CN110697438B - A Neural Network-Based Loss-in-Weight Material Unloader Controller - Google Patents

Abstract

The invention discloses a straight-fall loss-in-weight material unloading machine controller based on a neural network, which comprises a signal acquisition module, a processing module, a neural network module, a storage module and an output module. The neural network module predicts the weight loss value of the material based on the material level, blanking rate, material density and opening diameter of the feeding valve, so as to adjust the closing time of the feeding valve. Modeling the weighing behavior in blanking based on the neural network, the trained network of the present invention can accurately predict the weight loss value of falling materials under different blanking states, so that direct and precise blanking control can be realized and is suitable for small batches It also combines the detection of the position sensor and the control of the agitator to adjust the material accumulation form in the unloading silo, reducing the fluctuation of the blanking rate; and through the control of the cumulative error of the blanking, it reduces the total amount of batch blanking. error.

Description

A Neural Network-Based Loss-in-Weight Material Unloader Controller

This application is a divisional application with the application number 201710863074.1, the application date on September 19, 2017, and the invention title "Neural Network-Based Loss-in-Weight Feeder and Controller".

technical field

The invention relates to the field of quantitative cutting, in particular to a controller of a straight-fall loss-in-weight type material cutting machine based on a neural network.

Background technique

In industrial and agricultural manufacturing and commodity packaging, there are a large number of powder materials, such as iron ore concentrate, coal powder and other iron-making raw materials, polypropylene, polystyrene, polyvinyl chloride, light methyl cellulose, polyacrylonitrile, ring Oxygen resin powder coating and other chemical raw materials, quartz sand, cement and other building materials raw materials, daily chemical products such as washing powder, millet, soybean and other grain and legume agricultural products, or powder, slag, granular processed food, feed, fertilizer, pesticide and other agricultural production Materials, as well as powdered and granular health products, Chinese and Western medicines, condiments, etc. all require automatic quantitative packaging or ingredient manufacturing.

At present, many enterprises in my country still use manual quantitative ingredients or packaging. On the one hand, the labor intensity is high, the speed is slow, and the economic benefit is poor; It is easy to cause harm to the human body. Therefore, for production enterprises, it is urgent to provide inexpensive multi-component automatic quantitative feeding equipment or devices with high speed and accuracy to meet the requirements of quantitative packaging or batching manufacturing of a large number of materials.

At present, there are two common methods of automatic quantitative feeding device for powder and granular materials at home and abroad, volumetric and weighing. Volumetric dosing is based on the volume of the material for metered filling or feeding, and the quantitative feeding is rapid, but the quality of the quantitative material is changed by the density of the material. For example, the Chinese patent with the application number of 200920248298.2 considers that it is difficult to control the quantification during fast cutting, and adopts the method of first fast and then slow to reduce the influence of the feeding drop, but the final value of the cutting can only be close to the expected value, and the accuracy is not high.

Weighing type quantitative filling or feeding is carried out according to the quality of the material, which can be divided into incremental type and weightless type according to the different weighing measurement methods. Incremental weighing of the materials falling into the weighing hopper. This method requires continuous weighing during the feeding process, and feedback control of the feeding amount according to the weighing result. Since the material is continuously falling, when the feeding valve is closed , there is still some material in the air. In order to compensate for the interference of the aerial material on the measurement accuracy, many schemes use the technology of closing the valve in advance. For example, the Chinese patent application number 201410230888.8 divides the batching weighing process into three stages, and uses an iterative learning control method to calculate the final stage. Turn off the advance control amount.

Compared with the incremental type, the loss-in-weight weighing method measures the weight of the falling materials by continuously weighing the weight of the silo, thereby avoiding the problem of aerial materials. For example, the Chinese patents with application numbers of 200710142591.6, 201010108011.3 and 201310178558.4 all measure the falling materials through the calculation of the weight reduction of the weighing bin. Although these schemes do not need to consider the air volume, they do not consider the weight loss when the material falls from the feeding valve. It affects the accuracy of weighing and measurement, cannot meet the requirements of high-precision blanking, and these solutions can only be continuously blanked and cannot be directly applied to batch blanking.

Compared with the previous weightless weighing and blanking, if a nonlinear mapping can be constructed through the analysis of various factors that affect the weightless equivalent value of the falling material, the actual weight of the material in the weightless weighing process can be determined based on this mapping. The amount of feeding is measured.

SUMMARY OF THE INVENTION

The traditional weight loss scale realizes measurement by the principle of controlling the weight loss during operation. The discharge device and the weighing hopper are weighed, and the weight loss is calculated according to the reduction of the material weight per unit time in the weighing hopper of the weight loss scale. ΔG/Δt The feed flow of the scale. In the previous weightless weighing method, although the flow rate is obtained by the differential method, the change of the blanking flow rate between the two differentials, the influence of material bonding, and the environment such as vibration will affect the accuracy of the differential result. .

