belt;

Step S22: Obtain the material amount of the first conveyor belt as follows:

Among them, N=t ₁ /t _s , l is the length of the conveyor belt section, v is the speed of the conveyor belt, t _l =l/v, t _s is the sampling period;

时间后获得输送带上的物料分布向量:Step S23: After

The material distribution vector on the conveyor belt is obtained after time:

_x =(x1, _x2 ... _xm , _{xm+1 ...} xn),

Among them, s is the belt length between the middle drive winding point and the blanking point, and v is the speed of the conveyor belt.

后，每个段的装载量都将变成已知量。向量x＝(x₁,x₂…x_m,x_m+1…x_n)描述了输送带上的物料分布，在本说明书中称之为物料分布向量。周期性获取机尾胶带图像，计算带式输送机瞬时装载量。数据驱动建模获得装载量的趋势，不需要其绝对值，故无需对测量系统进行标定，实现更加方便。Over time, the material will be passed backwards on the conveyor belt in sequence, forming a linked list: x ₁ →x ₂ →...→x _m →x _m+ 1→...→x _n . When the conveyor running time is greater than

After that, the loading amount of each segment will become a known amount. The vector x=(x ₁ , x ₂ . . . x _m , x _{m +1} . Periodically obtain the image of the tail tape, and calculate the instantaneous loading of the belt conveyor. Data-driven modeling obtains the trend of loading capacity, and does not need its absolute value, so it is not necessary to calibrate the measurement system, which is more convenient to implement.

Preferably, the establishment and training of the prediction model described in step S4 includes the following steps:

S41: Set the head drive unit, the middle drive unit and the tail drive unit to the speed control mode;

S42: Set the online load measurement period _ts ', and start the online measurement of the load; set the variable sampling period T _S ' to obtain v, x, T _ET , T _EZ , T _EW , T _S '>t _s ';

S43: Add T _ET , T _EW , v into x to form the input vector of the middle drive unit, X _Z =(x ₁ , x ₂ . . . x _m , x _m+1 . . . x _n , v, T _EW , T _ET ); add T _EZ , T _ET , v into x to form the input vector of the tail drive unit, X _W = (x ₁ , x ₂ . . . x _m , x _{m + 1 .} ;

和尾部驱动单元学习样本

存入样本集

和

其中k为第k个T_S’周期，N^Z表示中部驱动单元样本数量，N^W表示尾部驱动单元样本数量；S44: Collect the data of various working conditions of the transporter, and obtain the learning sample of the central drive unit through processing

and tail drive unit learning samples

save sample set

and

Where k is the kth T _S ' period, N ^Z represents the number of samples of the middle drive unit, and N ^W represents the number of samples of the tail drive unit;

S45: Model and train the sample set through the LSSVM algorithm;

S46: After the training is completed, the parameters of the LSSVM model are stored in the PLC for online use.

Among them, N ^ε is the sample capacity. Since the sparse algorithm needs to be used to eliminate the learning samples in the LSSVM training process, N ^ε is smaller than the original number of samples N ^Z .

The motor torque TE of the conveyor driving unit is proportional to the driving circular force F _u of the driving wheel, that is, _{TE =} k·F _u , and the driving circular force is equal to the difference between the tension at the entry point and the tension at the separation point. Because the conveyor belt forms a loop, the tension of the main driving wheel, the middle driving wheel and the tail driving wheel is coupled with the tension of the separation point, and the resistance of the belt is also affected by the running state, so it is impossible to obtain a more accurate multi-drive unit. Analytical relationship for torque. However, there is a definite relationship between the torque of the main drive unit, the torque of the slave drive unit and the real-time operating state of the conveyor, so the torque deviation model can be obtained by means of learning modeling.

If the main drive unit adopts speed control, the middle slave drive unit and the tail slave drive unit also adopt speed control, and the speed setting value tracks the main drive inverter. At this time, there is no coupling relationship between the torques of the drive units, and the difference mainly depends on the operating parameters such as the operating speed of the conveyor and the distribution of materials. Using the operating data obtained in this working mode, a torque prediction model can be established.

