CN110697449A

CN110697449A - Screw weightless formula material blanking machine controller based on neural network

Info

Publication number: CN110697449A
Application number: CN201910845752.0A
Authority: CN
Inventors: 邹细勇; 王月真; 穆成银
Original assignee: China University of Metrology
Current assignee: China Jiliang University; China University of Metrology
Priority date: 2017-09-19
Filing date: 2017-09-19
Publication date: 2020-01-17
Anticipated expiration: 2037-09-19
Also published as: CN107720311A; CN107720311B; CN110697449B

Abstract

The invention discloses a screw weightless material blanking 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 a material weight loss value based on the material level, the blanking rate and the material density of the blanking bin, the diameter of a helical blade of the screw conveyor, the screw pitch and the maximum rotating speed of a screw, and therefore the closing time of the screw conveyor is adjusted. The model is built for blanking weighing based on the neural network, and the trained network can accurately predict the weight loss value of falling materials in different blanking states, so that direct and accurate blanking control can be realized, and the method is suitable for small-batch production; the material accumulation form in the blanking bin is adjusted by combining the detection of the bin level sensor and the control of the vibrating rod, so that the fluctuation of the blanking rate is reduced; the total error of batch blanking is reduced by controlling the accumulated error of blanking; and the feeding efficiency is improved because the screw can keep high running speed.

Description

Screw weightless formula material blanking machine controller based on neural network

The application is divisional application with application number 201710863072.2, application date 2017, 09 and 19 and invention name 'screw weight loss type material blanking machine based on neural network and controller thereof'.

Technical Field

The invention relates to the field of quantitative blanking, in particular to a screw weightless material blanking machine controller based on a neural network.

Background

In industrial and agricultural manufacturing and commodity packaging, a large amount of powder materials, such as iron-making raw materials including iron ore concentrate, coal powder and the like, chemical raw materials including polypropylene, polystyrene, polyvinyl chloride, light methyl cellulose, polyacrylonitrile, epoxy resin powder coating and the like, building material raw materials including quartz sand, cement and the like, daily chemical products including washing powder and the like, grain and bean agricultural products including millet, soybean and the like, or agricultural production materials including powder, slag and granular processed food, feed, chemical fertilizer, pesticide and the like, and granular health care products, Chinese and western medicaments, seasonings and the like need to be automatically quantitatively packaged or prepared by batching.

At present, many enterprises in China still adopt manual quantitative batching or packaging, so that on one hand, the labor intensity is high, the speed is low, and the economic benefit is poor; on the other hand, manual quantification of food, medicine and the like often cannot meet the sanitary requirements, toxic and harmful materials are used, and manual quantification is easy to cause harm to human bodies. Therefore, for the manufacturing enterprises, it is urgently needed to provide a cheap multi-component automatic quantitative blanking device or device with higher speed and accuracy, so as to meet the requirements of quantitative packaging of a large amount of materials or manufacturing of ingredients.

At present, two common methods, namely a positive displacement type and a weighing type, are adopted for automatic quantitative powder material feeding devices at home and abroad. The volumetric quantification is used for metering filling or feeding according to the volume of the material, the quantitative feeding is rapid, but the quality of the quantified material is changed by the change of the density of the material. In order to improve the blanking precision, various adjusting methods are provided, for example, in the chinese patent with application number 201320001933.3, the variable frequency speed regulation is adopted for the screw, the feeding speed is gradually reduced when the target value is approached, and the air drop value is reduced; in the Chinese patent with the application number of 201310234280.8, a large screw and a small screw are adopted in a three-speed variable-frequency feeding process of a soda packing machine for feeding materials in multiple stages; the Chinese patent with the application number of 200920248298.2 reduces the influence of the fall of the feeding by a method of first quick and then slow considering that the quantitative control is difficult to control when the fast feeding is carried out; the final blanking value of the non-weighing schemes can only be close to the expected value, and the accuracy is not high.

The weighing type quantification is used for metering filling or feeding according to the mass of the material, and the weighing type quantification can be divided into an incremental type and a weightless type from the difference of the weighing and metering method. The incremental type is weighed to the material that constantly falls into the weighing hopper, and this kind of mode needs constantly weigh at the unloading in-process, according to weighing result feedback control unloading volume, because the material is the continuous whereabouts, when unloading valve closed, still partial material is in the air. In order to compensate the interference of the materials in the air to the metering precision, a technology of closing a valve in advance is adopted in many schemes, for example, the Chinese patent with the application number of 201410230888.8 divides the ingredient weighing process into three stages, and an iterative learning control mode is adopted in the last stage to calculate the closing advance control quantity.

Compared with the incremental type, the weightless type weighing mode measures the weight of falling materials by constantly weighing the weight of the stock bin, thereby avoiding the problem of the materials in the air. Such as chinese patent nos. 200710142591.6, 201010108011.3 and 201310178558.4, which measure falling materials by calculating the weight reduction of the weighing bin, these schemes do not need to consider the empty space amount, but cannot satisfy the requirement of high-precision blanking because the weight loss effect when the materials fall from the blanking valve is not considered, and these schemes can only continuously blank and cannot be directly applied to batch blanking.

Compared with the traditional weightless type metering blanking, if a nonlinear mapping can be constructed by analyzing various factors influencing the weightless equivalent value of a falling material, the actual blanking amount of the material in the weightless type weighing process can be metered based on the mapping.