From the analysis of the blanking method, the general loss-in-weight weighing measurement generally uses a screw conveyor as the discharging device, which can only dynamically adjust the blanking rate during continuous operation, but cannot directly perform batch intermittent blanking; loss-in-weight weighing The two parameters of weighing accuracy and batching speed are two contradictory control quantities. To improve the weighing accuracy, it is hoped that the more stable the scale body, the better, that is, the slower the feeding speed, the better, but the batching time is bound to increase. Low efficiency; on the contrary, if the feeding speed is too fast, the accuracy is difficult to guarantee.

Considering the advantage that the loss-in-weight weighing does not need to consider the air volume, the solution of the present invention integrates it into the control of batch blanking. However, due to the non-zero speed of the material falling during the feeding process, it will affect the weighing measurement, which makes the weighing reading different from the static weighing. The dynamic impact caused by the non-zero speed of the falling material, that is, the weight loss value of the falling material, has many influencing factors, such as the closing speed of the conveying device, the falling shape of the material, the flow rate, etc., so the static weight is obtained by dynamic weighing. It is difficult to determine at one time through offline experiments.

According to the in-depth testing and analysis of the loss-in-weight weighing and unloading process, it is concluded that the most important factors affecting the weight loss value of the falling material of the straight-fall loss-in-weight material unloading machine include: the material level of the unloading bin, the blanking rate, the material density and the falling material. Material valve opening diameter. The weight loss value of the falling material is a complex nonlinear mapping of these physical quantities. In order to predict the weight loss value of the falling material and then adjust the valve closing time for accurate feeding, it is necessary to identify and express the mapping relationship.

The method of identifying the system and correcting the parameters based on the linear system theory can be well applied to the linear system, but it cannot be applied to the complex nonlinear system. Artificial neural network is a network that is widely interconnected by a large number of processing units. It has strong self-adaptive, self-organizing and self-learning capabilities. It is widely valued in system modeling, identification and control. Characteristics provide an effective method for system identification, especially nonlinear system identification.

At present, the multi-layer forward network is the most widely used in the identification of nonlinear systems. The multi-layer forward network has the ability to approximate any continuous nonlinear function, but the network structure is generally static. From the analysis of the material falling process, it can be seen that Since the material level of the unloading bin changes gradually, there is also a close relationship between the weight loss values of falling materials in two consecutive sampling periods. To this end, a dynamic recurrent neural network is used in the controller of the present invention to model the system. Different from the static feedforward neural network, the dynamic recurrent network has the function of mapping dynamic features by storing the internal state, so that the system has the ability to adapt to the time-varying characteristics, and is more suitable for the identification of nonlinear dynamic systems. In the scheme of the present invention, based on the dynamic recursive Elman neural network, the mapping relationship between the weight loss value of the falling material and the material level h of the unloading bin, the blanking rate d, the material density ρ and the opening aperture D of the unloading valve is identified, and the unloading During the process, the material distribution in the unloading silo is detected and dynamically adjusted, so that the trained neural network can accurately predict the weight loss value of the falling material in different states, so as to achieve high-precision cutting.

The technical solution of the present invention is to provide a straight loss-in-weight material unloading machine based on a neural network with the following structure, including: a frame, a cutting bin, a cutting valve, a mixing hopper, a weighing module, and a blanking valve , mixing silo and controller;

The feeding valve is located at the bottom opening of the feeding bin, and the feeding bin and feeding valve are 2 to 6 groups.

The unloading bin is installed on the weighing module fixed on the rack, and there is a bin position sensor inside it.

The bottom opening of the mixing hopper located under the blanking valve is controlled by the blanking valve, and a mixer is installed on its inner wall;

The mixing bin is located below the blanking valve, and has a push plate at the bottom;

The controller contains a neural network module using dynamic recursive E1man neural network, and each blanking valve has a corresponding neural network module. The four input quantities of the opening diameter of the feeding valve and the opening diameter of the feeding valve are mapped to the weight loss value of the falling material; the controller predicts the weight loss value of the falling material through the neural network module and adjusts the closing time of the feeding valve after correcting the feeding volume based on the predicted value;

The controller controls the action of each blanking valve in turn. After finishing one batch blanking, the blanking valve is opened, and after detecting that the materials in the mixing bin have accumulated to the set value, the push plate is opened to mix the materials evenly. discharge.

Preferably, it also includes a storage bin and a feed pump, the outlet of the feed pipe at the rear end of the feed pump is provided with a material spray head, and the material spray head is spherical cap-shaped with small circular holes distributed on its surface.

Preferably, the bin position sensor is installed on a top corner of the unloading bin near the center of the frame, and has a rotating base at the bottom.

