RU2812444C1 - System and method for controlling material distribution based on predicting material layer thickness - Google Patents

Abstract

FIELD: ferrous metallurgy.

SUBSTANCE: group of inventions relates to the field of cast iron and steel smelting and can be used to control the distribution of powdered material during its sintering in a sintering system, which includes a drum feeder, a roller-type distribution device having an auxiliary shutter, and a sintering cart, wherein the drum feeder is made with the ability to supply the mixed material to the said distribution device, which is configured to supply the material to the sintering car. The proposed system and control method are based on predicting the thickness of the material layer, in which the values of the bulk density of the mixed material, the drum rotation speed, the distribution roller rotation speed, the degree of opening of the auxiliary gate and the speed of the sintering cart are determined. The characteristic value of the material layer thickness is generated using a predefined dynamic model. The material layer thickness deviation value is calculated in accordance with the predicted layer thickness value and the target layer thickness value. The deviation value is entered into the optimization model to obtain the values of the drum rotation speed, distribution roller rotation speed, auxiliary gate opening degree and sintering car speed to be adjusted.

EFFECT: improved quality of the resulting material.

10 cl, 7 dwg, 1 tbl

Description

FIELD OF THE INVENTION

[0001] The present invention relates to the technical field of iron and steel smelting, in particular, to a system and method for controlling material distribution based on predicting the thickness of a layer of material.

BACKGROUND OF THE INVENTION

[0002] Sintering is the process of converting powdered material into a solid body, and in the production of iron and steel, it means that iron ore powder, pulverized coal and lime are uniformly mixed according to a certain ratio and sintered to produce sinter ore that meets the requirements of the blast furnace production process furnaces with subsequent transportation of raw materials for blast furnace production. One of the important parameters affecting the productivity and quality of the sintering process is the material distribution process, during which it is determined whether the thickness of the material layer meets the necessary requirements. At present, in the material distribution process, the thickness of the layer of material in the area where the auxiliary gate is located is usually adjusted by adjusting the opening degree of the auxiliary gate, and the thickness of the layer of material located on the sintering car is changed over its entire width by adjusting the rotation speed of the drum. When the opening degree of the auxiliary gate is high, the material distribution device distributes a large amount of material onto the sintering car, and accordingly a thick layer of material is formed, otherwise a thin layer of material is formed.

[0003] In order to obtain a material layer of appropriate thickness, in the prior art, control of the material distribution is usually carried out by determining the thickness of the material layer on the sintering car in real time. Specifically, in the process of distributing material on the sintering cart, the thickness of the material layer on the sintering cart is controlled downstream of the moving direction of the sintering cart, and feedback control of the material distribution is performed in accordance with the obtained material layer thickness value.

[0004] During actual operation, the thickness of the material layer on the sintering car is determined using the specified method for determining the thickness of the material layer after the completion of the material distribution process for 2-3 minutes from the time of adjusting the auxiliary shutter to the time of determining the thickness of the material layer, while using the control with feedback can have a large lag effect. At the same time, the value of the material layer thickness cannot be changed, even if it is currently found that it does not meet the requirements, which is not conducive to timely and stable control of the material distribution in the sintering system.

BRIEF DESCRIPTION OF THE PRESENT INVENTION

[0005] The present invention provides a material distribution control system and a method based on material layer thickness prediction to solve the problem of lag caused by determining the material layer thickness on a sintering car when controlling the material distribution of a sintering system in the prior art.

[0006] The first aspect of the present invention provides a material distribution control system based on prediction of the material layer thickness. The material distribution control system based on prediction of the thickness of the material layer includes a drum feeder, a roller-type distribution device and a sintering cart, the drum feeder is configured to supply mixed material to the roller-type distribution device, the roller-type distribution device is configured to supply material to the sintering cart ; wherein the material distribution control system further includes a mixture analyzer, a drum feeder controller connected to the drum feeder, a roller-type distributor controller and an auxiliary gate controller connected to the roller-type distributor, a sintering car controller connected to the sintering cart, and a central a processing unit connected to the mixture analyzer, drum controller, distribution roller controller, auxiliary gate controller and sintering car controller;.

[0007] wherein the central processing unit is configured to perform the following steps:

[0008] obtaining values of the bulk density of the mixed material from the mixture analyzer, the drum rotation speed from the drum feeder controller, the distribution roller rotation speed from the roller type distributor controller, the auxiliary gate opening degree from the auxiliary gate controller, and the sintering car speed from the sintering car controller;

[0009] predicting the thickness of the material layer in accordance with the values of the bulk density of the mixed material, the rotation speed of the drum, the rotation speed of the distribution roller, the opening degree of the auxiliary gate and the speed of the sintering car to determine the predicted value of the thickness of the material layer;

[0010] calculating a material layer thickness deviation amount in accordance with the predicted material layer thickness value and the target material layer thickness value;

[0011] inputting the material layer thickness deviation amount into the moving average optimization model to obtain values of the drum rotation speed to be adjusted, the distribution roller rotation speed to be adjusted, the auxiliary gate opening degree to be adjusted, and the sintering car speed to be adjusted; And

[0012] driving the drum feeder controller to adjust the rotation speed of the drum in accordance with the rotation speed of the drum to be adjusted; operating a roller-type distribution device controller to regulate the rotation speed of the distribution roller in accordance with the rotation speed of the distribution roller to be adjusted; driving the auxiliary gate controller to regulate the opening degree of the auxiliary gate in accordance with the opening degree of the auxiliary gate to be adjusted, and driving the sintering car controller to regulate the speed of the sintering car in accordance with the speed of the sintering car to be adjusted,

[0013] wherein the moving average optimization model is configured to calculate the drum rotation speed, the distribution roller rotation speed, the degree of opening of the auxiliary gate and the speed of the sintering car, in accordance with the minimum standard deviation of the thickness of the material layer, subject to a constant value of the bulk density of the mixed material

[0014] If necessary, the standard deviation of the material layer thickness is determined as follows:

[0015]

[0016] where σ is the standard deviation of the thickness of the material layer, E(k) is the deviation value of the thickness of the material layer, R(k) is the specified sequence of the target value of the thickness of the material layer, and Y(k) is the sequence of the predicted value of the thickness of the material layer.

[0017] If necessary, the step of predicting the thickness of the material layer in accordance with the values of the bulk density of the mixed material, the rotation speed of the drum, the rotation speed of the distribution roller, the opening degree of the auxiliary gate and the speed of the sintering car to obtain the predicted value of the thickness of the material layer is particularly carried out with using the following steps:

[0018] quantifying the values of the bulk density of the mixed material, the rotation speed of the drum, the rotation speed of the distribution roller and the speed of the sintering car at the same time in the same time interval in accordance with a given compression ratio and in sum with the value of the opening degree of the auxiliary gate, obtaining a predicted characteristic vector thickness of the material layer;

[0019] inputting the predicted material layer thickness characteristic vector into a predefined dynamic material layer thickness prediction model to obtain a material layer thickness characteristic value, wherein the dynamic material layer thickness prediction model includes displaying a relationship between the predicted material layer thickness characteristic vector and the characteristic thickness value layer of material; And

[0020] performing a material layer thickness characteristic value data recovery to obtain a predicted material layer thickness value.

[0021] Optionally, the material distribution control system further comprises a material layer thickness determining device located above the sintering car, wherein the material layer thickness determining device is connected to the central data processing unit and is configured to determine the thickness of the material layer on the sintering car in accordance with a given time interval and obtaining the measured value of the thickness of the material layer; and the step of inputting the predicted material layer thickness characteristic vector into the predefined dynamic material layer thickness prediction model to generate the material layer thickness characteristic value further includes:

[0022] obtaining time sequence data approximating the predicted material layer thickness characteristic vector as samples, the samples including input samples and a material layer thickness characteristic value corresponding to the input samples; and

[0023] updating the dynamic material layer thickness prediction model in real time using the training samples to obtain an updated dynamic material layer thickness prediction model.