Disclosure of Invention

The traditional weightless scale realizes metering by a principle of controlling weight loss during working, a discharging device and a weighing hopper are weighed, and the feeding flow of the weightless scale is calculated according to the reduction delta G/delta t of the weight of materials in the weighing hopper of the weightless scale in unit time. In the conventional weightless weighing method, although the flow is obtained by a difference method, the accuracy of difference results is influenced by the influence of factors such as the change of blanking flow rate between two differences, material adhesion, environment such as vibration and the like.

The method is analyzed in a blanking mode, a screw conveyor is generally adopted as a discharging device in the ordinary weightlessness weighing and metering, the blanking rate in continuous operation can be only dynamically adjusted, and quantitative blanking in batches cannot be directly carried out; the two parameters of the weighing precision and the batching speed of the weightless weighing are two mutually contradictory control quantities, and for improving the weighing precision, the more stable the weighing body is, the better the feeding speed is, but the batching time must be increased, and the efficiency is low; on the contrary, if the feeding speed is too fast, the precision is difficult to guarantee.

The simple screw feeder belongs to the volumetric quantitative category, has simple structure and low cost, but the stability and the precision of the quantitative filling speed depend on the stability of the material according to specific gravity, and are greatly influenced by the loosening degree of the material, the uniformity degree of particles and the like. Since the ordinary positive displacement type is conversion type in nature and cannot grasp the exact quality of the blanking like a weighing type, although a scheme combining weighing is provided later, the precision can be ensured only by depending on the extremely low feeding speed at the final stage of the blanking because no empty space amount prediction exists.

Considering the advantage that the weight loss type weighing does not need to consider the empty space, the scheme of the invention integrates the weight loss type weighing into the control of batch blanking. But the falling of the material with non-zero speed in the blanking process can affect the weighing, so that the weighing reading is different from static weighing. The dynamic impact caused by the falling of the non-zero speed of the material, namely the weight loss value of the falling material, has many influence factors, such as the closing speed of the conveying device, the falling state of the material, the flow rate and the like, so that the conversion scheme of acquiring the static weight by dynamic weighing is difficult to determine at one time through off-line experiments.

According to the deep test and analysis of the weightless weighing and feeding process, the most main influence factors of the falling material loss value of the screw weightless material feeding machine are summarized as follows: the material level of the blanking bin, the blanking rate, the material density, the diameter of a helical blade of the helical conveyor, the pitch and the maximum rotating speed of the screw. The falling material loss value is a complex non-linear mapping of these physical quantities, and in order to predict the falling material loss value and thus to perform accurate blanking by adjusting the screw closing time, the mapping relationship needs to be identified and expressed.

The method for identifying and correcting parameters of the system based on the linear system theory can be well applied to the linear system, but cannot be applied to a complex nonlinear system. The artificial neural network is a network formed by widely interconnecting a large number of processing units, has strong self-adaption, self-organization and self-learning capabilities, is generally emphasized in system modeling, identification and control, and has a nonlinear transformation characteristic, so that an effective method is provided for system identification, particularly for identification of a nonlinear system.

At present, the most applied in nonlinear system identification is a multilayer forward network which has the capability of approximating any continuous nonlinear function, but the network structure is generally static, and as can be seen from the analysis of the material falling process, because the material level of a blanking bin is gradually changed, the weight loss value of falling materials in two continuous sampling periods is also closely related. To this end, a dynamic recurrent neural network is employed in the controller of the present invention to model the system. Unlike static feedforward neural networks, dynamic recursive networks have the function of mapping dynamic features by storing internal states, so that the system has the capability of adapting to time-varying characteristics and is more suitable for identification of nonlinear dynamic systems. In the scheme of the invention, based on a dynamic recursion Elman neural network, the weight loss value of falling materials, the material level h of a blanking bin, the blanking rate D, the material density rho, the diameter D of a helical blade of a helical conveyor, the pitch S and the maximum rotating speed v of a screw are measured_RThe mapping relation between the two is identified, and the distribution of the materials in the blanking bin is detected and dynamically adjusted in the blanking process, so that the trained neural network can accurately predict the weight loss values of the falling materials in different states, and high-precision blanking is realized.

The technical scheme of the invention is to provide a screw weightless material blanking machine based on a neural network with the following structure, which comprises: the device comprises a rack, a blanking bin, a spiral conveyor, a mixing hopper, a weighing module, a blanking valve, a mixing bin and a controller;

the blanking bin is arranged on a weighing module fixed on the frame, and a bin position sensor is arranged in the blanking bin;

the spiral conveyor is positioned at an opening at the bottom of the discharging bin, and the discharging bin and the spiral conveyor are 2-6 groups;

the bottom opening of the mixing hopper positioned below the screw conveyor is controlled by a blanking valve, and a mixer is arranged on the inner wall of the mixing hopper; the mixing bin is positioned below the blanking valve, and the bottom of the mixing bin is provided with a push plate;

the controller comprises a neural network module adopting a dynamic recursive Elman neural network, each screw conveyer corresponds to one neural network module, and each neural network module maps the material level, the blanking rate and the material density of a corresponding blanking bin, the diameter of a screw blade of the screw conveyer, the screw pitch and the maximum rotating speed of the screw 6 input quantities as a falling material loss value; the controller predicts the weight loss value of the falling material through the neural network module and adjusts the closing time of the spiral conveyor after correcting the blanking amount based on the predicted value;

the controller controls the spiral conveyors to act in sequence, after the formula amount blanking is completed for one time, the blanking valve is opened, then the push plate is opened after the materials in the mixing bin are detected to be accumulated to a set value, and the uniformly mixed materials are discharged.