Preferably, the position sensor adopts a distance sensor.

Preferably, a stirrer is also installed on the side wall of the unloading bin, and the stirrer includes a base, two support arms, a support arm rotating shaft connecting the two support arms, a claw hand rotating shaft and a claw hand.

Preferably, a mixing material level sensor is installed on the side wall of the mixing silo, and there is a homogenizer inside. Tube.

Preferably, the mixer includes a mixing base, two mixing support arms, and a mixing support arm rotating shaft, a mixing claw rotating shaft and a mixing claw connecting the two mixing support arms in sequence.

Preferably, the mixer includes a mixing rotating shaft, a mixing rotating disc and a screw blade mounted on the mixing rotating shaft, and a mixing support frame supporting the mixing rotating shaft.

Preferably, the model of the neural network is:

xc _k (t)=x _k (t-mod(k,q)-1),

Among them, mod is the remainder function, f() function is taken as the sigmoid function; xc _k (t) is the output of the successor layer, x _j (t) is the output of the hidden layer, u _i (t-1) and y(t) are the input layer input and the output layer output, respectively, ω _j , ω _jk and ω _ji are the connection weights from the hidden layer to the output layer, the connection weights from the successor layer to the hidden layer, and the connection from the input layer to the hidden layer, respectively Weights, θ and θ _j are the thresholds of the output layer and the hidden layer respectively; k=1, 2...m, q is the selected regression delay scale, which is preferred according to the sampling period and cutting rate, such as optional q=3; j==1, 2...m, i=1, 2...4, the number m of nodes in the hidden layer and the successor layer can be selected from 11 to 20, for example, 16 is preferred.

Preferably, in addition to the predicted value of the weight loss value of the falling material, the current accumulated blanking error should be taken into account when correcting the blanking amount.

Another technical solution of the present invention is to provide a loss-in-weight material unloading machine controller based on neural network, which includes a signal acquisition module, a processing module, a neural network module, an iterative learning module, a storage module, and a first connection array. , the second connection array and the output module, the signal acquisition module collects the sensing signals of the material level of the unloading bin and the weight of the unloading bin in real time through the position sensor in the unloading bin and the weighing module carrying the unloading bin and transmits them to The processing module is used for data processing and analysis, and the memory is used for data storage;

The neural network module adopts a dynamic recursive Elman neural network, and its input layer receives four inputs from the processing module: the material level of the unloading silo, the blanking rate, the material density and the opening aperture of the unloading valve, and the output of the output layer passes through the 4th input respectively. A connection matrix and a second connection matrix are transmitted to the iterative learning module and the processing module;

When training the neural network offline, the iterative learning module adjusts the connection weights of the neural network according to the actual weight loss value of the falling material and the network output value input by the processing module and the neural network through the first connection array respectively;

When the material is controlled online, the first connection array is disconnected, and the neural network predicts the weight loss value of the falling material and outputs it to the processing module through the second connection array. The blanking valve is controlled by closing the valve.

Compared with the prior art, the scheme of the present invention has the following advantages: the present invention adopts a nonlinear network to model the relationship between the weight loss value of the falling material and its influencing factors, and can predict the static weight according to the dynamic weighing reading, so that it can be Accurate feeding is achieved by adjusting the closing time of the feeding valve. Compared with the traditional loss-in-weight scale scheme, this scheme can be used to accurately discharge materials in batches and is suitable for small batch production; the storage position sensor and agitator are used to detect and adjust the material accumulation form in the unloading bin to reduce The fluctuation of the blanking rate is reduced; the total error of batch blanking is also reduced by controlling the cumulative error of blanking.

Description of drawings

Figure 1 is a structural diagram of a straight-fall loss-in-weight material unloading machine based on a neural network;

Figure 2 is the outline structure diagram of the straight-fall loss-in-weight material unloading machine based on neural network;

Figure 3 is a schematic diagram of the weight loss effect of material falling;

Figure 4 is a schematic diagram of the composition of the controller;

Fig. 5 is a schematic diagram of Elman neural network structure;

Figure 6 is a schematic diagram of the partial structure of the bottom of the unloading silo;

Fig. 7 is a schematic diagram of the structure of the agitator in the unloading bin;

Figure 8 is a schematic diagram of the partial structure of the storage bin and the unloading bin;

Fig. 9 is the schematic diagram of material flow laminar flow in the lower silo;

Figure 10 is a schematic diagram of the distribution of multi-component materials in the mixing hopper;

Figure 11 is a schematic diagram of the structure of the mixing hopper in Example 1;

12 is a schematic diagram of the structure of the mixing hopper in Example 2.