Preferably, the step S44 further includes the following steps

Step S441 : establish a prediction model f ₁ of the torque difference ΔTE _(TZ) between the head drive unit and the middle drive unit, and a prediction model f ₂ of the torque difference ΔTE _(ZW) between the middle drive unit and the tail drive unit:

ΔTE _(TZ) = f ₁ (X _z ),

ΔTE _(ZW) = f ₂ (X _W );

加入模型f₁样本集中，形成样本

加入f₂样本集中,其中K为第K采集周期；Step _S442 : Collect T _ET , _TEZ , TEW and v with T _S ′ as a cycle, and form samples through steps S21 and S22

Join the model f ₁ sample set to form a sample

Join the _f2 sample set, where K is the Kth collection period;

和

Step S443: Obtain a sample set

and

Set the master drive unit as speed control, the middle slave drive unit and the tail slave drive unit also use speed control, and the speed set value tracks the master drive controller, set the data sampling period to T _s , and T _S >t _s , according to t _s Periodically performs on-line measurement of the loading amount, and periodically updates the material distribution phasor X.

Preferably, the step S45 performs modeling training on the sample set by using the LSSVM algorithm, including: establishing a torque deviation decision function of the central drive unit:

Tail drive unit torque deviation decision function:

和

为LSSVM算法的核函数。建模训练完成后，需要将学习样本集中的支持向量

支持值向量α、偏置b、核参数σ、正规化参数c等存入PLC。利用学习样本集通过训练来逼近非线性模型f₁、f₂，LSSVM模型对存储空间需求不大，对计算量要求较小，适合在PLC等工业控制器中使用。Among them, α is the support value vector, b is the bias, σ is the kernel parameter, c is the normalization parameter,

and

is the kernel function of the LSSVM algorithm. After the modeling training is completed, the support vectors in the learning sample set need to be

Support value vector α, bias b, kernel parameter σ, normalization parameter c, etc. are stored in PLC. Using the learning sample set to approximate the nonlinear models f ₁ and f ₂ through training, the LSSVM model does not require much storage space and requires less computation, so it is suitable for use in industrial controllers such as PLC.

Preferably, the parameters of the LSSVM model described in step S46 include: a support value vector α, a bias b, and a kernel parameter σ, which is a normalization parameter c. In the online application stage, the above parameters are extracted for prediction calculation. In actual use, the kernel function K can be used for calculation, and the model prediction value can be obtained by accumulative summation.

Preferably, the kernel function is a radial basis kernel function RBF.

其中α_k为对应样本的支持值，b为偏置。模型参数α_k和b可以通过如下方程组求解：

where α _k is the support value of the corresponding sample, and b is the bias. The model parameters α _k and b can be solved by the following system of equations:

in,

U=(K+c ⁻¹ I) ⁻¹ ;

In the U expression, I is the identity matrix, σ is the kernel parameter, and c is the normalization parameter, that is, the penalty coefficient;

Preferably, the step S6 further includes superimposing the manual correction value ΔT _Z on the middle torque deviation ΔT _E(TZ) to form the torque adjustment value of the middle drive unit, and superimposing the manual correction on the tail torque deviation ΔT _E(ZW) The value _ΔTW forms the torque adjustment value of the tail drive unit. Ideally, ΔT _Z = 0, ΔT _W = 0, there is a problem of model mismatch in actual operation, appropriately increase or decrease the output value of the predicted model, realize the mixing of the model output value and the correction value, and improve the control effect and control. performance.

Preferably, the step S6 further includes a limiter module, which respectively judges whether the torque adjustment value of the middle drive unit and the torque adjustment value of the tail drive unit exceed the limit. Decrease the torque change rate, if not, output the torque adjustment value of the middle drive unit and the torque adjustment value of the tail drive unit. The upper limit of the limiter is the increase of Δ on the basis of the current torque, and the lower limit of the limiter is the decrease of Δ on the basis of the current actual torque. The torque setting value cannot be too different from the current torque. If the difference between the torque adjustment value and the current torque exceeds Δ, the torque limiter will be used to appropriately reduce the adjustment value change rate, and gradually complete the adjustment of the middle drive unit and the tail. Torque adjustment of the drive unit.

A data-driven multi-drive conveyor torque control device adopts the data-driven multi-drive conveyor torque control method. There is a main drive unit at the head of the conveyor, a middle drive unit at the middle, and a tail drive unit at the tail. Each drive unit includes two coaxially installed drive motors. The motor control system controller adopts PLC and frequency converter. Drive, PLC and inverter use field bus protocol to communicate to realize coordinated control between drive units. The bus is made of fiber optic material to prevent electromagnetic interference and prevent lightning strikes.