Preferably, the device also comprises a storage bin and a feeding pump, wherein a material spray head is arranged at the outlet of a feeding pipe at the rear end of the feeding pump, the material spray head is in a spherical cap shape, and round small holes are distributed on the surface of the material spray head;

the rotating speed of the feeding pump is controlled according to the following formula:

wherein, V_{0 feeding}A set maximum feeding speed, l is the feed binFront stock level, L_MAnd L_mRespectively the preset highest and lowest feeding bin material positions.

Preferably, the bin level sensor is arranged on a vertex angle of the lower bin close to the center of the rack, and the bottom of the bin level sensor is provided with a rotating base;

the vibrating rod is installed near the side wall of the blanking bin on the rack and comprises a supporting column, a holder, a vibrator and a vibrating rod which are sequentially connected, a spring buffer is arranged at the bottom of the vibrator, and particle protrusions are distributed on the surface of the vibrating rod.

Preferably, a mixing material level sensor is installed on the side wall of the mixing bin, a material homogenizer is arranged in the mixing bin, the material homogenizer adopts a spiral blade, and a material conveying pipe is arranged below the push plate.

Preferably, the blender comprises a blending base, two blending support arms, a blending support arm rotating shaft, a blending claw rotating shaft and a blending claw which are connected in sequence.

Preferably, the mixer comprises a mixing rotating shaft, a mixing rotating disc and a helical blade which are arranged on the mixing rotating shaft, and a mixing support frame for supporting the mixing rotating shaft.

Preferably, the bottom of the blanking bin is provided with a drawing plate; the screw conveyor comprises a screw box, a conveying screw, a connector and a motor, wherein a motor shell is connected with the screw box through the connector, the conveying screw in the screw box is connected with a motor shaft through a shaft sleeve, a feeding hole is formed in the upper surface of the screw box relative to the opening at the bottom of the discharging bin, and the other end, opposite to the motor, of the screw box is further connected with a vertically-placed discharging pipe.

Preferably, the controller controls the operating speed of the screw conveyor in the following manner:

A. from a stopped state at a rate of mu a_maxStarting at acceleration, when the speed reaches lambda.v_RKeeping the speed unchanged;

B. when the closing time is up, the value is expressed in mu a_maxThe acceleration starts to decelerate until stopping;

wherein, a_maxRated maximum acceleration, v, of the screw conveyor_RThe maximum speed is mu is an acceleration coefficient between 0.5 and 0.9, and the lambda is a speed coefficient between 0.85 and 1.0;

the closing time means that the current baiting weight read from the weighing module is equal to:

wherein Ws and Wa are respectively the predicted values of the once feeding amount and the weight loss of the falling material of the current material, d is the feeding rate of the screw conveyer when the screw rotates at the maximum speed, and t_sFor the deceleration stop time length: t is t_s＝λ·v_R/μ·a_max。

Preferably, the model of the neural network is:

xc_k(t)＝x_k(t-mod(k，q)-1)，

wherein mod is a remainder function, and f () is a sigmoid function; xc_k(t) is the carry layer output, x_j(t) is the hidden layer output, u_i(t-1) and y (t) are input layer input and output layer output, ω_j、ω_jkAnd ω_jiRespectively, the connection weight from the hidden layer to the output layer, the connection weight from the accepting layer to the hidden layer and the connection weight from the input layer to the hidden layer, theta and theta_jOutput layer and hidden layer thresholds, respectively; m, q is a selected regression delay scale, and is preferably selected according to the sampling period and the blanking rate, and if q is optionally 3; j is 1, 2.. m, i is 1, 2.. 6, and the number m of hidden layer and receiving layer nodes can be selected from 11 to 20, such as 16.

Preferably, in addition to the predicted value of the weight loss value of the falling material, the current accumulated blanking error is also taken into account when the blanking amount is corrected.

The invention provides another technical solution, which is to provide a screw weightless material blanking machine controller based on a neural network, comprising a signal acquisition module, a processing module, a neural network module, an iterative learning module, a storage module, a first connection array, a second connection array and an output module, wherein the signal acquisition module is used for acquiring sensing signals of the material level and the weight of a blanking bin in real time through a bin position sensor in the blanking bin and a weighing module bearing the blanking bin and transmitting the sensing signals to the processing module for data processing and analysis, and a memory is used for data storage;

the neural network module adopts a dynamic recursive Elman neural network, an input layer of the neural network module respectively receives 6 input quantities of material level, blanking rate and material density of a blanking bin, the diameter and the pitch of a helical blade of a helical conveyor and the maximum rotating speed of a screw from the processing module, and the output quantity of an output layer is respectively transmitted to the iterative learning module and the processing module through a first connecting array and a second connecting array;

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 feeding is controlled on line, the first connection array is disconnected, the 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, and the processing module processes and analyzes the weight loss value and controls the closing of the screw conveyor at the opening at the bottom of the feeding bin through the output module.

Compared with the prior art, the scheme of the invention has the following advantages: the invention adopts a nonlinear network to model the relation between the weight loss value of the falling material and the influence factors thereof, and can predict the static weight according to the dynamic weighing reading, thereby realizing accurate blanking by adjusting the closing time of the screw conveyor. Compared with the traditional weightless scale scheme, the scheme can be used for batch quantitative blanking of materials and is suitable for small-batch production; the bin level sensor and the vibrating rod are adopted to detect and adjust the material accumulation form in the discharging bin, so that the fluctuation of the blanking rate is reduced; the total error of batch blanking is reduced by controlling the accumulated error of blanking; compared with a common screw type blanking device, the screw conveyor can keep higher running speed, so that the blanking efficiency is improved.

