Airflow cleaning device for material in cleaning screen chute and detection method

Abstract

The application provides a kind of airflow cleaning device and detection method of material in cleaning screen chute, including rack, cleaning screen, chute slide rail, duck type jet head, gas control valve, signal acquisition module and diagnostic alarm module;The both sides of the rack are respectively provided with chute slide rail, and the cleaning screen is movably installed in the chute slide rail;Several duck type jet heads are installed on the side wall of the rack above the chute slide rail, and the several duck type jet heads are communicated with gas source through the gas control valve;The signal acquisition module includes infrared sensor and strain gauge;The detection point of the infrared sensor and the strain gauge is correspondingly provided with a duck type jet head;The diagnostic alarm module collects the signal detected by the infrared sensor and the strain gauge, analyzes the collected signal, and determines the point where the material in the chute slide rail exists;The diagnostic alarm module controls the gas control valve to make the duck type jet head near the point work;The application can clean the material in the chute slide rail.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of agricultural machinery technology or intelligent agricultural machinery, in particular to a device for removing materials in a chute of a cleaning screen and a detection method. BACKGROUND

[0002] The track-type combine harvester is a commonly used agricultural harvesting equipment, and the wind screen type cleaning device is one of the important devices of the track-type combine harvester, which completes the cleaning work. The cleaning screen is a working part of the wind screen type cleaning device. The pulleys at the front end of the cleaning screen are arranged in the chute on both sides through the slide rail on the side wall of the machine box. The reciprocating movement of the cleaning screen is realized by the eccentric device at the rear end. The longitudinal axis flow threshing device is generally used for the threshing cylinder of the combine harvester, which is easy to cause the uneven distribution of the threshing cylinder. The uneven distribution of the materials on both sides of the cleaning screen is caused, and the grain and light impurities are inevitably thrown out of the screen box. The grain and stem materials are easily ejected to both sides of the screen box. Therefore, the side of the cleaning screen with more materials may have a greater probability of causing the materials to be ejected out of the screen box, enter the slide rail and chute fixed on the side wall of the machine box, cause the pulley to bounce or be blocked, and cause the imbalance of the pulley on the left and right sides under the long-term work of the cleaning screen, thereby reducing the service life of the screen box.

[0003] The invention patent CN106717594B relates to a combine harvester cleaning screen inclination automatic adjusting device, which is composed of a linear slide, a servo motor cleaning screen, an inclination sensor device, a control system and the like. The inclination sensor is installed at the tail of the cleaning screen, and the linear slide is installed vertically. The roller guide plate is installed on the working slide block of the linear slide. Under the driving of the servo motor, the working slide block is moved, and the roller guide plate is further moved, thereby realizing the adjustment of the inclination of the cleaning screen. The inclination sensor, the control system and the linear slide of the invention can realize the real-time regulation and control of the inclination of the cleaning screen, can make the cleaning screen adapt to the harvesting conditions of the field, especially the undulating terrain in the hilly area, can ensure the normal work of the cleaning screen, and is conducive to improving the cleaning efficiency and the overall efficiency of the machine. However, the change of the distance between the screen surface with the inclination of the cleaning screen and the threshing cylinder is not considered, which easily causes the materials to enter the chute slide rail, causes the imbalance of the pulley work, and finally affects the stability of the cleaning screen operation.

[0004] The invention patent CN110294286A provides a material blockage detection device, wherein the material blockage detection device comprises a sensing probe rod with one end extending into the material flow; a rotating shaft rotationally connected with the sensing probe rod; a proximity switch selectively in contact with the other end of the sensing probe rod; a controller electrically connected with the proximity switch and comprising a timing module and an alarm module. The falling material can move one end of the sensing probe rod to make the other end of the sensing probe rod out of contact with the proximity switch. The invention converts the impact force generated by the material during falling into an electrical signal. Once blockage or material falling is blocked, and the blockage or blockage time is longer than the preset value, the alarm module alarms, which is beneficial to timely discover abnormal conditions at the material blockage detection device, avoid accident expansion, improve the smoothness of the operation link, the sensing probe rod is not easy to damage, and the service life of the material blockage detection device is long. However, the material blockage detection device is not protected, and the space for material falling is occupied. The material blockage detection device will be affected by the material flow, which makes the sensing probe rod easy to rotate, causing the detection to be not sensitive or not accurate, thereby affecting the detection effect and accuracy.