[0024] If it is necessary that the data recovery for the characteristic value of the material layer thickness be performed using the following method:

[0025] h _i =k _i × N,

[0026] where h _i is the predicted value of the material layer thickness obtained at time i, k _i is the characteristic value of the material layer thickness generated at time i, and H is the maximum allowable material layer thickness on the sintering car.

[0027] If necessary, the step of quantifying the values of the bulk density of the mixed material, the drum rotation speed, the distribution roller rotation speed and the sintering car speed simultaneously in the same time interval in accordance with a predetermined compression ratio is performed as follows:

[0028] the ratio of the bulk density value to the highest raw material density value among all components is calculated;

[0029] the ratio of the drum rotation speed to the maximum drum rotation speed is calculated;

[0030] the ratio of the rotation speed of the distribution roller to the maximum rotation speed of the distribution roller is calculated; And

[0031] The ratio of the speed of the sintering cart to the maximum turning speed of the sintering cart is calculated.

[0032] If necessary, a dynamic material layer thickness prediction model is generated based on the training of the neural network model and established according to the following steps

[0033] determining N groups of independent values of volumetric densities of the mixed material, drum rotation speed, distribution roller rotation speed, auxiliary gate opening degree and sintering car speed in accordance with a predetermined time interval;

[0034] quantifying N groups of independent values of the bulk densities of the mixed material, drum rotation speed, distribution roller rotation speed and sintering car speed in the same interval in accordance with a given compression ratio and in sum with the value of the auxiliary gate opening degree; generating input data N groups of training samples,

[0035] determining the actual thickness of the material layer on the sintering car in accordance with the input data of the N groups of training samples according to a predetermined time interval, and calculating the actual characteristic values of the actual thickness of the material layer, wherein the actual characteristic values serve as the output of the N groups of training samples;

[0036] using the input data of the training samples and the output data of the training samples to train the neural network model using the backpropagation method;

[0037] continuously updating the weight parameter, bias parameter and learning coefficient of the neural network model through iterative learning; and

[0038] If the predicted value and measured value of the neural network model reach a predetermined error limit or if the neural network model reaches a predetermined maximum number of iterations, training is completed and the final updated weight parameter, bias parameter, and training coefficient are stored to obtain a dynamic layer thickness prediction model material.

[0039] Optionally, the dynamic material layer thickness prediction model may be a material layer thickness prediction table, which is established in accordance with the following steps:

[0040] Determining N groups of independent values of bulk densities of the mixed material, drum rotation speed, distribution roller rotation speed, auxiliary gate opening degree and sintering car speed in accordance with a predetermined time interval;

[0041] determining the actual thickness of the material layer on the sintering car in accordance with the input data of N groups of training samples according to a given time interval and calculating the actual characteristic values of the actual thickness of the material layer, and

[0042] performing a statistical analysis of N groups of independent values of the bulk densities of the mixed material, the drum rotation speed, the distribution roller rotation speed, the auxiliary gate opening degree and the sintering car speed, and the corresponding measured values of the material layer thickness to create a material layer thickness prediction table.

[0043] A second aspect of the present invention provides a method for controlling material distribution based on predicting the thickness of a layer of material, wherein the method for controlling material distribution includes:

[0044] determining the values of the bulk density of the mixed material, the rotation speed of the drum, the rotation speed of the distribution roller, the opening degree of the auxiliary gate and the speed of the sintering car;

[0045] predicting the thickness of the material layer according to the values of the bulk density of the mixed material, the rotation speed of the drum, the rotation speed of the distribution roller, the opening degree of the auxiliary gate and the speed of the sintering car to obtain a predicted value of the thickness of the material layer,

[0046] calculating a material layer thickness deviation amount in accordance with the predicted material layer thickness value and the target material layer thickness value;

[0047] inputting the deviation amount of the material layer thickness into the moving average optimization model to obtain the value of the drum rotation speed to be adjusted, the distribution roller rotation speed to be adjusted, the auxiliary gate opening degree to be adjusted, and the sintering car speed to be adjusted; and

[0048] adjusting the rotation speed of the drum in accordance with the rotation speed of the drum to be adjusted, the rotation speed of the distribution roller in accordance with the rotation speed of the distribution roller to be adjusted, the opening degree of the auxiliary gate in accordance with the opening degree of the auxiliary gate to be adjusted, the speed of the sintering cart according to the speed of the sintering car to be adjusted,

[0049] wherein the moving average optimization model is configured to calculate the drum rotation speed, the distribution roller rotation speed, the auxiliary gate opening degree and the sintering car speed, in accordance with the minimum standard deviation of the material layer thickness, subject to a constant value of the bulk density of the mixed material.

[0050] Optionally, the step of predicting the thickness of the material layer in accordance with the values of the bulk density of the mixed material, the rotation speed of the drum, the rotation speed of the distribution roller, the opening degree of the auxiliary gate and the speed of the sintering car to obtain a predicted value of the thickness of the material layer includes:

[0051] quantifying the values of the bulk density of the mixed material, the drum rotation speed, the distribution roller rotation speed, the auxiliary gate opening degree and the sintering car speed simultaneously in the same time interval according to a given compression ratio and in the amount of the auxiliary opening degree value shutter obtaining the predicted characteristic vector of the thickness of the material layer;

[0052] inputting the predicted material layer thickness characteristic vector into a predefined dynamic material layer thickness prediction model to obtain a material layer thickness characteristic value, wherein the dynamic material layer thickness prediction model includes displaying a relationship between the predicted material layer thickness characteristic vector and the characteristic thickness value layer of material; and

[0053] performing a recovery of the material layer thickness characteristic value data to obtain a predicted material layer thickness value.

[0054] Known from the above-mentioned technical solutions, embodiments of the present invention provide a system and method for controlling material distribution based on predicting the thickness of a layer of material. The material distribution control system includes a drum feeder, a roller type distributor and a sintering cart, wherein the drum feeder is configured to supply mixed material to the roller type distribution device, and the roller type distributor device is configured to supply material to the sintering cart; The material distribution control system further includes a mixture analyzer, a drum feeder controller coupled to the drum feeder, a roller-type distributor controller and an auxiliary gate controller coupled to the roller-type distributor, a sintering car controller coupled to the sintering cart, and a central processing unit. , connected to the mixture analyzer, drum feeder controller, roller type distributor controller, auxiliary gate controller and sintering cart controller.

[0055] In practice, the values of the bulk density of the mixed material, the rotation speed of the drum, the rotation speed of the distribution roller, the opening degree of the auxiliary gate, and the speed of the sintering car are first calculated. Then, the values of bulk density, drum rotation speed, distribution roller rotation speed, auxiliary gate opening degree and sintering car speed are quantified simultaneously in the same time interval according to the specified compression ratio to obtain the predicted characteristic vector of the material layer thickness. The predicted material layer thickness characteristic vector is then input into a predefined dynamic material layer thickness prediction model to generate a material layer thickness characteristic value. The characteristic value of the material layer thickness data is reconstructed to obtain the predicted value of the material layer thickness. Then, the material layer thickness deviation amount is calculated in accordance with the predicted material layer thickness value and the target material layer thickness value. Finally, the deviation value of the material layer thickness is input into the moving average optimization model to obtain the values of drum rotation speed, distribution roller rotation speed, auxiliary gate opening degree and sintering car speed to be adjusted, and thus achieve material distribution control in the system sintering based on predicting the thickness of the material layer. According to the material distribution control system based on material layer thickness prediction provided in the embodiment of the present invention, the thickness of the material layer on the sintering car can be predicted in advance by obtaining the mixed material parameters and the state parameters of the sintering system, so that the sintering system achieves timely and stable feedback control of material distribution according to the predicted material layer thickness and real-time adjustment of relevant key parameters affecting material distribution.