Drawings

FIG. 1 is a composition structure diagram of a screw weight loss type material blanking machine based on a neural network;

FIG. 2 is a diagram of an external structure of a screw weightless material blanking machine based on a neural network;

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

FIG. 4 is a schematic diagram of the controller;

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

FIG. 6 is a schematic view of a partial structure of a bottom of a blanking bin;

FIG. 7 is a schematic view of a vibrating rod;

FIG. 8 is a schematic view of a partial structure of a storage silo and a feed silo;

FIG. 9 is a schematic view of laminar flow of the material in the blanking bin;

FIG. 10 is a schematic view of the distribution of multicomponent materials in the mixing hopper;

FIG. 11 is a schematic view of the structure of a mixing hopper in example 1;

fig. 12 is a schematic view of the structure of the mixing hopper in example 2.

Wherein: 1. the device comprises a discharging bin 2, a screw conveyor 3, a mixing hopper 4, a weighing module 5, a discharging valve 6, a mixing bin 7, a push plate 8, a conveying pipe 9, a controller 10, a storage bin 11, a feeding pump 12, a bin level sensor 13, a mixer 14, a vibrating table 15, a feeding pipe 16, a material spray nozzle 17, a small hole 18, a vibrating rod 19, a material level surface 20, a stopping point 21, a scanning line 22, a material homogenizing device 23, a mixing material level sensor 24, a discharging pipe

30. Rack

91. The device comprises a signal acquisition module 92, a processing module 93, a neural network module 94, an iterative learning module 95, a storage module 96, a first connection array 97, a second connection array 98 and an output module

101. Buffer pool 102, umbrella-shaped body 103, damper 104, umbrella cap 105, umbrella stand 106 and drawing plate

131. Mixing base 132, mixing arm 133, mixing arm rotating shaft 134, mixing claw rotating shaft 135, mixing claw 136, mixing support 137, mixing rotating shaft 138, mixing turntable 139, helical blade

181. Pillar 182, cradle head 183, vibrator 184, vibration rod 185, particle protrusion 186, vibration rod track

201. Screw box 202, conveying screw 203, connector 204 and motor

301. Arc wedge block

Detailed Description

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 only these embodiments. The invention is intended to cover alternatives, modifications, equivalents and alternatives which may be included within the spirit and scope of the invention.

In the following description of the preferred embodiments of the present invention, specific details are set forth in order to provide a thorough understanding of the present invention, and it will be apparent to those skilled in the art that the present invention may be practiced without these specific details.

The invention is described in more detail in the following paragraphs by way of example with reference to the accompanying drawings. It should be noted that the drawings are in simplified form and are not to precise scale, which is only used for convenience and clarity to assist in describing the embodiments of the present invention.

Example 1:

as shown in fig. 1 and 2, the screw weight loss type material blanking machine based on the neural network comprises a blanking bin 1, a screw conveyor 2, a mixing hopper 3, a weighing module 4, a blanking valve 5, a mixing bin 6 and a controller 9, wherein each group of blanking bin 1 corresponds to the screw conveyor 2, the types of commonly used components are 2-6, and the types of components can be increased according to needs.

The bottom of the lower bunker 1 is provided with a drawing plate 106, the screw conveyor 2 comprises a screw box 201, a conveying screw 202, a connector 203 and a motor 204, the shell of the motor 204 is connected with the screw box 201 through the connector 203, and the conveying screw 202 in the screw box 201 is connected with the shaft of the motor 204 through a shaft sleeve; the upper surface of the screw box 201 is provided with a feeding hole corresponding to the bottom opening of the discharging bin 1, the other end of the screw box opposite to the motor is connected to a discharging pipe 24, and the discharging pipe 24 is fixed on the frame 30.

In the blanking process, as shown in fig. 1 and fig. 2, the control drawing plate 106 is opened, the material falls into the screw box 201 of the screw conveyor 2 from the blanking bin 1, the motor 204 is started, the conveying screw 202 rotates along with the motor, the material is conveyed to the blanking pipe 24 at the end part, and falls into the mixing hopper 3 below from the blanking pipe 24.

The housing 30 serves as a frame of the apparatus for fixing and supporting the other respective components. The weighing module 4 is fixed on the frame 30, and the discharging bin 1 is arranged on the weighing module 4. The bottom of the mixing hopper 3 is provided with an opening, and the opening and the closing of the opening are controlled by a blanking valve 5. The mixing hopper 4 is positioned at the lower part of the blanking bin 1, and the centers of the plurality of screw conveyors 2 are distributed in a circular arc shape relative to the center of the mixing hopper 4. The inner wall of the mixing hopper 4 is provided with a mixer 13 for uniformly mixing the multi-component materials.

Referring to fig. 2 and 4, the controller 9 is operated in a touch mode, and the touch screen thereof is provided with a human-computer interface for setting the formula of the multi-component material and other parameters, wherein the formula comprises the total weight of one blanking and the percentage of each component in the weight. The controller 9 is connected with each sensor and the action part through a signal acquisition module 91 and an output module 98 respectively.

The mixing bunker 6 is located below the discharge valve 5 and has a push plate 7 at its bottom, below which a feed pipe 8 is connected, which feeds the multicomponent mixture to a packaging bag or a production facility.