Among them: 1, unloading bin 2, unloading valve 3, mixing hopper 4, weighing module 5, blanking valve 6, mixing bin 7, push plate 8, feeding pipe 9, controller 10, storage bin 11 , feed pump 12, position sensor 13, mixer 14, vibrator 15, feed pipe 16, material nozzle 17, small hole 18, agitator 19, material level surface 20, docking point 21, scan line 22, Homogenizer 23, mixed material level sensor

30. Rack

91. Signal acquisition module 92, processing module 93, neural network module 94, iterative learning module 95, storage module 96, first connection array 97, second connection array 98, output module

101. Buffer pool 102, umbrella body 103, damper 104, umbrella cap 105, umbrella stand

131, mixing base 132, mixing support arm 133, mixing support arm shaft 134, mixing claw hand shaft 135, mixing claw hand 136, mixing support frame 137, mixing shaft 138, mixing turntable 139, screw blade

181, base 182, arm 183, arm shaft 184, claw hand shaft 185, claw hand

301. Arc wedge

Detailed ways

The preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings, but the present invention is not limited to these embodiments. The present invention covers any alternatives, modifications, equivalent methods and arrangements made within the spirit and scope of the present invention.

In order to give the public a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, and those skilled in the art can fully understand the present invention without the description of these details.

The invention is described in more detail by way of example in the following paragraphs with reference to the accompanying drawings. It should be noted that the accompanying drawings are all in a relatively simplified form and in an inaccurate scale, and are only used to facilitate and clearly assist in explaining the purpose of the embodiments of the present invention.

Example 1:

As shown in Fig. 1 and Fig. 2, the neural network-based straight-fall loss-in-weight material unloading machine of the present invention includes unloading bin 1, unloading valve 2, mixing hopper 3, weighing module 4, blanking valve 5, Mixing silo 6 and controller 9, in which each component of material has a set of unloading silo 1 and unloading valve 2 corresponding, the commonly used component categories are 2 to 6, and component categories can be added as needed. Preferably, the feeding bin 1 adopts a bin-shaped structure composed of right-angled trapezoids and rectangles.

The rack 30 serves as a frame of the equipment for fixing and supporting various other components. The weighing module 4 is fixed on the frame 30 , the unloading bin 1 is installed on the weighing module 4 , the bottom of the mixing hopper 3 has an opening, and the opening and closing of the opening is controlled by the blanking valve 5 . The mixing hopper 4 is located at the lower part of the lowering bin 1 , and the centers of the plurality of feeding valves 2 are distributed in an arc shape relative to the center of the mixing hopper 4 . A mixer 13 for uniformly mixing multi-component materials is installed on the inner wall of the mixing hopper 4 .

The controller 9 adopts the touch operation mode, and the human-machine interface on the touch screen can set the formula and other parameters of the multi-component material. As shown in FIG. 1 and FIG. 4 , the controller 9 is connected to each sensor and action components through a signal acquisition module 91 and an output module 98 respectively.

The mixing bin 6 is located below the blanking valve 5, and has a push plate 7 at the bottom, and a feeding pipe 8 is connected below the push plate, which transports the multi-component mixed material to the packaging bag or production equipment.

Preferably, a mixing material level sensor 23 is installed on the side wall of the mixing material bin 6, and there is also a homogenizer 22 inside, and the homogenizer 22 adopts a helical blade. The capacity of the mixing bin 6 is several times, such as 15 times, that of the mixing hopper 3. After completing multiple batches of material, the controller 9 reads the state of the mixed material level sensor 23. If it is detected that the material level of the mixed material exceeds the set value If the threshold value is reached, the homogenizer 22 is controlled to rotate to stir the mixture evenly again. Under the control of the controller 9 , the push plate 7 is opened, and the mixture is output from the feeding pipe 8 .

Figure 3 illustrates the influence of the weight loss effect of the material on the weight during the material falling process. The material falls from the unloading valve 2 at a speed v _0. The unloading valve 2 and the unloading bin 1 are carried on the weighing module 4 together. The weighing module 4. The mass equivalent change of the material in the lower silo 1 can be measured by the following formula:

Among them, Gs is the initial weight at zero time, at time t, dm is the blanking mass per unit time (g/s) at the outlet of the unloading valve, v0 is the initial speed of the material falling, and the material with Δm leaves the bottom within Δt time. material valve 2.

As shown in Figure 2 and Figure 3, the controller can dynamically read the current reading of the weighing module in real time, but during the feeding process, the reduced value read by the controller is not the actual weight of the material falling into the mixing hopper, but It is the reverse impact effect including the falling of the material. Therefore, the impact of this impact should be deducted when calculating the material blanking amount. But in practice, how to accurately obtain the equivalent weight value of the impact is a difficult problem that must be solved.