BRIEF DESCRIPTION OF THE DRAWINGS

[0056] The following is a brief description of the accompanying drawings for a more detailed description of embodiments of the present invention. It should be understood that other accompanying drawings may be produced by one skilled in the art in accordance with these accompanying drawings without introducing creative work.

[0057] FIG. 1 is a block diagram of a material distribution control system based on material layer thickness prediction provided in an embodiment of the present invention;

[0058] FIG. 2 is a flowchart of operation of a material distribution control system based on material layer thickness prediction provided in an embodiment of the present invention;

[0059] FIG. 3 is a block diagram of material layer thickness prediction in a material distribution control system provided in an embodiment of the present invention;

[0060] FIG. 4 is a flowchart of real-time updating of a dynamic material layer thickness prediction model provided in an embodiment of the present invention;

[0061] FIG. 5 is a flowchart for generating a dynamic material layer thickness prediction model provided in an embodiment of the present invention;

[0062] FIG. 6 is a block diagram of an LSTM neural network model provided in an embodiment of the present invention; and

[0063] FIG. 7 is a flowchart for constructing another dynamic material layer thickness prediction model provided in an embodiment of the present invention.

[0064] List of reference numbers in the drawings:

1 material pre-processing workshop

2 mixer

3 drum feeder

4 roller type distributor

5 sintering trolley

6 incendiary horn

7 exhauster

8 single roll crusher

9 circulation cooler

10 device for determining the thickness of the material layer

101 mixture analyzer

102 drum feeder controller

103 roller type switchgear controller

104 auxiliary gate controller

105 sintering trolley controller

106 central processing unit.

DESCRIPTION OF IMPLEMENTATION OPTIONS

[0065] The following is a detailed description of embodiments of the invention and examples of embodiments of the invention shown in the accompanying drawings. When describing the accompanying drawings, the same reference numbers are used throughout the various accompanying drawings to designate the same or similar elements unless otherwise indicated. The present invention is not limited to the embodiments specified in this description. These embodiments are examples of the system and method that are described in detail in the claims and are consistent with certain aspects of the present invention.

[0066] The sintering system preferably includes a plurality of equipments such as a sintering car, a mixer, an exhauster drive shaft and a circulation cooler, and the general process of the sintering system is shown in FIG. 1. Various raw materials are prepared in the material pre-processing workshop 1 to form a mixed material. After entering the mixer 2 for uniform mixing and granulation, the mixed material is evenly distributed on the sintering car 5 using the drum feeder 3 and the roller type distribution device 4 to form a layer of mixed material. Next, the incendiary forge 6 and exhauster 7 are started to heat the mixed material and begin the sintering process. The agglomerated ore obtained after completion of agglomeration is crushed by single roll crusher 8, enters circulation cooler 9 for cooling, and finally delivered to the blast furnace or finished ore storage bin after screening and granulation.

[0067] In order to timely determine the thickness of the material layer on the sintering car 5 according to the current state of material distribution and realize real-time control of the material distribution on the sintering car 5, a material distribution control system based on predicting the thickness of the material layer is presented. In fig. 1 is a schematic structural view of a material distribution control system based on material layer thickness prediction shown in an embodiment of the present invention. The material distribution control system includes a drum feeder 3, a roller type distribution device 4 and a sintering cart 5, wherein the drum feeder 3 is designed to supply mixed material to the roller type distribution device 4, and the roller type distribution device 4 is configured to supply material to the sintering cart 5.

[0068] The material layer thickness prediction system further includes a mixture analyzer 101 including a sampler and a self-contained analyzer. The sampler is configured to receive granular mixed material coming from the sintering system and input the resulting mixed material into an off-line analyzer. The autonomous analyzer is designed to measure the density of all components of the detected samples and the bulk density of the mixed material. The mixture analyzer 101 performs step S201, as a result of which the obtained bulk density value of the mixed material is sent to the central processing unit 106.

[0069] It should be noted that the mixture bulk density and the mixed material bulk density referred to in the present invention have the same meaning, namely the mixed material bulk density.

[0070] The system also includes a drum feeder controller 102. The drum feeder controller 102 is coupled to the drum feeder 3 and is configured to control the rotation speed of the drum feeder 3 and determine a value for the rotation speed of the drum feeder 3. For example, a rotation speed sensor. A rotation speed sensor converts the rotation speed of a rotating object into a frequency of change of electrical current, so, for example, measuring speed by digital-to-analog conversion can be done using a magnetically sensitive rotation speed sensor or a laser rotation speed sensor, or by installing an encoder at the rotating end of the object, or using high frequency position sensor. The drum feeder controller 102 is further configured to perform step S202, whereby the obtained drum rotation speed value is sent to the central processing unit 106.

[0071] In FIG. 2 is a block diagram of a material distribution control system based on material layer thickness prediction provided in an embodiment of the present invention.

[0072] The system includes a roller-type distributor controller 103. The roller-type distributor controller 103 is connected to the roller-type distributor 4 and is configured to control the rotation speed of the distribution roller of the roller-type distributor 4 and determine a value of the rotation speed of the roller-type distributor 4 4. The roller-type distributor controller 103 is further configured to perform step S203, whereby the obtained rotation speed value of the distributor roller is sent to the central processing unit 106.

[0073] The system includes a auxiliary gate controller 104. The auxiliary gate controller 104 is located above the drum feeder 3 and is configured to control the opening degree of the auxiliary gate of the roller type dispenser 4 and determine the opening degree value of the auxiliary gate of the roller type distributor 4. Sub gate controller 104 configured to carry out step S204, as a result of which the obtained value of the opening degree of the auxiliary shutter is sent to the central processing unit 106.

[0074] It should be noted that under each storage bin there are at least four auxiliary gates. In an embodiment of the present invention, the number of auxiliary gate controllers 104 used is identical to the number of auxiliary gates, and one auxiliary gate controller 104 controls the opening degree of one auxiliary gate.

[0075] The system includes a sintering cart controller 105. The sintering cart controller 105 is located on the sintering cart 5 and is configured to control the operating speed of the sintering cart 5 and determine the operating speed of the sintering cart 5. The sintering cart controller 105 is configured to perform step 205, at as a result of which the obtained speed value of the sintering cart is sent to the central processing unit 106.

[0076] The system includes a central processing unit 106. The central processing unit is connected to a mixture analyzer 101, a drum feeder controller 102, a roller type distributor controller 103, an auxiliary gate controller 104, and a sinter cart controller 105.

[0077] It should be noted that steps S201 to S205 are not in a specific order and may be performed in any order or simultaneously.

[0078] The central processing unit 106 is configured to perform steps S206 to S213.

[0079] In step S206, the value of the bulk density of the mixed material, the rotation speed of the drum, the rotation speed of the distribution roller, the opening degree of the auxiliary gate and the speed of the sintering car are determined.

[0080] In step S207, the predicted value of the thickness of the material layer is determined in accordance with the values of the bulk density of the mixed material, the rotation speed of the drum, the rotation speed of the distribution roller, the opening degree of the auxiliary gate and the speed of the sintering car.

[0081] With reference to FIG. 3, which is a flowchart for predicting the thickness of a material layer in a material distribution control system provided in an embodiment of the present invention, the prediction method includes the following steps.

[0082] In step S301, the values of the bulk density of the mixed material, the drum rotation speed, the distribution roller rotation speed and the sintering car speed are simultaneously quantified at the same time interval in accordance with the predetermined compression ratio, and in sum with the value of the auxiliary shutter opening degree, form characteristic vector of the predicted thickness of the material layer.