Preferably, a mixing level sensor 23 is mounted on the side wall of the mixing silo 6, and a refiner 22 is arranged in the mixing silo, wherein the refiner 22 is a spiral blade. The capacity of the mixing bunker 6 is 15 times of that of the mixing hopper 3, after a plurality of one-time baiting is completed, the controller 9 reads the state of the mixing material level sensor 23, if the material level of the mixture is detected to exceed a set threshold value, the homogenizer 22 is controlled to rotate to stir the mixture uniformly again, and under the control of the controller 9, the push plate 7 is opened, and the mixture is output from the material conveying pipe 8.

FIG. 3 illustrates the effect of weight loss on weighing during material fall at a velocity v₀The screw conveyor 2 falls down from the screw conveyor 2, the screw conveyor 2 and the lower bin 1 are carried on a weighing module 4, and the change of the mass equivalent of the material in the lower bin 1 measured by the weighing module 4 can be represented by the following formula:

wherein Gs is the initial weight at zero time, and at t time, dm is the blanking mass (g/s) per unit time of the outlet of the screw conveyor, v₀At an initial velocity of the material as it falls, Δ m of material leaves the screw conveyor 2 within a time Δ t.

As shown in fig. 2 and 3, the controller can dynamically read the current reading of the weighing module in real time, but during the feeding process, the reading is reduced by a value which is not the weight of the material actually falling into the mixing hopper, but includes the reverse impact effect of the falling material. The impact effect is therefore deducted when calculating the amount of material discharged. However, in practice, how to accurately obtain the equivalent weight value of the impact is a difficult problem to solve.

As can be seen from equation (1), the mass of the material detected by the weighing module not only includes the second term in the actually falling material, but also is affected by the momentum reverse impact caused by the falling of the material with non-zero velocity, i.e. the third term in the equation, wherein the third term is the weight loss effect term. Therefore, the method of directly reading the weight in the conventional weightless scale cannot obtain the actual weight at a certain moment in the discharge bin.

In order to obtain the exact mass of the material in the blanking bin at the present moment, the influence of the weight loss effect needs to be considered, and a third term value in the formula (1), namely an equivalent weight loss value of the falling material, needs to be obtained.

Through carrying out repeated experiment test and analysis to weightless formula unloading process, conclude screw weightless formula material blanking machine, the most main influence factor of its unloading in-process whereabouts material loss weight value includes: discharging binMaterial level h, blanking rate D, material density rho, diameter D of helical blade of helical conveyor, pitch S and maximum rotating speed v of screw_R. Falling material loss values are complex non-linear mappings of these physical quantities. In order to accurately predict the weight loss value of falling materials in different states and accurately feed materials by adjusting the closing time of the valve, the mapping relation needs to be identified.

Based on the mapped complex nonlinear characteristics and considering the close connection existing between the weight loss values of falling materials in two continuous sampling periods, the invention adopts dynamic recursive Elman neural network modeling to carry out dynamic recursive Elman neural network modeling on the weight loss values of the falling materials, the material level h of a blanking bin, the blanking rate D, the material density rho, the diameter D of a helical blade of a helical conveyor, the pitch S and the maximum rotating speed v of a screw_RThe mapping relationship between the two is identified.

As shown in fig. 4, the controller 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. The neural network module 93 adopts an Elman neural network, and the storage module 95, i.e. a memory, is used for storing data.

As shown in fig. 5, the used Elman neural network has a recursive structure, compared with the BP neural network, the Elman neural network further includes a receiving layer in addition to an input layer, a hidden layer and an output layer, and the receiving layer is used for feedback connection between layers, so that the receiving layer can express time delay between input and output and parameter timing characteristics, and the network has a memory function. In fig. 4, the input layer of the established neural network has 6 units, the number m of nodes of the hidden layer and the accepting layer can be selected from 11 to 20, for example, 16 is selected, and the output layer has only one unit.

The neural network model is:

xc_k(t)＝x_k(t-mod(k，q)-1) (2)，

wherein mod is a remainder function, and f () is a sigmoid function; xc_k(t) is the carry layer output, x_j(t) is the hidden layer output, u_i(t-1) and y (t) are input layer input and output layer output, ω_j、ω_jkAnd ω_jiRespectively, the connection weight from the hidden layer to the output layer, the connection weight from the accepting layer to the hidden layer and the connection weight from the input layer to the hidden layer, theta and theta_jOutput layer and hidden layer thresholds, respectively; k is 1, 2 … m, q is the selected regression delay scale, and is preferred according to the sampling period and the blanking rate, and if the selectable q is 3; j is 1, 2.. m, i is 1, 2.. 6.

As shown in fig. 2 and 4, preferably, the controller 9 may further implement switching between the first connection array and the second connection array through touch screen operation, so that the controller operates in an offline training mode or an online prediction mode.

In the neural network established by combining fig. 4 and fig. 5, the input layer comprises 6 nodes, wherein the material density ρ, the diameter D of the spiral blade of the spiral conveyor, the pitch S and the maximum rotating speed of the screw are determined values, and the other 2 quantities need to be dynamically acquired in real time through the signal acquisition module. By introducing a plurality of different delay regression carrying layer nodes into the network, the network structure is more appropriate to the blanking process, and the network training is faster to converge.

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.