It can be seen from formula (1) that the material quality detected by the weighing module not only includes the actual falling material, which is the second item in the formula, but also receives the momentum reverse impact caused by the material falling at a non-zero speed, which is the first item in the formula. Three effects, the third of which is the weightless effect. Therefore, the method of directly reading the weight in the traditional loss-in-weight scale cannot obtain the actual weight at a certain moment in the unloading bin.

In order to obtain the exact mass of the material in the silo at the current moment, it is necessary to consider the influence of this weightless effect. First, the third value in formula (1), which is the equivalent weight loss value of the falling material, must be obtained.

Through repeated experimental testing and analysis of the loss-in-weight blanking process, it is concluded that the most important factors affecting the weight-loss value of falling materials during the blanking process of the straight-fall loss-in-weight material blanking machine include: the material level h of the blanking bin, the falling material The material rate d, the material density ρ and the opening diameter D of the feeding valve. The weight loss value of falling material is a complex nonlinear mapping of these physical quantities. In order to accurately predict the weight loss value of falling materials in different states, so as to adjust the valve closing time to carry out accurate feeding, it is necessary to identify the mapping relationship.

Based on the complex nonlinear characteristics of the mapping, and considering the close relationship between the weight loss values of falling materials in two consecutive sampling periods, the present invention adopts dynamic recursive Elman neural network modeling to analyze the relationship between the weight loss values of falling materials and unloading bins. The mapping relationship between the position h, the blanking rate d, the material density ρ and the opening aperture D of the blanking valve is identified.

As shown in FIG. 4, the controller in the present invention includes a signal acquisition module 91, a processing module 92, a neural network module 93, an iterative learning module 94, a storage module 95, a first connection array 96, a second connection array 97 and an output module 98. Wherein, the neural network module 93 adopts Elman neural network, and the storage module 95, ie, the memory, is used to store data.

As shown in Figure 5, the Elman neural network used has a recursive structure. Compared with the BP neural network, the Elman neural network includes a succession layer in addition to the input layer, the hidden layer and the output layer. The succession layer is used between layers. The feedback connection of the network enables it to express the time delay and parameter timing characteristics between the input and output, so that the network has a memory function. In Figure 5, the input layer of the established neural network has 4 units, and the number of nodes m in the hidden layer and the successor layer can be selected from 11 to 20. If 16 is selected, the output layer has only one unit.

The model of the neural network is:

xc _k (t)=x _k (t-mod(k,q)-1) (2),

Among them, mod is the remainder function, f() function is taken as the sigmoid function; xc _k (t) is the output of the successor layer, x _j (t) is the output of the hidden layer, u _i (t-1) and y(t) are the input layer input and the output layer output, respectively, ω _j , ω _jk and ω _ji are the connection weights from the hidden layer to the output layer, the connection weights from the successor layer to the hidden layer, and the connection from the input layer to the hidden layer, respectively Weights, θ and θ _j are the thresholds of the output layer and the hidden layer respectively; k=1, 2...m, q is the selected regression delay scale, which is preferred according to the sampling period and cutting rate, such as optional q=3; j==1, 2...m, i=1, 2...4.

As shown in FIG. 2 and FIG. 4 , preferably, the controller 9 in the present invention can also realize the switching of the first connection array and the second connection array through the touch screen operation, so that the controller works in offline training or online prediction. model.

Combined with Figure 4 and Figure 5, the established neural network, the input layer includes 4 nodes, of which the material density ρ and the opening aperture D of the feeding valve are determined values, and the other two quantities need to be dynamically real-time through the signal acquisition module. collection. By introducing multiple nodes in the succession layer with different delay regressions into the network, the network structure is more in line with the blanking process, so that the network training can converge faster.

When training the neural network offline, the iterative learning module adjusts the connection weight of the neural network according to the actual weight loss value of the falling material and the network output value input by the processing module and the neural network through the first connection matrix respectively.

In order to obtain training samples, after the start of feeding, when the material forms a continuous material flow from the feeding valve at the bottom of the feeding bin to the mixing hopper, continue feeding for a period of time, and read the dynamic weight of the weighing module in real time when the valve is closed Read W, wait for the material to fall and read the static weight reading WD of the weighing module, then the actual value of the falling material weight loss when the valve is closed is L=WD-W, which is the actual value of the sample output value y That is, the expected value y _d .

The neural network training adopts the gradient descent method, and the weight and threshold adjustment methods in the training are as follows.