[0083] The data volume and data type of the bulk density of the mixed material, the drum rotation speed, the distribution roller rotation speed, the opening degree of the auxiliary gate and the speed of the sintering car at the same time are different, and therefore these data volumes cannot be quickly calculated. In an embodiment of the present invention, the values of the bulk density of the mixed material, the drum rotation speed, the distribution roller rotation speed and the sintering car speed are quantified at the same time in the same interval according to a predetermined compression ratio, the interval being ( 0, 1).

[0084] Thus, the ratio of the bulk density value to the highest raw material density value among all components is calculated; the ratio of the drum rotation speed to the nominal drum rotation speed is calculated; the ratio of the rotation speed of the distribution roller to the nominal rotation speed of the distribution roller is calculated; and the ratio of the speed of the sintering cart to the maximum turning speed of the sintering cart is calculated.

[0085] The mathematical model for determining bulk density is expressed as: the mathematical model for determining the drum rotation speed is expressed as: the mathematical model for determining the rotation speed of the distribution roller is expressed as and the mathematical model for determining the speed of the sintering cart is expressed as .

[0086] Norm(ρ) represents the quantity determined bulk density, ρ represents the bulk density, and represents the density of the component with the highest density among all components; Norm(n ₁ ) represents the amount of rotation speed of the drum, n ₁ represents the rotation speed of the drum, and represents the nominal rotation speed of the drum; Norm(n ₂ ) represents the amount of rotation speed of the distribution roller, n ₂ represents the rotation speed of the distribution roller, and represents the nominal rotation speed of the distribution roller; and Norm(s) is the amount of the specified speed of the sintering cart, v is the speed of the sintering cart, and represents the maximum speed of the sintering car.

[0087] Predicting the characteristic vector of the thickness of a material layer consists of integrating the factors affecting the thickness of the material layer in accordance with a certain rule, such as:

[0088]

[0089] where X (k) represents the characteristic vector of the material layer thickness, x ₁ (k), x ₂ (k), x ₃ (k), x ₄ (k) and x ₅ (k) respectively represent the bulk density, speed drum rotation, distribution roller rotation speed, auxiliary gate opening degree and sintering car speed.

[0090] In step S302, the predicted material layer thickness characteristic vector is input into a predetermined dynamic material layer thickness prediction model to generate a material layer thickness characteristic value, wherein the dynamic material layer thickness prediction model includes displaying a relationship between the predicted material layer thickness characteristic vector and characteristic value of the thickness of the material layer.

[0091] The relationship display includes a predicted material layer thickness characteristic vector and a corresponding material layer thickness characteristic value, namely

[0092] y(k)=f _k (x (k)),

[0093] where y(k) is a characteristic value affecting the material layer thickness, af _k is a display of the relationship between the predicted material layer thickness characteristic vector and the material layer thickness characteristic value.

[0094] In step S303, the material layer thickness characteristic value data is restored to obtain a predicted material layer thickness value.

[0095] The final predicted value of the material layer thickness can be obtained by performing certain data processing on the characteristic value output by the dynamic material layer thickness prediction model. In particular, the product of the characteristic value of the thickness of the material layer and the maximum permissible thickness of the material layer on the sintering car is calculated to obtain the predicted value of the thickness of the material layer.

[0096] In particular, data recovery is performed from the characteristic value of the material layer thickness using the following method:

[0097] h _i =k _i × N,

[0098] where h _i is the predicted material layer thickness value generated at time i, k _k is the characteristic material layer thickness value generated at time i, and H is the maximum allowable material layer thickness on the sintering car.

[0099] In step S208, a material layer thickness deviation amount is calculated in accordance with the predicted material layer thickness value and the target material layer thickness value.

[00100] E(k)=R(k)-Y(k),

[00101] where Y(k) is the sequence of the predicted value of the thickness of the material layer obtained at time k, R(k) is the sequence of the target value of the thickness of the material layer obtained at time k, and E(k) is the sequence of the thickness deviation value layer of material at time k.

[00102] In step S209, the deviation amount of the material layer thickness is input into the moving average optimization model to obtain the values of the drum rotation speed to be adjusted, the distribution roller rotation speed to be adjusted, the auxiliary gate opening degree to be adjusted, and the sintering car speed to be adjusted. adjustment.

[00103] The moving average optimization model is configured to calculate the drum rotation speed, the distribution roller rotation speed, the auxiliary gate opening degree and the sintering car speed, in accordance with the minimum standard deviation of the thickness of the material layer, subject to a constant value of the bulk density of the mixed material.

[00104] The standard deviation of the material layer thickness is determined as follows:

[00105]

[00106] where α is the standard deviation of the material layer thickness, E(k) is the deviation value of the material layer thickness at time k, R(k) is the specified sequence of the target value of the material layer thickness at time k, and Y(k) is the sequence of the predicted values of the thickness of the material layer at moment k.

[00107] Using the k-th moment as an example, the central processing unit 106 obtains the values of the bulk density of the mixed material, the drum rotation speed, the distribution roller rotation speed, the opening degree of the auxiliary gate, and the speed of the sintering car at the k-th moment and calculates the sequence of the predicted value of the layer thickness material at the kth moment:

[00108]

[00109] where f _k represents a display of the relationship between the predicted characteristic vector of the thickness of the material layer and the characteristic value of the thickness of the material layer at the k-th moment; and at the kth moment, the predicted sequence of the magnitude of the deviation is expressed as follows:

[00110]

[00111] At the kth moment, the standard deviation of the material layer thickness is expressed as follows:

[00112]

[00113] Equation is substituted into the above equation to calculate the standard deviation, and thus the values of drum rotation speed, distribution roller rotation speed, auxiliary gate opening degree and sintering car speed can be calculated to be adjusted according to the standard deviation value σ using the optimization model moving average, provided that the bulk density of the mixed material remains constant.

[00114] The method for controlling the rotation speed of the drum, the rotation speed of the distribution roller, the opening degree of the auxiliary gate, and the speed of the sintering car at the kth moment is described above. The control methods at the (k+1)th moment,....., (k+j)th moment are similar, and their descriptions are not given here.

[00115] After the central processing unit 106 obtains the values of the drum rotation speed, the distribution roller rotation speed, the auxiliary gate opening degree and the sintering car speed to be adjusted, the following steps S210 to S213 are performed.

[00116] In step S210, the speed value of the sintering car to be adjusted is sent. The sintering cart controller 105, receiving the speed value of the sintering cart to be adjusted, is driven to adjust the speed value of the sintering cart to be adjusted.

[00117] In step S211, the auxiliary gate opening degree value to be adjusted is sent. The auxiliary shutter controller 104, receiving the auxiliary shutter opening degree value to be adjusted, is driven to adjust the auxiliary shutter opening degree value to be adjusted.

[00118] In step S212, the distribution roller rotation speed value to be corrected is sent. The roller-type distribution device controller 103, receiving the rotation speed value of the distribution roller to be adjusted, is driven.

[00119] In step S213, the drum rotation speed value to be adjusted is sent. The drum feeder controller 102, receiving the drum rotation speed value to be adjusted, is driven.

[00120] It should be noted that steps S210 to S213 may be performed in any order or simultaneously. However, the steps in which the speed of the sintering car is adjusted according to the sintering speed to be adjusted, the opening degree of the auxiliary gate is adjusted according to the opening degree of the auxiliary gate to be adjusted, the rotation speed of the distribution roller is adjusted according to the rotation speed of the distribution roller, to be adjusted, and the drum rotation speed is adjusted in accordance with the drum rotation speed to be adjusted, must be performed simultaneously.

[00121] The system includes a central processing unit 106 coupled to a mixture analyzer 101, a drum feeder controller 102, a roller type distributor controller 103, an auxiliary gate controller 104, and a sinter cart controller 105.