In order to obtain a training sample, after the feeding is started, when a continuous material flow is formed between materials from a screw conveyor at the bottom of a feeding bin and a mixing hopper, the feeding is continued for a period of time, a dynamic weight reading W of a weighing module is read in real time when a valve is closed, a static weight reading WD of the weighing module is read after the materials fall, and the actual weight loss value of the falling materials under the state of the moment of closing the valve is L-WD-W, wherein the actual weight loss value of the falling materials under the state of closing the valve is L-WD-The actual value of the value, i.e. the sample output value y, i.e. the desired value y_d。

The neural network training adopts a gradient descent method, and the weight and threshold value adjusting method in the training is as follows.

Assuming a total of P training samples, let the error function be:

then the adjustment of the weight from the hidden layer to the output layer is shown as follows:

ω_j(t+1)＝ω_j(t)+Δω_j(t+1) (6)，

wherein the content of the first and second substances,

δ_y＝-(y_d-y)·y·(1-y) (8)，

the adjustment formula of the output layer threshold is as follows:

θ(t+1)＝θ(t)+Δθ(t+1) (9)，

wherein the content of the first and second substances,

similarly, the adjustment formula of the connection weight from the input layer to the hidden layer is:

ω_ji(t+1)＝ω_ji(t)+Δω_ji(t+1) (11)，

wherein the content of the first and second substances,

δ_j＝δ_y·ω_j·x_j(t)·(1-x_j(t)) (13)，

the adjustment formula of the hidden layer threshold is as follows:

θ_j(t+1)＝θ_j(t)+Δθ_j(t+1) (14)，

wherein the content of the first and second substances,

irrespective of xc_k(t) pairs of connection weights ω_jkThe adjustment formula of the connection weight from the bearer layer to the hidden layer is as follows:

ω_jk(t+1)＝ω_jk(t)+Δω_jk(t+1) (16)，

wherein the content of the first and second substances,

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

The training end condition may be set to a total error or a variation thereof smaller than a set value or a number of times of training up to a certain amount.

Preferably, in order to make the training sample cover more cases, each time the valve is closed can be set to a random value from the moment the auger is started or the moment the weight of the weighing module is read to some determined value.

Before network training, 6 input quantities and 1 output quantity are subjected to normalization preprocessing:

r′＝r-r_min/r_max-r_min(18)，

wherein r is an unprocessed physical quantity, r' is a normalized physical quantity, r_maxAnd r_minRespectively the maximum and minimum values of the sample data set.

When calculating the weight loss predicted value of the falling material, converting the network output quantity into the weight loss value of the falling material by the following formula:

r＝r_min+r′·(r_max-r_min) (19)。

when the feeding is controlled on line, the first connection array is disconnected, the neural network predicts the weight loss value yL of falling materials and outputs the weight loss value yL to the processing module through the second connection array, and the processing module processes and analyzes the weight loss value yL and then closes the valve of the spiral conveyor at the opening at the bottom of the feeding bin through the output module:

assuming that the one-time blanking amount of the current components is Ws, when blanking is started, the controller obtains the initial static weight of the blanking bin as G0 by reading the sensing value of the weighing module; the controller then continuously reads the sensor value of the weighing module and shuts off the auger when the dynamic weight reading reaches (G0-Ws-yL).

Preferably, in addition to the predicted value of the weight loss value of the falling material, the current accumulated blanking error of the material of the present component is compensated, that is, when the dynamic weight reading of the blanking bin is detected to reach (G0-Ws-yL + E), the screw conveyor is closed, wherein E is the current accumulated blanking error of the present component, and when E is positive, the excessive blanking is indicated.

Preferably, the controller controls the operating speed of the screw conveyor as follows:

wherein Ws and Wa are respectively the predicted values of the once feeding amount and the weight loss of the falling material of the current material, d is the feeding rate of the screw conveyer when the screw rotates at the maximum speed, and t_sFor the length of deceleration stop time, t_s＝λ·v_R/μ·a_max。

When the training sample is obtained off-line and the neural network is quoted on-line, the same rotating speed control is adopted for the spiral conveyor.

The signal acquisition module respectively acquires sensing signals of the material level and the weight of falling materials of the blanking bin in real time through a bin level sensor in the blanking bin and a weighing module bearing the blanking bin, the sensing signals are transmitted to the processing module for data preprocessing, then the data are input to the neural network, the output value of the neural network and the expected output value preprocessed by the processing module are transmitted to the iterative learning module through the first connecting array, and the iterative learning module transmits the adjusted weight back to the neural network according to a gradient descent method.

The bin level sensor adopts a distance sensor to detect the material level height of materials in the lower bin, and the processing module can calculate the falling mass equivalent of the materials in unit time, namely the falling rate, by periodically and continuously acquiring signals of the weighing module.

As shown in fig. 2 and 6, preferably, in order to reduce the influence of the action of the screw conveyor 2 on the weighing, a buffer tank 101 is arranged at the bottom of the lower bin, and comprises a damper 103 and an umbrella-shaped body 102. The damper 103 is a flexible connection section that reduces the vibration transmitted to the weighing module when the screw conveyor 2 is operated. The umbrella 102, in turn, includes a cap 104 and a frame 105 supporting the cap.

The formula (1) can also be used for analyzing that the weight loss value of falling materials is closely related to the falling conditions and is influenced by the shape distribution of the materials in the lower storage bin 1.