Assuming there are P training samples in total, let the error function be:

The adjustment formula of the connection weights from the hidden layer to the output layer is as follows:

ω _j (t+1)=ω _j (t)+Δω _j (t+1) (6),

in,

δ _y =-(y _d -y)·y·(1-y) (8),

The adjustment formula of the output layer threshold is:

θ(t+1)=θ(t)+Δθ(t+1) (9),

in,

Similarly, the adjustment formula of the connection weights from the input layer to the hidden layer is:

ω _ji (t+1)=ω _ji (t)+Δω _ji (t+1) (11),

in,

δ _j = δ _y ·ω _j ·x _j (t) · (1-x _j (t)) (13),

The adjustment formula of the hidden layer threshold is:

θ _j (t+1)=θ _j (t)+Δθ _j (t+1) (14),

in,

Without considering the dependence of xc _k (t) on the connection weight ω _jk , the adjustment formula of the connection weight from the successor layer to the hidden layer is:

ω _jk (t+1)=ω _jk (t)+Δω _jk (t+1) (16),

in,

The initial value range of each weight is taken as the (-0.1, 0.1) interval, and the learning rate η is a decimal less than 1, which can be adjusted dynamically by using a fixed rate or according to the total error of the current network output.

The training end condition can be set as the total error or its variation is less than a set value or the number of training times reaches a certain amount.

Preferably, in order to make the training samples cover more situations, each time the valve is closed can be set as a random value after the time when the weight reading of the weighing module becomes a certain value.

Before network training, normalize preprocessing on 4 inputs and 1 output:

r′=rr _min /r _max -r _min (18),

Among them, r is the unprocessed physical quantity, r' is the normalized physical quantity, and r _max and r _min are the maximum and minimum values of the sample data set, respectively.

When calculating the weight loss prediction value of falling materials, use the following formula to convert the network output back to the weight loss value of falling materials:

r=r _min +r'·(r _max -r _min ) (19).

When the material is controlled online, the first connection array is disconnected, and the neural network predicts the weight loss value yL of the falling material and outputs it to the processing module through the second connection array. The blanking valve is closed valve control:

Assuming that the one-time unloading amount of the current component is Ws, when starting to unload, the controller obtains the initial static weight of the unloading bin as G0 by reading the sensing value of the weighing module; then, the controller continuously reads the weighing The sensing value of the module, when the dynamic weight reading reaches (G0-Ws-yL), close the unloading valve.

As an option, in addition to the predicted value of the weight loss value of the falling material, the current accumulated blanking error of this component should be compensated, that is, when it is detected that the dynamic weight reading of the blanking bin reaches (G0-Ws-yL+E), the blanking will be closed. valve, where E is the current accumulated blanking error of the component, and E is a timing that indicates too much blanking.

The signal acquisition module separately collects the sensing signals of the material level and the weight of the falling material in real time through the position sensor in the unloading silo and the weighing module carrying the unloading silo, respectively, and transmits them to the processing module for data preprocessing. The neural network, the output value of the neural network and the expected output value preprocessed by the processing module are all transmitted to the iterative learning module through the first connection matrix, and the iterative learning module returns the adjusted weights to the neural network according to the gradient descent method.

The storage position sensor adopts a distance sensor to detect the material level height of the material in the unloading silo. By periodically collecting the signals of the weighing module, the processing module can calculate the mass equivalent of the falling material per unit time, that is, the blanking rate.

2 and 6 , preferably, in order to reduce the influence of the action of the feeding valve 2 on the weight, a buffer pool 101 is provided at the bottom of the feeding silo, which includes a damper 103 and an umbrella body 102 . The damper 103 adopts a soft connection section, which can reduce the vibration transmitted to the weighing module when the feeding valve 2 acts. The umbrella body 102 further includes an umbrella cap 104 and an umbrella stand 105 supporting the umbrella cap.

It can also be analyzed from formula (1) that the weight loss value of falling materials is closely related to the blanking condition, which is affected by the shape distribution of materials in the lowering bin 1.

Under the action of gravity, the outflow form of particulate matter from the lower silo mainly includes two types: bulk flow and central flow. In the flow pattern of bulk flow, the entire particle layer in the silo can flow out approximately uniformly, and basically every particle is moving; while in the flow pattern of central flow, some particles are static, and there is a gap between the flowing and static particles. Flow channel boundaries. The overall feeding rate of the bulk flow is larger than that of the central flow, and the fluctuation of the feeding rate is small and the flow is stable. In the actual production process, the material in the lower silo may have a flow pattern of central flow, so that when the material port begins to discharge, the material is solidified into a plate due to the compaction stress generated by the silo pressure.

As shown in Figures 6 to 8, as an option, in order to reduce the fluctuation range of the blanking rate in the blanking bin, so as to better predict the weight loss value of falling materials, the blanking machine adopts a distance sensor type storage position sensor and a manipulator agitator. Detect and adjust the material accumulation pattern in the unloading silo, so that the dynamic material arch formation and collapse alternately appear above the unloading port, so as to ensure that the blanking shape is a stable overall flow pattern.