[00122] Known from the above-mentioned technical solution, an embodiment of the present invention provides a material distribution control system based on prediction of the material layer thickness. In practice, the values of bulk density of the mixed material, drum rotation speed, distribution roller rotation speed, auxiliary gate opening degree and sintering car speed are first determined. Then, the values of bulk density, drum rotation speed, distribution roller rotation speed and sintering cart speed are simultaneously calculated in the same time interval in accordance with a given compression ratio and, in sum with the value of the degree of opening of the auxiliary shutter, form the predicted characteristic vector of the thickness of the material layer. Then, the predicted material layer thickness characteristic vector is input into the preset dynamic material layer thickness prediction model to obtain the material layer thickness characteristic value. Next, the characteristic value of the material layer thickness data is restored to obtain the predicted value of the material layer thickness. Then, the material layer thickness deviation amount is calculated according to the predicted material layer thickness value and the target material layer thickness value. Finally, the deviation value of the material layer thickness is input into the moving average optimization model to obtain the values of the drum rotation speed to be adjusted, the distribution roller rotation speed to be adjusted, the auxiliary gate opening degree to be adjusted, and the sintering car speed to be adjusted, and, Thus, control of the material distribution in the sintering system is achieved based on predicting the thickness of the material layer. According to the material distribution control system based on material layer thickness prediction provided in an embodiment of the present invention, the thickness of the material layer on the sintering car can be determined in advance by obtaining the mixed material parameters and the state parameters of the sintering system, so that accurate material distribution control is achieved sintering systems in accordance with the predicted thickness of the material layer.

[00123] It should be noted that the mixed material parameters include the bulk density of the mixed material, and the sintering system state parameters include the rotation speed of the drum, the rotation speed of the distribution roller, the opening degree of the auxiliary gate, and the speed of the sintering car. The thickness of the material layer on the sintering cart is influenced by many parameters, such as the degree of opening of the main gate, the angle of inclination of the drum, the width of the sintering cart and the number of auxiliary gates. However, parameters such as the width of the sintering car and the number of auxiliary gates are usually constant; Moreover, parameters such as the degree of opening of the main shutter and the angle of inclination of the drum are variable, but cannot change frequently, and therefore remain constant in the long term. Therefore, in the material distribution control system based on material layer thickness prediction in the embodiment of the present invention, these influencing parameters are not taken into account. That is, the indicated influencing parameters do not change in real application and when creating a dynamic model for predicting the thickness of the material layer.

[00124] A dynamic material layer thickness prediction model is set in advance according to the actual production data of the sintering system equipment. In the application stage, the dynamic material layer thickness prediction model can be used in all sintering systems, and there is a certain difference between the actual operating conditions of different sintering systems. Over time, the operating state of the same sintering system may also change to a certain extent during long-term use. In this case, if the dynamic material layer thickness prediction model is not adjusted adaptively, a situation may arise where there is a large deviation between the predicted value of the material layer thickness and the actual result. In order to overcome such a technical problem, in some embodiments of the present invention, the material layer thickness prediction system further includes a material layer thickness determining device 10 capable of determining the thickness of the material layer on the sintering car in real time. A material layer thickness determining device 10 is located above the sintering cart to determine the thickness of the material layer on the sintering cart. For example, a level indicator can be used to determine the thickness of a layer of material.

[00125] As shown in FIG. 4, some embodiments of the present invention provide a method for updating a dynamic material layer thickness prediction model in real time.

[00126] Before step S302, in which the predicted material layer thickness characteristic vector is input into a preset dynamic material layer thickness prediction model to generate a material layer thickness characteristic value, the method further includes the following steps.

[00127] In step S401, time sequence training samples approximating the predicted material layer thickness characteristic vector are obtained, wherein the training samples include input samples and a material layer thickness characteristic value corresponding to the input samples.

[00128] It should be noted that samples and training samples referred to in the present invention have the same meaning and refer to training samples.

[00129] The training samples include not only the mixed material parameters of the sintering system and the state parameters of the sintering system, but also include a corresponding measured material layer thickness value. A specific method for obtaining training samples is as follows: training samples are obtained in a time sequence approximating the predicted characteristic vector of the material layer thickness, wherein the training samples include input samples and a characteristic value of the material layer thickness corresponding to the input samples. It should be noted that the measured value of the material layer thickness can be adjusted to update or train the dynamic material layer thickness prediction model after the parameters are quantified, and the normalization is performed by adopting the following model:

[00130]

[00131] where h _i is the predicted value of the material layer thickness obtained at time i, k _i is the characteristic value of the material layer thickness obtained at time i, and H is the maximum permissible material layer thickness on the sintering car.

[00132] The resulting statistical predictive samples are used as training samples. The sintering process of the sintering system is a long continuous process, the sintering system performs the mixing operation all the time until the predicted characteristic vector of the material layer thickness is obtained. By obtaining statistical predicted samples during a given time interval, close to the predicted characteristic vector of the material layer thickness, it is possible to guarantee that the operating state of the sintering system corresponding to the obtained training samples corresponds to the operating state of the sintering system at the time of obtaining the predicted characteristic vector of the material layer thickness.

[00133] In step S402, the dynamic material layer thickness prediction model is updated in real time using the training samples to obtain an updated dynamic material layer thickness prediction model.

[00134] With the help of the training samples obtained in step S401, the dynamic material layer thickness prediction model can be updated in real time, and thus it is further ensured that the characteristic value of the material layer thickness output by the dynamic material layer thickness prediction model is is more accurate. In a particular updating method, a predicted amount of deviation between the predicted value of the material layer thickness and the characteristic value of the material layer thickness corresponding to the measured value is determined in accordance with the characteristic value of the material layer thickness predicted by the training samples. There are two ways to update a dynamic forecasting model. If the predicted deviation value is small, relatively constant, and within the acceptable error range of the model quantity, the predicted deviation value is directly added to the material layer thickness characteristic value of the dynamic prediction model, and the result is used as an updated material layer thickness characteristic value. If the predicted deviation amount is large and the dependency display is determined to change according to the model score, the dependency display included in the dynamic material layer thickness prediction model is updated according to the training samples and the characteristic value of the material layer thickness corresponding to the measured value.

[00135] It should be noted that the standard deviation can be taken as a quantitative indicator. The statistical confidence limit is set in accordance with the statistical distribution rule of the quantitative indicator, and thus the need to initiate an update and the required update method are determined. The characteristic value of the material layer thickness corresponding to the measured value is not suitable for solving the closed-loop control problem in the mixing process due to the time lag, but can be used as a reference value. That is, when the sintering system reaches a stable state at steady state, the thickness of the material layer can be maintained at a certain level. Although the measured value fluctuates under normal conditions, the distribution of the measured value generally does not deviate from the confidence interval, and when the distribution exceeds the confidence interval, the model update mechanism is triggered. If it is determined from the results of the index analysis that the process characteristic is gradually changing, the recurrent method for estimating the model parameters is selected, and the recurrent identification in the moving window is used to update the dynamic material layer thickness prediction model according to the following steps.

[00136] The set of samples of the initial dynamic forecasting model is specified as where t is the total number of samples. When a new measured value [X ^m ,Y ^m ] is determined, the new measured value [X ^m ,Y ^m ] is added to the sample set and the original sample is discarded, resulting in a new sample set (training samples) expressed as:

[00137]

[00138] The dynamic material layer thickness prediction model is then updated in real time using the training samples to obtain an updated dynamic material layer thickness prediction model. If the results of the indicator analysis determine that the process characteristic changes suddenly, the just-in-time method is selected to restore the dynamic model for predicting the thickness of the material layer.