The granular materials flow out of the feed bin under the action of gravity mainly in two types of bulk flow and central flow. The whole particle layer in the overall flow type middle storage bin can flow out approximately uniformly, and basically every particle moves; while some particles in the flow pattern of the core stream are stationary, there is a flow channel boundary between the flowing and stationary particles. The integral blanking rate of the integral flow is larger than that of the central flow, the fluctuation of the blanking rate is small, and the flow is stable. In the actual production process, the material in the discharging bin may have a central flow pattern, so that when the material is discharged from the material opening, the material is firm and plate-formed due to the compaction stress effect generated by bin pressure.

As shown in the combined drawings of 6-8, preferably, in order to reduce the fluctuation range of the blanking rate in the blanking bin and better predict the loss of weight value of the falling materials, a distance sensor type bin sensor and a rotatable vibrating rod are adopted in the blanking machine to detect and adjust the material accumulation form in the blanking bin, so that the dynamic material arch is alternately formed and collapsed above the blanking opening, and the stability of the material compactness and the blanking form is ensured.

As shown in FIG. 8, the discharging bin 1 continuously discharges materials, and when the material level in the bin is reduced to a certain value, the materials need to be supplemented. For this purpose, a storage bin 10 is arranged above the lower bin 1, and the material in the storage bin 10 is fed into the lower bin 1 through a feed pump 11 and a feed pipe 15. In order to uniformly feed material particles, a material spray nozzle 16 is arranged at an outlet at the tail end of the feeding pipe 15, the surface of the material spray nozzle 16 is in a spherical shape, round small holes 17 are distributed on the surface of the material spray nozzle, and the aperture of each small hole is optimized according to the granularity of the material. The feed pump 11 is a screw conveyor, and its operation is controlled by a controller.

In the blanking process of the blanking bin 1, along with the reduction of the material level surface 19, the feeding pump 11 acts under the control of the controller, so that the material level of the top surface of the material in the blanking bin is kept near a preset value, and the rotating speed of the feeding pump is controlled according to the following formula:

wherein, V_{0 feeding}A set maximum feeding speed, L is the current material level of the feed bin, L_MAnd L_mAre respectively preset at L_TThe highest and lowest levels nearby.

In fig. 8, the two drawings are respectively observed from the side view and the top view of the lower silo 1, as shown in fig. 8a and 8b, a bin sensor 12 is installed on a vertex angle of the lower silo 1 near the center of the frame, and the bin sensor has a rotatable base capable of pitching and rotating, so that the bin sensor can detect materials in different directions of the stop pointing points 20, and each stop pointing point 20 forms a scanning line 21 close to a concentric circle, thereby judging the distribution of the discharge level surface 19.

As shown in fig. 7, the controller improves the distribution of the material by installing a vibrating rod 18 on the sidewall of the lower bin 1 by the blanking machine. The vibrating rod 18 comprises a support 181, a holder 182, a vibrator 183 and a vibrating rod 184 which are connected in sequence, a spring buffer is arranged at the bottom of the vibrator 183, particle protrusions 185 are distributed on the surface of the vibrating rod 184, and the holder 182 can pitch and rotate so that the vibrating rod 184 can do curvilinear motion in the lower storage bin 1.

As shown in the combined drawings of fig. 7 and 6, the umbrella-shaped body in the buffer pool at the bottom of the discharging bin can bear the pressure of the upper bin, weaken the action of larger compaction force near the discharging port, greatly reduce the pressure of the bin below the umbrella cap, and simultaneously form an annular material flow port at the periphery of the umbrella-shaped body, so that the materials in the bin tend to flow integrally, and the arching and blocking of the materials can be further prevented.

In the blanking process, the controller judges the distribution of the materials in the blanking bin through the detection of the bin level sensor and the tracking of the blanking rate in unit time respectively, so that the material level in the blanking bin keeps an approximate parabolic surface shape. With reference to fig. 8 to 9, when the materials are uniformly distributed, the distance values of the materials detected by the bin sensors in different directions are approximately concentrated in a smaller range after geometric transformation of the detection ray and the inclination angle in the vertical direction. When the material is locally hardened or is arched stably, the detected distance value exceeds the range. Meanwhile, the feeding speed of each feeding bin is tracked in real time through the weighing module. When the distance sensor detects the abnormal state or finds that the fluctuation of the blanking amount in unit time exceeds a set threshold value, such as 5%, the controller commands the vibrating rod to act, the vibrating rod starts to pass through a high-point area of the material level to a low-point area of the material level from a starting point through the operation of the pan-tilt head to perform snake-shaped stirring, and a vibrating rod track 126 of the tail end of the vibrating rod 124 in the blanking bin 1 is shown in fig. 7; meanwhile, the vibrator starts to vibrate, and the particle protrusions on the vibration rod drive peripheral particles, so that hardening or material arch which is occasionally formed is broken, and the material distribution is recovered to be uniform. Through dynamic detection and control of material distribution, fluctuation of material compactness is reduced, and therefore stability of filling amount in unit time is guaranteed. And stopping feeding while the vibrating rod acts, and closing the drawing plate.

As shown in figure 9, the invention greatly weakens the compaction force effect generated by the charging impact by the detection and action coordination of the distance sensor and the vibrating rod, effectively prevents the granularity segregation of the materials in the bin, activates the materials in the lower bin and improves the distribution of the materials.

Fig. 10 is a supplementary record of the multi-component material feeding process, in which a schematic diagram of the material distribution in the mixing hopper at the time of feeding 4 components is shown.