Therefore, the present invention has the following beneficial effects: (1) The present invention adopts the data-driven method to obtain the torque deviation prediction model of the main drive unit, the intermediate drive unit and the tail drive unit in combination with the on-line measurement method of the load, so as to realize the master-slave drive unit. coordination control between them. (2) The data-driven modeling method obtains the torque setting value from the drive group, improves the control performance of the unloaded multi-point drive conveyor, and improves the safety of the conveyor. (3) LSSVM is used to complete the modeling training. LSSVM has low requirements on storage space and online computing power, and is suitable for industrial controllers such as PLC to solve the problem of tape tension at the unloading point of the unloading multi-point drive belt conveyor.

Description of drawings

FIG. 1 is a model diagram of a multi-drive conveyor according to an embodiment of the present invention.

FIG. 2 is a schematic diagram of the tension of a multi-drive conveyor according to an embodiment of the present invention.

FIG. 3 is a control block diagram of a multi-drive conveyor according to an embodiment of the present invention.

4 is a schematic diagram of a material distribution phasor of a multi-drive conveyor according to an embodiment of the present invention.

FIG. 5 is a schematic diagram of an input vector of a multi-drive conveyor according to an embodiment of the present invention.

FIG. 6 is a schematic diagram of the mechanism of a multi-drive conveyor control system according to an embodiment of the present invention.

FIG. 7 is a flow chart of modeling and use of a torque control method for a multi-drive conveyor according to an embodiment of the present invention.

In the figure: 1. Main drive unit 2, middle drive unit 3, tail drive unit 4, camera and light source 5, materials.

Detailed ways

The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

Example:

As shown in Figures 1 to 7, a data-driven unloading multi-drive conveyor slave drive unit control method and device, the head, middle and tail of the conveyor are equipped with three drive units, each drive unit includes two A coaxially mounted drive motor. The controller of the motor control system adopts PLC, which is driven by inverter, and the field bus protocol is used for communication between PLC and inverter to realize coordinated control between drive units. The bus is made of fiber optic material to prevent electromagnetic interference and prevent lightning strikes.

The two motors in each group of drive units are installed coaxially and belong to rigid body connection. They are controlled by master and slave. The inverter of the slave motor works in the direct torque control mode and receives the torque set value sent by the master motor. In the following description, the drive unit composed of two coaxially mounted motors will be described as a whole.

When the conveyor is running, the torque of each drive unit is related to the material distribution on the belt, belt speed and other factors. In this embodiment, the method of image processing is used to realize the real-time measurement of the instantaneous conveying capacity. As shown in Figure 1, the industrial camera periodically obtains the tape image at the head drive unit at t _s , and calculates the instantaneous loading capacity of the belt conveyor, including the following steps : As shown in Figure 4, divide the carrying section of the tape into n equal parts of fixed length, each equal part has a length l, in the inclined carrying section, the length is calculated by the actual belt length, not its horizontal projection, in the middle unloading station The back end takes the blanking point as the starting point for the calculation. Calculating the instantaneous loading of the belt conveyor includes the following steps: the industrial camera periodically obtains the image of the tape at the tail of the machine, and calculates the instantaneous loading of the belt conveyor. Data-driven modeling obtains the trend of loading capacity, and does not need its absolute value, so it is not necessary to calibrate the measurement system, which is more convenient to implement.

Step S21: Divide the carrying section of the conveyor belt into n equal parts of the conveyor belt section of fixed length, and obtain the instantaneous load q of the conveyor belt;

Step S22: Obtain the material amount of the first conveyor belt as follows:

Among them, N=t ₁ /t _s , l is the length of the conveyor belt section, v is the speed of the conveyor belt, t _l =l/v, t _s is the sampling period;

时间后获得输送带上的物料分布向量:Step S23: After

The material distribution vector on the conveyor belt is obtained after time:

_x =(x1, _x2 ... _xm , _{xm+1 ...} xn),

Among them, s is the belt length between the middle drive winding point and the blanking point, and v is the speed of the conveyor belt.

后，每个段的装载量都将变成已知量。向量x＝(x₁,x₂…x_m,x_m+1…x_n)描述了输送带上的物料分布，在本说明书中称之为物料分布向量。Over time, the material will be passed backwards on the conveyor belt in sequence, forming a linked list: x ₁ →x ₂ →...→x _m →x _m+1 →...→x _n . When the conveyor running time is greater than

After that, the loading amount of each segment will become a known amount. The vector x=(x ₁ , x ₂ . . . x _m , x _{m +1} .