[00139] A dynamic material layer thickness prediction model is established using a plurality of groups of known mixed material parameters, sintering system state parameters, and corresponding measured material layer thickness values. An embodiment of the present invention provides a method for creating a dynamic material layer thickness prediction model by training a neural network model. Specifically, the neural network model is trained by taking multiple groups of known mixed material parameters and sintering system state parameters as inputs to the neural network model and corresponding measured material layer thickness values as outputs of the neural network model. That is, the weight matrices and bias conditions corresponding to the intermediate layers of the neural network model are adjusted, and thus the relationship mapping between the parameters of the mixed material, the corresponding parameters of the mixer state and the homogeneity of mixing is established. In fig. 5 is a flowchart for constructing a dynamic material layer thickness prediction model according to an embodiment of the present invention, including the following specific steps.

[00140] In step S501, N groups of independent values of the bulk densities of the mixed material, the drum rotation speed, the distribution roller rotation speed, the auxiliary gate opening degree, and the sintering car speed are determined.

[00141] The N groups of independent parameters of the mixed material and the corresponding state parameters of the sintering system may be the data of the same sintering system or the data of multiple sintering systems, and are divided into the same group according to the corresponding ratio, that is, the data of the same mixer in the same moment are used by the same group of data.

[00142] In step S502, N groups of independent values of volumetric densities, drum rotation speeds, distribution roller rotation speeds, and sintering car speeds are quantified in the same interval in accordance with a given compression ratio, and in sum with the value of the auxiliary gate opening degree, the input data of N groups of training samples, while the bulk density in the parameters of the mixed material, as well as the rotation speed of the drum, the rotation speed of the distribution roller and the speed of the sintering cart in the parameters of the state of the sintering system should be expressed quantitatively in the interval (0, 1), and the value of the degree of opening auxiliary shutter is expressed as a percentage.

[00143] In step S503, N groups of independent material layer thickness measured values are obtained from the material layer thickness determining device 10, and the measured material layer thickness values are quantified according to the determined compression ratio to generate output data of the N groups of training samples.

[00144] The measured thickness values of the material layer are quantified in accordance with the determined compression ratio. In particular, the ratio of the measured value of the thickness of the material layer and the maximum permissible thickness of the material layer on the sintering car is calculated.

[00145] In step S504, the input data of the training samples and the output data of the training samples are used to train the neural network model using the backpropagation method.

[00146] The dynamic prediction model uses the input data of the training samples and the output data of the training samples to train the neural network model by applying backpropagation, where backpropagation is a learning algorithm suitable for a multilayer neural network that directs the response (output ) multilayer neural network inputs to achieve a given target range through repeated cyclic iterations of signal transmission and modification of weights.

[00147] In step S505, the weight parameter, bias parameter, and learning coefficient of the neural network model are continuously updated through iterative learning.

[00148] In step S506, if the predicted characteristic value of the material layer thickness of the neural network model and the measured characteristic value of the material layer thickness reach a predetermined error limit, or if the neural network model reaches a predetermined maximum number of iterations, the training is completed and the weight parameter is finally updated, The bias parameter and learning coefficient are stored to obtain a dynamic material layer thickness prediction model.

[00149] By continuously updating the weight parameter, bias parameter, and learning coefficient of the neural network model, a relationship mapping between the predicted material layer thickness characteristic vector and the material layer thickness characteristic value in the neural network model is established until the characteristic value of the material layer thickness specified by the dependence mapping will not meet production needs. The specific determination method includes: determining whether the predicted material layer thickness characteristic value of the neural network model and the obtained material layer thickness characteristic value reach a predetermined error limit or whether the neural network model reaches a predetermined maximum number of iterations; If the predicted characteristic value of the material layer thickness of the neural network model and the obtained characteristic value of the material layer thickness reach the specified error limit or the neural network model reaches the specified maximum number of iterations, the training ends and the finally updated weight parameter, bias parameter and learning coefficient are saved to obtain the dynamic models for predicting material layer thickness.

[00150] The dynamic material layer thickness prediction model created by training a neural network model, in an embodiment of the present invention may, in particular, use an RNN (recurrent neural network) or LSTM (long short-term memory) neural network model. Fig. 6 is a schematic diagram of the structure of an LSTM neural network model presented in an embodiment of the present invention, in which the structure of each intermediate layer is mainly composed of a forgetting gate, an input gate and an output gate, and each gate has a corresponding weight matrix and bias terms. The multilayer neural network is trained using training samples, while the weight parameter, bias parameter and learning coefficient are constantly updated, and thus a dynamic model for predicting the thickness of the material layer is formed.

[00151] In a specific process of training a neural network model in an embodiment of the present invention, the neural network model divides a plurality of training sample groups into two parts, namely 2/3 training data groups and 1/3 test data groups, wherein the input and output data 2/3 groups of training samples are used as the training data of the neural network model to continuously update the weight parameter, bias parameter and learning coefficient, and the input and output data of 1/3 group of test data are used as the test data of the permissible error of the neural network model.

[00152] With reference to FIG. 7, which is a flowchart for constructing another dynamic material layer thickness prediction model in an embodiment of the present invention, the dynamic material layer thickness prediction model can be supplemented with a mixing uniformity prediction table, which is set according to the following steps:

[00153] In step S701, N groups of independent values of volumetric densities of the mixed material, drum rotation speed, distribution roller rotation speed, auxiliary gate opening degree, and sintering car speed are determined;

[00154] In step S702, N groups of independent material layer thickness measured values are obtained from the material layer thickness detecting device; and

[00155] In step S703, N groups of independent values of the bulk densities of the mixed material, the drum rotation speed, the distribution roller rotation speed, the auxiliary gate opening degree and the sintering car speed, as well as the corresponding measured values of the material layer thickness, are statistically analyzed to create a thickness prediction table layer of material.

[00156] It should be noted that the layer thickness prediction table and the material layer thickness prediction table mentioned in the present application have the same meaning and refer to the material layer thickness prediction table.

[00157] Statistical analysis involves data pre-processing. During data pre-processing, the bulk density, drum rotation speed, distribution roller rotation speed and sintering car speed are first quantified in the same interval according to the specified compression ratio. Then, multiple data groups in which the mixed material parameters and sintering system state parameters are the same are extracted, and different groups of corresponding measured material layer thickness values are compared. If the corresponding measured material thickness values also match, only one group of data is retained and other groups of the same data are deleted. If the corresponding mixing uniformity is different, but the deviation is within the acceptable range, the average mixing uniformity value corresponding to multiple data groups is used as the corresponding material layer thickness, and only one data group is stored, and the acceptable deviation range is set to 0.5% -2% depending on manufacturing accuracy requirements. For a plurality of data groups in which the mixed material parameters and the sintering system state parameters are the same, the corresponding measured mixing uniformity values are different, and the deviation exceeds the acceptable range, the sources of the plurality of data groups are marked and stored in an isolated area. The data sources described here refer to the sintering system corresponding to the data collection and collection time, and the isolation area refers to the independent area in the layer thickness prediction table, which is configured to record abnormal data as reference data for equipment overhaul.

[00158] It should be clarified that the measured values of the thickness of the material layer also require preliminary processing, in particular, the ratio of the measured value of the thickness of the material layer and the maximum permissible thickness of the material layer on the sintering car is calculated.

[00159] A plurality of independent parameter groups of the mixed material, state parameters of the sintering system, and corresponding measured material layer thicknesses that are subject to data preprocessing are ordered according to a plurality of predetermined index scores, and all index scores are ordered from smallest to largest. There are at least five index indicators. For example, in an embodiment of the present invention, five index values and index orders are established, where the first level index value represents the bulk density of the mixed material, the second level index value represents the rotation speed of the drum, the third level index value represents the rotation speed of the distribution roller, the index value the fourth level represents the opening degree of the auxiliary gate, and the fifth level index value represents the speed of the sintering car. A mixing uniformity prediction table is created in accordance with the above rule, wherein the mixing uniformity prediction table includes displaying a relationship between a predicted material layer thickness characteristic vector and a material layer thickness characteristic value. For example, some mixing uniformity prediction table data is shown in Table 1.