As shown in Figure 8, the unloading bin 1 is continuously discharging, and when the material level in the bin drops to a certain value, it needs to be replenished. To this end, a storage bin 10 is set above the lowering bin 1 , and the materials in the storage bin 10 enter the lowering bin 1 through the feeding pump 11 and the feeding pipe 15 . In order to make the material particles evenly unloaded, a material nozzle 16 is provided at the end outlet of the feeding pipe 15. The surface of the material nozzle 16 is spherical, and its surface is distributed with circular small holes 17. The diameter of the small holes is based on the particle size of the material. Make optimization. The feed pump 11 adopts a screw conveyor, and its action is controlled by a controller. During the unloading process of the unloading bin 1, with the lowering of the material level surface 19, the feed pump 11 operates under the control of the controller, so that the material level of the top surface of the material in the unloading bin 1 is kept near the preset value.

The two figures in FIG. 8 are viewed from the side view and top view of the lowering bin 1, respectively. As shown in FIGS. 8a and 8b, a bin position sensor 12 is installed on a top corner of the lowering bin 1 near the center of the frame. The rotating base can be pitched and rotated, so that the position sensor can detect materials in the directions of different docking points 20. Each docking point 20 forms scan lines 21 close to concentric circles, so as to determine the distribution of the material level surface 19.

As shown in FIG. 7 , the controller improves the distribution of the material through an agitator 18 installed on the side wall of the unloading bin 1 by the unloader. The agitator 18 includes a base 181, two supporting arms 182, a supporting arm rotating shaft 183 connecting the two supporting arms, a claw hand rotating shaft 184 and a claw hand 185, wherein the base 181 also contains a rotating shaft.

As shown in Figure 7 and Figure 6, the umbrella body in the buffer pool at the bottom of the lower silo can bear the pressure of the upper silo, weaken the effect of the large compaction force near the discharge port, and greatly reduce the silo pressure under the umbrella cap. At the same time, an annular material flow port is formed around the silo, so that the material in the silo tends to flow as a whole, which can further prevent the material from being blocked by arching.

During the unloading process, the controller judges the distribution of the materials in the unloading bin through the detection of the bin position sensor and the tracking of the unloading rate per unit time, so that the material level surface in the unloading bin maintains an approximate parabolic shape. As shown in Figures 8-9, when the materials are evenly distributed, the distance values of the materials detected by the bin sensor in different directions are approximately concentrated in a small range after the geometric transformation of the ray and the inclination of the vertical direction. When the material is locally hardened or stable, the detected distance value exceeds this range. At the same time, the feeding rate of each feeding bin is tracked in real time through the weighing module. When the distance sensor detects the above abnormal state or finds that the fluctuation of the feeding amount per unit time exceeds the set threshold value such as 5%, the controller commands the agitator to act. In the low point area of the material level, do a helical overturn, so as to break the occasional hardening or material arch, restore the flow of the material, and maintain the laminar flow state of the overall flow.

As shown in Fig. 9, the present invention greatly reduces the effect of the compaction force generated by the impact of the charging through the detection and action coordination of the distance sensor and the agitator, effectively preventing the particle size segregation of the material in the bin, and making the lower bin hopper The material inside is activated, which improves the flow of the material. During the continuous feeding and unloading process, all the particles are flowing in an orderly manner. With the outflow of the particles in the bin, the particle group presents the laminar flow state of the overall flow.

Figure 10 supplements the record of the multi-component material unloading process, which shows the schematic diagram of the material distribution in the mixing hopper when the 4 components are unloaded.

As shown in Figure 2 and Figure 11, when the multi-component material falls into the mixing hopper 3, the mixer in the mixing hopper 3 moves to mix the materials evenly. As shown in FIG. 11 , the mixer 13 includes a mixing base 131 , two mixing support arms 132 , a mixing support arm rotating shaft 133 and a mixing claw hand rotating shaft 134 connecting the two mixing support arms 132 in sequence. and mixing claw hand 135, wherein the mixing base 131 also contains a rotating shaft.

There is an arc-shaped wedge 301 on the bottom inner side of the openable and closable component at the bottom opening of the mixing hopper 3, and the mixing claw 135 is made of semi-hard flexible material. Under the action of the controller, the mixing claw 135 of the mixer 13 repeats the spiral turning from the left side of the mixing hopper to the bottom side to the right side, from high to low to high, to mix and stir multi-component materials. On the outside of the non-open side of the mixing hopper, there is also a vibrator 14. The controller also controls the vibrator 14 to start vibrating while the mixer 13 is operating. Under the action of the device 14, it can be fully and uniformly mixed.

The blanking machine of the present invention is used for blanking, and the blanking behavior of each blanking valve is firstly modeled off-line, and the blanking of each component of materials is carried out separately in the process of collecting samples, so that all materials can be recovered without causing damage. waste. During actual cutting, the material level and the blanking rate of the blanking bin are periodically collected, and the current weight loss value of the falling material can be predicted in real time. Therefore, starting from the first batch, the material can be cut accurately and other methods such as online iterative learning can be avoided. The blanking error in the scheme fluctuates.