[0005] The invention patent CN117036362A provides a material blockage detection method, device, equipment and storage medium. The material blockage detection method first acquires detection data in the charging device and target image data of a specified area. The detection data in the charging device includes material height data at the last time, material height real-time data and material detection real-time data in the charging device. The specified area includes an area outside the charging device. The first identification result is obtained by comparing and analyzing the detection data in the charging device. The target recognition model is called to perform image recognition on the target image data, and the second identification result is generated. If at least one of the first identification result and the second identification result indicates blockage, the target blockage information is generated. The method realizes multi-angle and multi-direction detection, reduces the adverse effects of single detection factors, and improves the accuracy of blockage detection. However, the invention does not filter the original data, which may affect the accuracy of the detection data. Directly using image processing in the relatively harsh cleaning room of the combine harvester is not easy to realize, has low reliability, greatly affects the detection result, and thus it is difficult to discover and timely handle the blockage condition.

[0006] The utility model patent CN206927269U discloses a kind of pneumatic unblocking device for avoiding and eliminating inclined chute blockage, it is by gas source system, control airflow pipeline system, working airflow pipeline system, electric control system and jet system combination constitutes, and unblocking principle is that compressed air is delivered to jet system from gas source system, and the bottom plate nozzle that is parallel to the bottom plate of inclined chute is downwardly jetted and the side plate nozzle that is vertically side plate is inwardly jetted, compressed air is jetted to the inside of inclined chute, to instantaneously eliminate the friction of material and the inner wall in the inclined chute and blow apart cohesion between material, so that material is released from restraint and freely falls under its own gravity, with The characteristics that it is easy to operate and reliable in performance, avoid the security risk brought by manual cleaning, effectively prevent the problem of blockage caused by material sticking to the wall in the inclined chute, but the pneumatic unblocking device needs larger air pressure, needs compressed air machine to supply air, without using detachable gas tank to supply air, And larger air pressure can cause damage to the structure and material of inclined chute, in addition, there is no material amount detection system for inclined chute blockage, to facilitate direct accurate point and release a certain amount of gas to remove material, in the case of frequent use of unblocking device, nozzle and pipeline parts need to be replaced and repaired regularly, The structure is relatively complex, and it is not convenient to replace.

[0007] The utility model patent CN214852915U provides a kind of corn kernel cleaning screen, including screen, frame, the screen is divided into multiple-stage screen, present ladder-shaped arrangement, screen hole is regular hexagon, upper screen and lower screen are connected using bent plate, and metal reinforcing rib is fixed in bent plate, with upper screen connection position downside;The frame is fixed with L-shaped reinforcing plate in the end of the uppermost screen, and C-shaped reinforcing plate is fixed in the end of the lowermost screen, both side frames are L-shaped, the screen hole of the utility model is regular hexagon, can enhance the cleaning function of screen, improve the cleaning efficiency of corn, effectively reduce the problems such as screen box throwing grain, corn straw blockage;Bent plate bottom is provided with metal reinforcing rib, can improve the service life of screen, the waist-shaped hole of both sides of frame can conveniently adjust and install screen position, but it is for object that hoist rod type cleaning screen is not considered, and it is inevitable that screen box throws grain under eccentric wheel driven slide groove type cleaning screen, material has great probability to bounce out of screen box, enter slide rail and slide groove, cause the gravity center of both sides of pulley to be not identical, imbalance of cleaning screen can be caused under long time work, further cause fault, reduce the service life of screen box.

[0008] The utility model discloses CN201309745Y provides a kind of adjustable belt conveyor chute material block detection device, including with one small switch plate with ear hole Angle with big switch plate connection integral, pass through a ladder shaft in the ear hole of the small switch plate, through the both ends of ladder shaft installation setting on a pair of support on chute, on the side of the big switch plate respectively be Angle with counterweight and switch support, then pass through a pair of support installation on chute, with not occupy other equipment space, also not affect the operation of other equipment, and convenient maintenance, this device can effectively detect chute material block, avoid due to chute material block and crush belt, cylinder and drive system, simultaneously also avoid due to chute material block and cause chute extrusion and deformation, but the adjustable belt conveyor chute material block detection device of this technology can be affected by material flow, leading to the rotation of small switch plate and big switch plate is not sensitive or inaccurate, without sensor detection precision is high, affect the effect and accuracy of detection, simultaneously can increase the complexity and instability of chute, leading to the chute of belt conveyor occurs greater vibration, affect the uniformity and stability of belt conveyor chute. SUMMARY

[0009] In view of the deficiencies in the prior art, the present application provides a gas flow cleaning device for cleaning materials in a cleaning screen chute and a detection method. The duck-shaped air jets are installed on the side walls of the frame above the chute slide rails, which can effectively and accurately blow air at the point to prevent materials from accumulating in the chute and slide rails and affecting the normal operation of the cleaning screen.