[00160]

[00161] The following description illustrates an embodiment of the method according to the present invention. Details not disclosed in the embodiment of the method of the present invention are indicated in the above embodiment of the system of the present invention.

[00162] An embodiment of the present invention provides a method for controlling material distribution based on prediction of the thickness of a layer of material, wherein the method for controlling material distribution includes the following steps:

[00163] determining the values of the bulk density of the mixed material, the rotation speed of the drum, the rotation speed of the distribution roller, the opening degree of the auxiliary gate and the speed of the sintering car;

[00164] predicting the thickness of the material layer according to the values of the bulk density of the mixed material, the rotation speed of the drum, the rotation speed of the distribution roller, the opening degree of the auxiliary gate and the speed of the sintering car to obtain a predicted value of the thickness of the material layer,

[00165] characterized in that the step of predicting the thickness of the material layer in accordance with the values of the bulk density of the mixed material, the rotation speed of the drum, the rotation speed of the distribution roller, the degree of opening of the auxiliary gate and the speed of the sintering car to obtain the predicted value of the thickness of the material layer includes:

[00166] quantifying the values of the bulk density of the mixed material, the rotation speed of the drum, the rotation speed of the distribution roller and the speed of the sintering car at the same time in the same time interval according to a given compression ratio and in sum with the value of the opening degree of the auxiliary gate, obtaining a predicted characteristic vector thickness of the material layer;

[00167] inputting the predicted material layer thickness characteristic vector into a predefined dynamic material layer thickness prediction model to obtain a material layer thickness characteristic value, wherein the dynamic material layer thickness prediction model includes displaying a relationship between the predicted material layer thickness characteristic vector and the characteristic thickness value layer of material;

[00168] performing data recovery of the characteristic value of the material layer thickness to obtain a predicted value of the material layer thickness;

[00169] calculating a material layer thickness deviation amount in accordance with the predicted material layer thickness value and the target material layer thickness value;

[00170] inputting the deviation amount of the material layer thickness into the moving average optimization model to obtain the values of the drum rotation speed to be adjusted, the distribution roller rotation speed to be adjusted, the auxiliary gate opening degree to be adjusted, and the sintering car speed to be adjusted.

[00171] Adjusting the drum rotation speed in accordance with the drum rotation speed to be adjusted, the distribution roller rotation speed in accordance with the distribution roller rotation speed to be adjusted, the auxiliary gate opening degree in accordance with the auxiliary gate opening degree to be adjusted, the sintering car speed according to the speed of the sintering car to be adjusted,

[00172] In this case, the moving average optimization model is configured to calculate the drum rotation speed, the distribution roller rotation speed, the degree of opening of the auxiliary gate and the speed of the sintering cart, in accordance with the minimum standard deviation of the thickness of the material layer, subject to a constant value of the bulk density of the mixed material.

[00173] Known from the above-mentioned technical solutions, embodiments of the present invention provide a system and method for controlling material distribution based on predicting the thickness of a layer of material. The material distribution control system includes a drum feeder 3, a roller type distribution device 4 and a sintering cart 5. The drum feeder 3 is configured to supply mixed material to the roller type distribution device 4, and the roller type distribution device 4 is configured to supply material to the sintering cart 5 The material distribution control system further includes a mixture analyzer 101, a drum feeder controller 102 connected to the drum feeder 3, a roller type distributor controller 103 and an auxiliary gate controller 104 connected to the roller type distributor 4, a sintering car controller 105 connected with a sintering cart 5, and a central processing unit 106 connected to a mixture analyzer 101, a drum feeder controller 102, a roller type distributor controller 103, an auxiliary gate controller 104 and a sintering cart controller 105.

[00174] In practice, the values of bulk density of the mixed material, drum rotation speed, distribution roller rotation speed, auxiliary gate opening degree, and sintering car speed are first calculated. Then, the values of bulk density, drum rotation speed, distribution roller rotation speed, auxiliary gate opening degree and sintering car speed are quantified simultaneously in the same time interval according to the specified compression ratio to obtain the predicted characteristic vector of the material layer thickness. The predicted material layer thickness characteristic vector is then input into a predefined dynamic material layer thickness prediction model to generate a material layer thickness characteristic value. The characteristic value of the material layer thickness data is reconstructed to obtain the predicted value of the material layer thickness. Then, the material layer thickness deviation amount is calculated in accordance with the predicted material layer thickness value and the target material layer thickness value. Finally, the deviation amount of the material layer thickness is input into the moving average optimization model to obtain the values of the drum rotation speed to be adjusted, the distribution roller rotation speed to be adjusted, the auxiliary gate opening degree to be adjusted, and the sintering car speed to be adjusted, and Thus, control of the distribution of material in the sintering system is achieved based on predicting the thickness of the material layer. According to the material distribution control system based on material layer thickness prediction provided in the embodiment of the present invention, the thickness of the material layer on the sintering car can be predicted in advance by obtaining the mixed material parameters and the state parameters of the sintering system, so that the sintering system achieves timely and stable feedback control of material distribution according to predicted material layer thickness.

[00175] Similar parts in the embodiments presented in the present invention may refer to each other. The specific embodiments provided as stated above are just a few examples in accordance with the general concept of the present invention, but do not limit the scope of protection of the present invention. Any other embodiments implemented by those skilled in the art in accordance with the embodiments of the present invention without the introduction of creative efforts fall within the scope of protection of the present invention.

Claims (52)