Offline establishment of a trained torque deviation prediction model, including the following steps:

S41: Set the head drive unit, the middle drive unit and the tail drive unit to the speed control mode;

S42: Set the online load measurement period _ts ', and start the online measurement of the load; set the variable sampling period T _S ' to obtain v, x, T _ET , T _EZ , T _EW , T _S '>t _s ;

S43: Add T _ET , T _EW , v into x to form the input vector of the middle drive unit, X _Z =(x ₁ , x ₂ . . . x _m , x _m+1 . . . x _n , v, T _EW , T _ET ); add T _EZ , T _ET , v into x to form the input vector of the tail drive unit, X _W = (x ₁ , x ₂ . . . x _m , x _{m + 1 .} ;

和尾部驱动单元学习样本

存入样本集

和

其中k为第k个T_S’周期，N^Z表示中部驱动单元样本数量，N^W表示尾部驱动单元样本数量；S44: Collect the data of various working conditions of the transporter, and obtain the learning sample of the central drive unit through processing

and tail drive unit learning samples

save sample set

and

Where k is the kth T _S ' period, N ^Z represents the number of samples of the middle drive unit, and N ^W represents the number of samples of the tail drive unit;

Step S441 : establish a prediction model f ₁ of the torque difference ΔTE _(TZ) between the head drive unit and the middle drive unit, and a prediction model f ₂ of the torque difference ΔTE _(ZW) between the middle drive unit and the tail drive unit:

ΔTE _(TZ) = f ₁ (X _z ),

ΔTE _(ZW) = f ₂ (X _W );

加入模型f₁样本集中，形成样本

加入f₂样本集中，其中K为第K采集周期；Step _S442 : Collect T _ET , _TEZ , TEW and v with T _S ′ as a cycle, and form samples through steps S21 and S22

Join the model f ₁ sample set to form a sample

Join the f ₂ sample set, where K is the K-th collection cycle;

和

Step S443: Obtain a sample set

and

Set the master drive unit as speed control, the middle slave drive unit and the tail slave drive unit also use speed control, and the speed set value tracks the master drive controller, set the data sampling period to T _s , and T _S >t _s , according to t _s periodically performs on-line measurement of loading, and periodically updates the material distribution phasor.

S45: Model and train the sample set through the LSSVM algorithm;

Establish the torque deviation decision function of the central drive unit:

Tail drive unit torque deviation decision function:

和

为LSSVM算法的核函数。建模训练完成后，需要将学习样本集中的支持向量

支持值向量α、偏置b、核参数σ、正规化参数c等存入PLC。利用学习样本集通过训练来逼近非线性模型f₁、f₂，LSSVM模型对存储空间需求不大，对计算量要求较小，适合在PLC等工业控制器中使用。Among them, α is the support value vector, b is the bias, σ is the kernel parameter, c is the normalization parameter,

and

is the kernel function of the LSSVM algorithm. After the modeling training is completed, the support vectors in the learning sample set need to be

Support value vector α, bias b, kernel parameter σ, normalization parameter c, etc. are stored in PLC. Using the learning sample set to approximate the nonlinear models f ₁ and f ₂ through training, the LSSVM model does not require much storage space and requires less computation, so it is suitable for use in industrial controllers such as PLC.

The kernel function is the radial basis kernel function RBF.

其中α_k为对应样本的支持值，b为偏置。模型参数α_k和b可以通过如下方程组求解：

where α _k is the support value of the corresponding sample, and b is the bias. The model parameters α _k and b can be solved by the following system of equations:

in,

U=(K+c ⁻¹ I) ⁻¹ ;

In the U expression, I is the identity matrix, σ is the kernel parameter, and c is the normalization parameter, that is, the penalty coefficient;

S46: After the training is completed, the supported value model parameters are stored in the PLC for online use.

Among them, N ^ε is the sample capacity. Since the sparse algorithm needs to be used to eliminate the learning samples in the LSSVM training process, N ^ε is smaller than the original number of samples N ^Z .

The motor torque TE of the conveyor driving unit is proportional to the driving circular force F _u of the driving wheel, that is, _{TE =} k·F _u , and the driving circular force is equal to the difference between the tension at the entry point and the tension at the separation point. Because the conveyor belt forms a loop, the tension of the main driving wheel, the middle driving wheel and the tail driving wheel is coupled with the tension of the separation point, and the resistance of the belt is also affected by the running state, so it is impossible to obtain a more accurate multi-drive unit. Analytical relationship for torque. However, there is a definite relationship between the torque of the main drive unit, the torque of the slave drive unit and the real-time operating state of the conveyor, so the torque deviation model can be obtained by means of learning modeling.