[0010] The present application achieves the above technical objectives through the following technical means.

[0011] A gas flow cleaning device for cleaning materials in a cleaning screen chute, comprising a frame, a cleaning screen, a chute slide rail, duck-shaped air jets, a gas control valve, a gas source, a signal acquisition module, and a diagnostic alarm module.

[0012] The chute slide rails are installed on both sides of the frame, and the cleaning screen is movably installed in the chute slide rails. The duck-shaped air jets are installed on the side walls of the frame above the chute slide rails. The duck-shaped air jets are connected to the gas source through the gas control valve, and are used to clean the materials in the chute slide rails.

[0013] The signal collection module includes infrared sensors and strain gauges. The infrared sensors are distributed on the side of the chute slide rail to detect whether there is material in the chute slide rail. The strain gauges are distributed on the bottom of the chute slide rail to detect the deformation of the bottom of the chute slide rail. The detection point of the infrared sensor and the strain gauge is near the corresponding duck jet head. The diagnostic alarm module collects the signals detected by the infrared sensor and the strain gauge. The diagnostic alarm module analyzes the collected signals to determine the point where the material in the chute slide rail is located. The diagnostic alarm module controls the gas valve to make the duck jet head near the point work to clean the material in the chute slide rail.

[0014] Further, the duck jet head includes an air inlet, a rectifier chamber, a grid strip and a duck nozzle. The rectifier chamber is provided with an air inlet which is gradually expanded along the flow direction. The air inlet is communicated with the gas valve. The rectifier chamber is provided with a plurality of grid strips which are distributed at intervals. The other side of the rectifier chamber is provided with a duck nozzle which is gradually tapered along the flow direction.

[0015] Further, one end of the duck nozzle is embedded in the internal space of the chute slide rail, and the distance from the bottom of the chute slide rail is not more than 5 mm.

[0016] Further, the gas valve includes a gas distribution head, a valve, a gas distribution chamber and a gas delivery head. A plurality of gas distribution heads are uniformly distributed on the gas delivery head. Each gas distribution head is provided with a valve. The inlet of the gas delivery head is communicated with the gas source. Each gas distribution head is communicated with the corresponding duck jet head. The diagnostic alarm module selectively controls the valve to open to make the duck jet head near the point work.

[0017] Further, the sensing head of the infrared sensor extends into the internal space of the slide rail. The infrared sensor is in a linear array and is installed above the duck jet head to detect whether there is material at each point of the slide rail. A plurality of strain gauges are installed at the bottom of the chute. When the material falls into the chute, the displacement of the pulley changes, causing the strain gauge to deform.

[0018] Further, the diagnostic alarm module includes a signal processing and diagnosis unit and a display alarm unit. The signal processing and diagnosis unit learns and diagnoses the signals collected by the infrared sensor and the strain gauge to establish a deep learning diagnosis model. The deep learning diagnosis model outputs the point information of the material. The display alarm unit alarms according to the point of the material output by the deep learning diagnosis model. The display alarm unit controls the duck jet head corresponding to the point to be communicated with the gas source for fixed-point removal of the material.

[0019] A detection method of a gas flow cleaning device for cleaning material in a cleaning screen chute, comprising the following steps:

[0020] Several infrared sensors are distributed on the sides of the chute rail to detect whether there is material in the chute rail; several strain gauges are distributed on the bottom of the chute rail to detect the deformation at the bottom of the chute rail; the infrared sensors and strain gauges constitute an n-point signal input diagnostic alarm module, where the original data matrix Z∈[X1,X2,…,X…] of the n-point signals is... j ,…,X n ], where X j Let j be the signal at point j, where j∈[1,n];

[0021] Normalize all the original data matrices Z so that the matrix elements are between [0, 1] to obtain the normalized data matrix M;

[0022] The normalized data matrix M is then subjected to noise processing by adding a Gaussian noise distribution q. D (μ0,σ), where μ0 is the distribution center and σ is the variance of the noise, to obtain the noisy data matrix.

[0023]

[0024] in: This represents the normalized voltage signal at the j-th point of the i-th sampling point after adding noise, where m represents the total number of sampling points and n represents the number of points.