1. A system for controlling the distribution of powdered material during sintering in a sintering system, which includes a drum feeder, a roller-type distributor device having an auxiliary gate, and a sintering cart, wherein the drum feeder is configured to supply the mixed material to said distributor device, which configured to supply material to a sintering cart containing a mixture analyzer, a drum feeder controller connected to the drum feeder, a roller-type distribution device controller and an auxiliary gate controller connected to the roller-type distribution device, a sintering cart controller connected to the sintering cart, and a central a processing unit connected to a mixture analyzer, a drum feeder controller, a roller type distributor controller, an auxiliary gate controller and a sintering cart controller, wherein the central processing unit is configured to: - obtaining the values of bulk density of the mixed material from the mixture analyzer, drum rotation speed from the drum feeder controller, distribution roller rotation speed from the roller type distributor controller, auxiliary gate opening degree from the auxiliary gate controller and sintering cart speed from the sintering cart controller, - predicting the thickness of the material layer in accordance with the values of the bulk density of the mixed material, drum rotation speed, distribution roller rotation speed, degree of opening of the auxiliary gate and the speed of the sintering cart to determine the predicted value of the material layer thickness, - calculating the deviation value of the thickness of the material layer in accordance with the predicted value of the thickness of the material layer and the target value of the thickness of the material layer, - inputting the deviation value of the thickness of the material layer into the optimization model to obtain the values of the drum rotation speed to be adjusted, the distribution roller rotation speed to be adjusted, the degree of opening of the auxiliary gate to be adjusted, and the sintering cart speed to be adjusted, and - actuating the drum feeder controller to regulate the speed of rotation of the drum in accordance with the speed of rotation of the drum to be adjusted, actuating the roller type distribution device controller to regulate the rotation speed of the distribution roller in accordance with the rotation speed of the distribution roller to be adjusted, driving the auxiliary shutter controller to adjust the opening degree of the auxiliary shutter in accordance with the opening degree of the auxiliary shutter to be adjusted, actuating the sintering cart controller to regulate the speed of the sintering cart in accordance with the speed of the sintering cart to be adjusted, wherein the optimization model is made with the ability to calculate the rotation speed of the drum, the rotation speed of the distribution roller, the degree of opening of the auxiliary gate and the speed of the sintering cart, in accordance with the minimum standard deviation of the thickness of the material layer, subject to a constant value of the bulk density of the mixed material. 2. The system according to claim 1, characterized in that the central processing unit is configured to determine the standard deviation of the thickness of the material layer according to the formula where σ is the standard deviation of the material layer thickness, E(k) is the deviation value of the material layer thickness, R(k) is the specified sequence of the target value of the material layer thickness, and Y(k) is the sequence of the predicted value of the material layer thickness. 3. The system according to claim 1, characterized in that in order to obtain a predicted value of the thickness of the layer of material in accordance with the values of the bulk density of the mixed material, the rotation speed of the drum, the rotation speed of the distribution roller, the degree of opening of the auxiliary gate and the speed of the sintering cart, the central processing unit is made with possibility: - quantitative determination of the volumetric density of the mixed material, drum rotation speed, distribution roller rotation speed and sintering cart speed simultaneously in the same time interval in accordance with a given compression ratio and in sum with the value of the degree of opening of the auxiliary gate to obtain the predicted characteristic vector of the layer thickness material, - inputting the predicted characteristic vector of the material layer thickness into a predefined dynamic material layer thickness prediction model to obtain the characteristic value of the material layer thickness, wherein the dynamic material layer thickness prediction model includes displaying the relationship between the predicted characteristic vector of the material layer thickness and the characteristic value of the material layer thickness , And - performing data recovery of the characteristic value of the thickness of the material layer to obtain the predicted value of the thickness of the material layer. 4. The system according to claim 3, characterized in that it additionally contains a device for determining the thickness of the material layer located above the sintering cart, which is connected to the central data processing unit and is configured to determine the thickness of the material layer on the sintering cart in accordance with a specified time interval and obtaining the measured value of the thickness of the layer of material, while for inputting the predicted material layer thickness characteristic vector into the preset dynamic material layer thickness prediction model to generate the material layer thickness characteristic value, the central processing unit is further configured to: - obtaining time-series data approximating the predicted material layer thickness characteristic vector as samples, the samples including input samples and a material layer thickness characteristic value corresponding to the input samples, and - updating the dynamic model for predicting the thickness of the material layer in real time using training samples to obtain an updated dynamic model for predicting the thickness of the material layer. 5. The system according to claim 3, characterized in that to restore data on the characteristic value of the material layer thickness, the central processing unit is configured to determine the predicted value of the material layer thickness using the formula where h _i is the predicted value of the thickness of the material layer obtained at time i, k _i is the characteristic value of the thickness of the material layer generated at time i, and H is the maximum permissible thickness of the material layer on the sintering car. 6. The system according to claim 3, characterized in that for quantitative determination of the values of the bulk density of the mixed material, the rotation speed of the drum, the rotation speed of the distribution roller and the speed of the sintering car simultaneously in the same time interval in accordance with a given compression ratio, the central processing unit made with the ability to: - calculating the ratio of the bulk density value to the highest density value of the raw material among all components, - calculating the ratio of the drum rotation speed to the maximum drum rotation speed, - calculating the ratio of the distribution roller rotation speed to the maximum distribution roller rotation speed, and - calculating the ratio of the speed of the sintering cart to the maximum rotation speed of the sintering cart. 7. The system according to claim 3, characterized in that the central processing unit is additionally configured to establish a dynamic model for predicting the thickness of the material layer based on training the neural network model by performing the following steps: - determination of N groups of independent values of volumetric densities of the mixed material, drum rotation speed, distribution roller rotation speed, degree of opening of the auxiliary gate and sintering cart speed in accordance with a given time interval, - quantitative determination of N groups of independent values of volumetric densities of the mixed material, drum rotation speed, distribution roller rotation speed and sintering car speed in the same interval in accordance with a given compression ratio and in sum with the value of the auxiliary gate opening degree; generation of input data N groups training samples, - determining the actual thickness of the material layer on the sintering car in accordance with the input data of N groups of training samples according to a given time interval and calculating the actual characteristic values of the actual thickness of the material layer, while the actual characteristic values serve as the output data of N groups of training samples, - use of input data of training samples and output data of training samples to train a neural network model using the backpropagation method, - continuous updating of the weight parameter, bias parameter and learning coefficient of the neural network model through iterative training, while If the predicted value and the measured value of the neural network model reach the specified error limit, or if the neural network model reaches the specified maximum number of iterations, the training is completed, and the final updated weight parameter, bias parameter, and learning coefficient are stored to obtain a dynamic material layer thickness prediction model. 8. The system according to claim 3, characterized in that the dynamic model for predicting the thickness of the material layer is a table for predicting the thickness of the material layer, and to establish the table for predicting the thickness of the material layer, the central processing unit is configured to: - determination of N groups of independent values of volume densities of the mixed material, drum rotation speed, distribution roller rotation speed, degree of opening of the auxiliary gate and sintering cart speed in accordance with a given time interval, - determining the actual thickness of the material layer on the sintering car in accordance with the input data of N groups of training samples according to a given time interval and calculating the actual characteristic values of the actual thickness of the material layer, and performing a statistical analysis of N groups of independent values of the bulk densities of the mixed material, drum rotation speed, distribution roller rotation speed, auxiliary gate opening degree and sintering car speed, as well as the corresponding actual values of the material layer thickness to create a material layer thickness prediction table. 9. A method for controlling the distribution of powdered material during its sintering in a sintering system containing a drum feeder, a roller-type distribution device having an auxiliary shutter, and a sintering cart, wherein the drum feeder is configured to supply mixed material to the roller-type distribution device, the roller-type distribution device type made with the ability to supply material to the sintering cart, including: - determination of the bulk density of the mixed material, drum rotation speed, distribution roller rotation speed, auxiliary gate opening degree and sintering cart speed, - predicting the thickness of the material layer in accordance with the values of the bulk density of the mixed material, drum rotation speed, distribution roller rotation speed, degree of opening of the auxiliary gate and the speed of the sintering cart to obtain the predicted value of the material layer thickness, - calculating the deviation value of the thickness of the material layer in accordance with the predicted value of the thickness of the material layer and the target value of the thickness of the material layer, - inputting the material layer thickness deviation value into the optimization model to obtain the value of the drum rotation speed to be adjusted, the distribution roller rotation speed to be adjusted, the auxiliary gate opening degree to be adjusted, and the sintering car speed to be adjusted, and - regulation of the drum rotation speed in accordance with the drum rotation speed to be adjusted, the distribution roller rotation speed in accordance with the distribution roller rotation speed to be adjusted, the auxiliary gate opening degree in accordance with the auxiliary gate opening degree to be adjusted, the sintering cart speed in accordance with the speed of the sintering cart subject to adjustment, in this case, the optimization model is made with the ability to calculate the rotation speed of the drum, the rotation speed of the distribution roller, the degree of opening of the auxiliary gate and the speed of the sintering cart, in accordance with the minimum standard deviation of the thickness of the material layer, subject to a constant value of the bulk density of the mixed material. 10. The method according to claim 9, characterized in that the step of predicting the thickness of the material layer in accordance with the values of the bulk density of the mixed material, the rotation speed of the drum, the rotation speed of the distribution roller, the degree of opening of the auxiliary gate and the speed of the sintering cart to obtain the predicted value of the thickness of the material layer includes: - quantitative determination of the volumetric density of the mixed material, drum rotation speed, distribution roller rotation speed, auxiliary gate opening degree and sintering cart speed simultaneously in the same time interval in accordance with a given compression ratio and in total with the auxiliary gate opening degree value obtaining predicted characteristic vector of the material layer thickness, - inputting the predicted characteristic vector of the material layer thickness into the predefined dynamic material layer thickness prediction model to obtain the characteristic value of the material layer thickness, wherein the dynamic material layer thickness prediction model includes displaying the relationship between the predicted characteristic vector of the material layer thickness and the characteristic value of the material layer thickness , And - performing data recovery of the characteristic value of the thickness of the material layer to obtain the predicted value of the thickness of the material layer.