Example 2:

As shown in Figure 2 and Figure 12, when the multi-component material falls into the mixing hopper 3, the mixer in the mixing hopper 3 moves to mix the materials evenly. As shown in FIG. 12 , the mixer 13 includes a mixing shaft 137 fixed in the mixing hopper, a mixing turntable 138 and a helical blade 139 installed on the mixing shaft 137, and a mixing support fixed on the inner wall of the mixing hopper 3 The frame 136 is used to support the mixing shaft 137 . The mixing turntable 138 is similar to the annular shape of a water wheel, and its outer ring is provided with rectangular blades that are substantially perpendicular to the circumference, and the blades can be perforated. The helical blade 139 adopts an irregular helical blade, and holes are distributed on the blade.

There is an arc-shaped wedge 301 on the bottom inner side of the openable component at the bottom opening of the mixing hopper 3. Under the action of the controller, the mixing shaft 137 of the mixer 13 rotates, and the inner rectangular blade and the spiral blade 139 together Turn the material over.

On the outside of the non-open side of the mixing hopper, there is also a vibrator 14. The controller also controls the vibrator 14 to start vibrating while the mixer 13 is operating. Under the action of the device 14, it is fully and uniformly mixed.

Example 3:

As shown in Figure 2, in order to measure the weight loss of the lower silo 1, two horizontal support parts can be drawn from the outer side wall of the lower silo; the weighing module is placed horizontally, and the weighing module supports the blanking from both sides in the vertical direction warehouse. Or two suspension parts are drawn from the top of the lower silo 1, the weighing module is placed horizontally, and the weighing module supports the lower silo from both sides in the vertical direction.

The above-mentioned embodiments do not constitute a limitation on the protection scope of the technical solution. Any modifications, equivalent replacements and improvements made within the spirit and principles of the above-mentioned embodiments shall be included within the protection scope of this technical solution.

Claims (6)

1. The signal acquisition module acquires sensing signals of the material level and the weight of the blanking bin in real time through a bin position sensor in the blanking bin and a weighing module bearing the blanking bin respectively and transmits the sensing signals to the processing module for data processing and analysis, and the memory is used for data storage;

the neural network module adopts a dynamic recursive Elman neural network, and an input layer of the neural network module receives 4 input quantities of material level, blanking rate, material density and opening aperture of a blanking valve from the processing module respectively;

when the blanking is controlled on line, the neural network predicts the weight loss value of falling materials and outputs the weight loss value to the processing module, and the processing module processes and analyzes the weight loss value and then controls the blanking valve at the opening at the bottom of the blanking bin to be closed through the output module:

continuously reading the sensing value of the weighing module, and closing the discharging valve when the dynamic weight reading reaches (G0-Ws-yL);

wherein Ws is the one-time feeding amount of the current components, G0 is the initial static weight of a feeding bin when feeding is started, and yL is the lost weight value of the falling materials predicted by a neural network;

and when the fluctuation of the blanking amount per unit time is found to exceed a set threshold value, the controller commands the stirrer on the side wall of the blanking bin to act.

2. The neural network-based direct falling weight loss type material blanking machine controller is characterized by further comprising an iterative learning module, a first connection array and a second connection array;

the output quantity of the output layer is transmitted to the iterative learning module and the processing module through the first connection array and the second connection array respectively;

when the neural network is trained off line, the iterative learning module adjusts the connection weight of the neural network according to the actual weight loss value of the falling material and the network output value which are input by the processing module and the neural network through the first connection array respectively;

when the blanking is controlled on line, the first connection array is disconnected, and the trained neural network predicts the weight loss value of the falling material and outputs the weight loss value to the processing module through the second connection array.

3. The neural network-based falling weight loss type material blanking machine controller is characterized in that in the valve closing control, the current accumulated blanking error of the material of the component is compensated, when the dynamic weight reading of the blanking bin is detected to reach (G0-Ws-yL + E), the blanking valve is closed, wherein E is the current accumulated blanking error of the component, and the positive E value is used for indicating that the blanking is excessive.

4. The neural network-based direct-falling weight loss type material blanking machine controller according to claim 1, wherein the number of the neural network modules is multiple, and one neural network module corresponds to each blanking valve.

5. The neural network-based direct-falling weight-loss material blanking machine controller is characterized in that when material distance values detected by the bin sensor in different directions are geometrically transformed to exceed a set range, the controller commands the stirrer to act.

6. The neural network-based fall weight loss material blanking machine controller according to claim 1, wherein the controller further controls the action of the feed pump so that the level of the top surface of the material in the blanking bin is kept near a preset value.

Images