[0025] According to the formula For the input noisy data respectively Encode to obtain high-level features h and f θ Let represent the functional relationship used in the encoding process, where s is the Sigmoid activation function, W is the weight matrix of the encoder, and b is the bias vector of the encoder.

[0026] According to the formula Decode the high-level features h of the noisy data to obtain the reconstructed output of the original data. Where g a The functional relationships used in the decoding process are: s is the Sigmoid activation function, C′ is the encoder weight matrix, and d′ is the encoder bias vector; C′=W T ;

[0027] Through the objective function For noisy data matrices and reconstructing the data matrix Minimize the error between them to obtain the optimal parameter r, r = {W, C′, b, d′};

[0028] The reconstructed data matrix As the filtered input signal, extract The peak value, mean value and standard deviation of the peak are taken as input features, and the feature matrix is obtained after splicing the features Wherein, fz n,k The kth peak value feature of the nth point is represented as jz n,k The kth mean value feature of the nth point is represented as bzc n,k The kth standard deviation feature of the nth point is represented as;

[0029] The feature matrix C is divided into a test set and a training set, and is used as the input layer of the BP neural network. The number of neurons is set to 3, the number of neurons in the hidden layer is (2 / 3)*(n+4), the activation function is ReLU, the output layer of the BP neural network is set to n+1 according to the point state label, and the point state label includes no material and jth point material. The BP neural network is performed to obtain a slide rail chute material deep learning diagnosis model.

[0030] The slide rail chute material deep learning diagnosis model is used to determine the point where the material is located, and the corresponding duck-shaped jet head is connected with the gas source to control the point removal of the material.

[0031] The beneficial effects of the present application are as follows:

[0032] 1. The air flow cleaning device and detection method for the material in the cleaning screen chute, the structure of the duckbill type air jet head in the cleaning screen chute slide rail anti-blocking device is small and convenient to install, does not occupy too much space, does not interfere with the movement of the cleaning screen, can effectively and accurately blow air at the point, prevents the material from appearing in the chute and the slide rail, and affects the normal work of the cleaning screen; the control valve and the motor are used to control the gas flow direction, the gas flow can be accurately controlled at the point, the gas tank is used for gas supply, the device is convenient to place on the side wall of the machine case, and the air supply is convenient.

[0033] 2. The air flow cleaning device and detection method for the material in the cleaning screen chute, by establishing a slide rail chute material deep learning diagnosis model, the original signal can be quickly filtered and diagnosed and classified, the signal real-time acquisition, processing diagnosis and alarm display system established by LabView can clearly understand the material position and timely blow air to prevent blocking, thereby reducing the failure rate of the cleaning screen and ensuring the work efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0034] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiment or prior art description. The drawings in the following description are some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0035] Figure 1Axonometric view of the airflow cleaning device for the material in the chute of the cleaning screen according to the present application.

[0036] Figure 2 Axonometric perspective view of the canard according to the present application.

[0037] Figure 3 Axonometric view of the gas control valve and motor according to the present application.

[0038] Figure 4 Front view of the cleaning screen box according to the present application.

[0039] Figure 5 Sectional view of the cleaning screen box according to the present application.

[0040] Figure 6 Partial enlarged view of the chute slide rail according to the present application.

[0041] Figure 7 Point distribution diagram of the acquisition module according to the present application.

[0042] Figure 8 Flow chart of the material detection system of the cleaning screen chute slide rail according to the present application.

[0043] Figure 9 DAE-BPNN diagnosis program flow chart according to the present application.

[0044] Figure 10 DAE-BPNN diagnosis model schematic diagram according to the present application.

[0045] Figure 11 Interactive interface of the material detection system of the cleaning screen chute slide rail according to the present application.

[0046] In the drawings:

[0047] 1 - rack; 2 - cleaning screen; 3 - chute slide rail; 3-1 - chute; 3-2 - slide rail; 4 - canard; 4-1 - air inlet; 4-2 - rectifier chamber; 4-3 - grid strip; 4-4 - canard nozzle; 5 - gas control valve; 5-1 - gas distribution head; 5-2 - valve; 5-3 - gas distribution chamber; 5-4 - gas delivery head; 6 - signal acquisition module; 6-1 - infrared sensor; 6-2 - strain gauge; 7 - diagnosis alarm module. DETAILED DESCRIPTION

[0048] Embodiments of the present application are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout the drawings denote the same or similar elements or elements having the same or similar functions. The embodiments described below by reference to the drawings are exemplary and are intended to explain the present application, and cannot be understood as a limitation of the present application.