Referring to fig. 2 and 11, when the multi-component material falls into the mixing hopper 3, the mixer in the mixing hopper 3 operates to mix the material. As shown in fig. 11, the mixer 13 includes a mixing base 131, two mixing arms 132, and a mixing arm rotating shaft 133, a mixing claw rotating shaft 134, and a mixing claw 135 which connect the two mixing arms 132 in sequence, wherein the mixing base 131 also includes a rotating shaft.

An arc wedge block 301 is arranged on the inner side of the bottom of the openable component at the opening at the bottom of the mixing bin 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 repeatedly turns in a spiral shape from the left side to the right side of the mixing hopper, from high to low and then to high, and mixes and stirs the multi-component material. And a vibrating table 14 is arranged outside the non-opening side of the mixing hopper, the controller controls the vibrating table 14 to vibrate while the mixer 13 acts, and the multi-component materials in the mixing hopper can be fully and uniformly mixed under the action of the mixer 13 and the vibrating table 14.

The blanking machine is applied to blanking, the blanking behaviors of the screw conveyors are modeled separately in an off-line mode, and each component of material is blanked separately in the process of collecting samples, so that all the materials can be recycled without waste. The material level and the blanking rate of the blanking bin are periodically acquired during actual blanking, and the weightlessness value of the current falling material can be forecasted in real time, so that accurate blanking can be realized from the first batch, and the fluctuation of other blanking errors in an online iterative learning scheme is avoided.

Example 2:

referring to fig. 2 and 12, when the multi-component material falls into the mixing hopper 3, the mixer in the mixing hopper 3 operates to mix the material. 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 mounted on the mixing shaft 137, and a mixing bracket 136 fixed on the inner wall of the mixing hopper 3 for supporting the mixing shaft 137. The mixing turntable 138 is similar to a water wheel vehicle in ring shape, and rectangular blades which are basically vertical to the circumference are arranged on the outer ring of the mixing turntable, and holes can be formed in the blades. The helical blades 139 are irregular helical blades with holes distributed in the blades.

An arc wedge block 301 is arranged on the inner side of the bottom of the openable component at the opening at the bottom of the mixing bin 3, under the action of the controller, the mixing rotating shaft 137 of the mixer 13 rotates, and the rectangular blade and the helical blade 139 in the mixer turn over the materials.

And a vibrating table 14 is arranged outside the non-opening side of the mixing hopper, the controller controls the vibrating table 14 to vibrate while the mixer 13 acts, and the multi-component materials in the mixing hopper are fully and uniformly mixed under the action of the mixer 13 and the vibrating table 14.

Example 3:

referring to fig. 2, in order to measure the weight loss of the blanking bin 1, two horizontal supporting parts can be led out from the outer side wall of the blanking bin; the weighing module is horizontally placed, and the weighing module supports the lower storage bin from two sides in the vertical direction. Or two suspension parts are led out from the top of the lower feed bin 1, the weighing module is horizontally placed, and the lower feed bin is supported by the weighing module from two sides in the vertical direction.

The above-described embodiments do not limit the scope of the present invention. Any modification, equivalent replacement, and improvement made within the spirit and principle of the above-described embodiments should be included in the protection scope of the technical solution.

Claims

1. A screw weightless material blanking machine controller based on a neural network comprises a signal acquisition module, a processing module, a neural network module, a storage module and an output module, wherein the signal acquisition module acquires sensing signals of the material level and the weight of a 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 a 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 6 input quantities of material level, blanking rate and material density of a blanking bin, the diameter and the pitch of a helical blade of a helical conveyor and the maximum rotating speed of a screw 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 predicted values to the processing module, and the processing module processes and analyzes the predicted values and controls the closing of the spiral conveyor at the opening at the bottom of the blanking bin through the output module:

continuously reading the sensing value of the weighing module, and closing the screw conveyor 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 the feeding bin when feeding is started, and yL is the lost weight value of the falling materials predicted by the neural network.

2. The screw weight loss type material blanking machine controller based on the neural network is characterized by further comprising an iterative learning module, a first connecting array and a second connecting 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 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 screw weight-loss type material blanking machine controller based on the neural network as claimed in claim 1, wherein the screw conveyor is controlled to close, the current accumulated blanking error of the material of the present component is compensated, and the screw conveyor is closed when the dynamic weight reading of the blanking bin is detected to reach (G0-Ws-yL + E), wherein E is the current accumulated blanking error of the present component, and the condition that E is positive indicates that blanking is excessive.

4. The screw weight-loss type material blanking machine controller based on the neural network as claimed in claim 1, wherein the number of the neural network modules is multiple, and each screw conveyor has a corresponding neural network module.

5. The screw rod weight loss type material blanking machine controller based on the neural network as claimed in any one of claims 1 to 4, wherein the model of the neural network is as follows:

xc_k(t)＝x_k(t-mod(k，q)-1)，

6. The screw weight loss type material blanking machine controller based on the neural network is characterized in that the controller also improves the distribution of the materials through a vibrating rod arranged on the side wall of the lower storage bin.

7. The screw weight-loss type material blanking machine controller based on the neural network is characterized in that when the distance values of the materials detected by the distance sensors in different directions are geometrically transformed and exceed a set range, the controller commands the vibrating rod to act.

8. The screw weight loss type material blanking machine controller based on the neural network as claimed in claim 6, wherein when the fluctuation of blanking amount per unit time is found to exceed a set threshold value, the controller commands the vibrating rod to act.

9. The neural network-based screw weight loss type material blanking machine controller is characterized in that the controller also controls the action of the feeding pump, so that the material level of the top surface of the material in the blanking bin is kept near a preset value.