If the main drive unit adopts speed control, the middle slave drive unit and the tail slave drive unit also adopt speed control, and the speed setting value tracks the main drive inverter. At this time, there is no coupling relationship between the torques of the drive units, and the difference mainly depends on the operating parameters such as the operating speed of the conveyor and the distribution of materials. Using the operating data obtained in this working mode, a torque prediction model can be established.

The online torque control of the multi-drive conveyor includes the following steps:

Step S1: the head drive unit is set to the speed drive mode, and both the middle drive unit and the tail drive unit are set to the torque control mode;

Step S2: set the main sampling period T _S , collect the material distribution vector x ^k , the main drive unit torque T _ET , the middle drive unit torque T _EZ , and the tail drive unit torque T _EW ;

Step S3: adding T _ET , T _EW , and v to x ^k to form the input vector of the middle drive unit

Add T _EZ , T _ET , and v to the material distribution vector x ^k to form the input vector of the tail drive unit

和尾部驱动单元的输入向量

输入到预测模型中进行计算，获得模型输出值即中部转矩偏差ΔT_E(TZ)和尾部转矩偏差ΔT_E(ZW)；Step S4: call the prediction model established and trained offline, read the pre-stored model parameters, and convert the input vector of the central drive unit

and the input vector of the tail drive unit

Input into the prediction model for calculation, and obtain the model output values, namely, the middle torque deviation ΔT _E(TZ) and the tail torque deviation ΔT _E(ZW) ;

Step S6: Set the torque adjustment value of the middle drive unit and the torque adjustment value of the tail drive unit according to the middle torque deviation ΔT _E(TZ) and the tail torque deviation ΔT _E(ZW) . After superimposing the artificial correction value ΔT _Z on the middle torque deviation ΔT _E(TZ) , the torque adjustment value of the middle drive unit is formed, and after superimposing the artificial correction value ΔT _W on the tail torque deviation ΔT _E(ZW) , the rear drive unit torque is formed. torque adjustment value. Determine whether the torque adjustment value of the middle drive unit and the torque adjustment value of the tail drive unit exceed the limit respectively. If so, limit the torque adjustment value to the upper limit and appropriately reduce the torque change rate. If not, output the middle drive Unit torque adjustment value and tail drive unit torque adjustment value.

Improve the control performance of the unloaded multi-point drive conveyor and improve the safety of the conveyor.

The specific embodiments described herein are merely illustrative of the spirit of the invention. Those skilled in the art to which the present invention pertains can make various modifications or additions to the described specific embodiments or substitute in similar manners, but will not deviate from the spirit of the present invention or go beyond the definitions of the appended claims range.

Although the terms such as torque adjustment value, prediction model, material distribution phasor, on-time loading amount, torque deviation, etc. are used frequently in this paper, the possibility of using other terms is not excluded. These terms are used only to more conveniently describe and explain the essence of the present invention; it is contrary to the spirit of the present invention to interpret them as any kind of additional limitation.

Claims (10)

1. A data-driven multi-drive conveyor torque control method is characterized by comprising a main drive unit, a middle drive unit and a tail drive unit, and the control method comprises the following steps:

step S1: the head driving unit is set to be in a speed driving mode, and the middle driving unit and the tail driving unit are both set to be in a torque control mode;

step S2: setting a main sampling period T _S Collecting material distribution vector x ^k Main drive unit torque T _ET Middle drive unit torque T _EZ Tail drive unit torque T _EW ；

And step S3: will T _ET 、T _EW The belt speed v of the conveyor belt is added with x ^k In forming the input vector of the middle drive unit

Will T _EZ 、T _ET V distribution vector x of added material ^k In forming the input vector of the tail drive unit

And step S4: calling the off-line built and trained prediction model, reading the pre-stored model parameters, and inputting the vector of the middle drive unit

And input vector of tail drive unit

Inputting the data into a prediction model for calculation to obtain a model output value, namely a middle torque deviation delta T _E(TZ) And tail torque deviation Δ T _E(ZW) ；

Step S6: according to the central torque deviation Delta T _E(TZ) And tail torque deviation Δ T _E(ZW) A mid-drive unit torque adjustment value and a tail-drive unit torque adjustment value are set.