[0049] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "axial," "radial," "vertical," "horizontal," "inner," and "outer," etc., indicating orientation or positional relationships based on the orientation or positional relationships shown in the accompanying drawings, are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0050] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0051] like Figure 1 As shown, the airflow removal device for materials in the cleaning screen chute of the present invention includes a frame 1, a cleaning screen 2, a chute slide rail 3, a duck jet nozzle 4, an air control valve 5, a signal acquisition module 6, and a diagnostic alarm module 7.

[0052] The frame 1 has chute rails 3 on both sides. The cleaning screen 2 is movably installed in the chute rail 3. Generally, the pulley at the front end of the cleaning screen 2 cooperates with the chute rail 3. The reciprocating motion of the cleaning screen 2 is achieved by the rotation of the eccentric wheel and the displacement of the pulley. Several duck jet nozzles 4 are installed on the side wall of the frame 1 above the chute rail 3. The front half of the duck jet nozzles 4 extends into the internal space of the chute rail 3 through several holes in the side wall of the machine box. The duck jet nozzles 4 are connected to the air tank through the air control valve 5 and are used to clean the material in the chute rail 3. The duck jet nozzles 4, the air control valve 5 and the air tank are connected by an air pipe, which can spray high-speed airflow into the chute rail 3 to blow away the material splashed into the chute rail 3 by the cleaning screen.

[0053] The signal acquisition module 6 includes infrared sensors 6-1 and strain gauges 6-2, a plurality of infrared sensors 6-1 are distributed on the side of the chute rail 3, for detecting whether there is material in the chute rail 3; a plurality of strain gauges 6-2 are distributed on the bottom of the chute rail 3, for detecting the deformation of the bottom of the chute rail 3; the detection point of the infrared sensor 6-1 and the strain gauge 6-2 is provided with a canard jet head 4; the diagnostic alarm module 7 collects the signals detected by the infrared sensor 6-1 and the strain gauge 6-2, the diagnostic alarm module 7 analyzes the collected signals, and determines the point where the material in the chute rail 3 exists; the diagnostic alarm module 7 controls the gas valve 5 to make the canard jet head 4 near the point work, for cleaning the material in the chute rail 3, preventing the existence of material in the chute rail 3, causing the displacement of the pulley to be blocked, causing the operation of the cleaning screen to be unbalanced, and affecting the service life of the cleaning screen.

[0054] As shown in Figure 2 The canard jet head 4 includes an air inlet 4-1, a rectifier chamber 4-2, a grid strip 4-3 and a canard nozzle 4-4; one side of the rectifier chamber 4-2 is provided with the air inlet 4-1 which is gradually expanded along the flow direction, the air inlet 4-1 is communicated with the gas valve 5, the rectifier chamber 4-2 is provided with a plurality of spaced grid strips 4-3, the other side of the rectifier chamber 4-2 is provided with the canard nozzle 4-4, and the canard nozzle 4-4 is gradually tapered along the flow direction. One end of the canard nozzle 4-4 is embedded in the internal space of the chute rail 3, and the distance from the bottom of the chute rail 3 is not more than 5mm.

[0055] In the embodiment, the canard jet head 4 is 40mm long and 24mm wide, a plurality of canard jet heads 4 are linearly arranged with a spacing of 40mm, and are installed by punching and welding on both sides of the rack 1. The 10mm length of the canard nozzle 4-4 is embedded in the internal space of the chute rail 3, and the canard jet head 4 is 5mm away from the bottom of the chute rail 3. It is small and convenient, simple in structure, and can accurately point to the material in the internal space of the chute rail 3. The air inlet 4-1 is connected with the gas valve 5 through the air pipe to supply air to the canard jet head 4. The rectifier chamber 4-2 is rectangular, a plurality of grid strips 4-2 are welded in the middle of the internal space, connected with the air inlet 4-1, and rectify the high-speed airflow passing through, output more uniform high-speed airflow. The canard nozzle 4-4 is connected with the rectifier chamber 4-2, which imitates the shape of the duck bill through the principle of profiling, establishes the canard nozzle 4-4 with flat front end and wide rear end, and makes the rectified airflow enter the flat front end from the wide rear end, increases the pressure and speed of the airflow jet, and can more effectively remove the material falling into the internal space of the chute rail 3.