2. The method as claimed in claim 1, wherein the step S2 of obtaining the material distribution vector comprises the steps of:

step S21: dividing the conveyor belt bearing section into n equal conveyor belt sections with fixed length to obtain the instantaneous loading q of the conveyor belt;

step S22: the material quantity of the first section of the conveying belt is obtained as follows:

wherein N = t ₁ /t _s L is the length of the conveyor belt, v is the speed of the conveyor belt, t _l ＝l/v，t _s Is a sampling period;

step S23: through a process

Obtaining a material distribution vector on the conveying belt after time:

x＝(x ₁ ,x ₂ ...x _m ,x _m+1 ...x _n )，

wherein s is the belt length between the middle driving winding-out point and the blanking point, and v is the belt speed of the conveying belt.

3. The method as claimed in claim 2, wherein the step of building and training the predictive model of step S4 comprises the steps of:

s41: setting the head driving unit, the middle driving unit and the tail driving unit to be in a speed control mode;

s42: setting on-line load measurement period t _s ', start the on-line measurement of the load; setting variable sampling period T _S ' obtaining v, x, T _ET 、T _EZ 、T _EW ，T _S ’＞t _s ’；

S43, adding T _ET 、T _EW V is added to X to form the input vector of the central drive unit, X _Z ＝(x ₁ ，x ₂ ...x _m ，x _m+1 ...x _n ，v，T _EW ，T _ET ) (ii) a Will T _EZ ，T _ET V is added to X to form the tail drive unit input vector, X _W ＝(x ₁ ，x ₂ ...x _m ，x _m+1 ...x _n ，v，T _EZ ，T _ET )；

S44: the method comprises the steps of collecting data of multiple working conditions of a conveyor, and processing the data to obtain a middle driving unit learning sample

And tail drive unit learning samples

Deposit sample set

And

where k is the kth T _S ' period, N ^Z Representing the number of samples of the middle drive unit, N ^W Representing the number of tail drive unit samples;

s45: modeling and training a sample set through an LSSVM algorithm;

s46: and after the training is finished, storing the LSSVM model parameters into a PLC (programmable logic controller) for online use.

4. The method as claimed in claim 3, wherein the step S44 further comprises the step of

Step S441: establishing a head drive unit to mid drive unit torque differential Δ T _E(TZ) Prediction model f ₁ Torque difference Δ T between the middle drive unit and the tail drive unit _E(ZW) Prediction model f ₂ ：

ΔT _E(TZ) ＝f ₁ (X _z )，

ΔT _E(ZW) ＝f ₂ (X _w )；

Step S442: by T _S ' for periodic acquisition of T _ET 、T _EZ 、T _EW And v, forming a sample through step S21 and step S22

Adding model f ₁ Collecting the samples to form samples

Adding f ₂ Collecting samples, wherein K is the Kth collection period;

step S443: obtaining a sample set

And

5. the method as claimed in claim 4, wherein the step S45 of performing modeling training on the sample set through the LSSVM algorithm comprises: establishing a torque deviation decision function of the middle driving unit:

tail drive unit torque deviation decision function:

wherein alpha is a vector of support values, b is an offset, sigma is a kernel parameter, c is a normalization parameter,

and

is the kernel function of the LSSVM algorithm.

6. The method of claim 5, wherein the LSSVM model parameters of step S46 comprise: the vector α of support values, offset b, and kernel parameter σ are normalization parameters c.

7. The method as claimed in claim 6, wherein said kernel function is a radial basis kernel function RBF.

8. The method as claimed in claim 7, wherein the step S6 further comprises a step of controlling the torque deviation Δ T in the middle _E(TZ) Upper superimposed artificial correction value delta T _Z Then forming the torque adjustment value of the middle driving unit and the torque deviation delta T at the tail part _E(ZW) Upper superimposed artificial correction value delta T _W And then forming a tail drive unit torque adjustment value.

9. The method as claimed in claim 8, wherein step S6 further comprises a limiting module for determining whether the torque adjustment value of the middle driving unit and the torque adjustment value of the tail driving unit exceed a limit, respectively, if yes, limiting the torque adjustment value to an upper limit value, and reducing the torque change rate appropriately, otherwise, outputting the torque adjustment value of the middle driving unit and the torque adjustment value of the tail driving unit.

10. A data-driven multi-drive conveyor torque control apparatus, characterized by adopting the data-driven multi-drive conveyor torque control method according to any one of claims 1 to 9.

Images