[0056] As shown in Figure 3As shown, the gas control valve 5 includes a gas distribution head 5-1, a valve 5-2, a gas distribution chamber 5-3 and a gas delivery head 5-4; the gas delivery head 5-4 is uniformly distributed with several gas distribution heads 5-1, each gas distribution head 5-1 is provided with a valve 5-2, and the inlet of the gas delivery head 5-4 is communicated with the gas source; each gas distribution head 5-1 is communicated with the corresponding duck-shaped jet head 4, and the diagnostic alarm module 7 selectively controls the valve 5-2 to open, so as to make the duck-shaped jet head 4 near the point work.

[0057] In the embodiment, the gas control valve 5 has a radius of 40 mm and a height of 40 mm, is connected with the several duck-shaped jet heads 4 and the gas tank through the air pipe respectively, the gas distribution chamber 5-3 is in the shape of a circular cake, the several gas distribution heads 5-1 are distributed around the side of the gas distribution chamber 5-3, the airflow in the gas distribution chamber 5-3 can basically keep consistent under the pressure field of the gas distribution head 5-1, the valve 5-2 is connected with the gas distribution chamber 5-3 through the bolt, and the airflow in the gas distribution chamber 5-3 is guided to the required gas distribution head 5-1 through the opening and closing of the valve 5-2, so as to realize the function of the gas control valve 5 controlling the airflow direction.

[0058] As shown in Figure 4 , Figure 5 , Figure 6 and Figure 7 , the signal acquisition module 6 includes an infrared sensor 6-1 and a strain gauge 6-2, the several infrared sensors 6-1 are distributed on the side of the chute slide rail 3, and are used for detecting whether there is material in the chute slide rail 3; the several strain gauges 6-2 are distributed on the bottom of the chute slide rail 3, and are used for detecting the deformation of the bottom of the chute slide rail 3; the sensing head of the infrared sensor 6-1 extends into the internal space of the slide rail 3-2, the infrared sensor 6-1 is in a linear array, is installed above the duck-shaped jet head 4, and is used for detecting whether there is material at each point of the slide rail 3-2; the several strain gauges 6-2 are installed on the bottom of the chute 3-1, and when the material falls into the chute 3-1, the displacement of the pulley changes, so that the strain gauge 6-2 deforms. The infrared sensor 6-1 is in a linear array, has a spacing of 40 mm, is installed 5 mm above the duck-shaped jet head 4, detects whether there is material at each point of the slide rail 3-2, and transmits to the diagnostic alarm module 7; the several strain gauges 6-2 are installed on the bottom of the chute 3-1, have a spacing of 20 mm, and when the material falls into the chute 3-1, the displacement of the pulley changes, so that the abnormal strain signal of the chute 3-1 is detected and transmitted to the diagnostic alarm module 7, so as to realize the acquisition of the multiple detection signals.

[0059] The diagnostic alarm module 7 comprises a signal processing diagnostic unit and a display alarm unit; the signal processing diagnostic unit learns and diagnoses the signals detected by the infrared sensors 6-1 and the strain gauges 6-2, establishes a deep learning diagnostic model, and the deep learning diagnostic model outputs the point information where the material is located; the display alarm unit alarms according to the point where the material is located output by the deep learning diagnostic model, and the display alarm unit controls the duck-shaped air jet head 4 corresponding to the point to be in communication with the air source for the point removal of the material.

[0060] As shown in Figure 8 , Figure 9 and Figure 10 , the detection method of the airflow removal device for materials in the cleaning screen chute of the application comprises the following steps:

[0061] Step S1: A plurality of infrared sensors 6-1 are distributed on the side of the chute slide rail 3 to detect whether there is material in the chute slide rail 3; a plurality of strain gauges 6-2 are distributed on the bottom of the chute slide rail 3 to detect the deformation of the bottom of the chute slide rail 3; the plurality of infrared sensors 6-1 and strain gauges 6-2 constitute 14 point signal input diagnostic alarm modules 7; a voltage signal is obtained through the 14 points at a sampling rate of 1 kHz, and the voltage signal is segmented at intervals of 1s to obtain an original data matrix Z ∈ [X1, X2, …, X 14 ] of each second, wherein X1, X2, …, X 14 represent the sampling voltage signal data within 1s at different points. Before data set training, all the original data matrices Z are normalized to make the matrix elements in [0, 1] to obtain a normalized data matrix M.

[0062] Step S2: The normalized data matrix M is subjected to noise adding processing to add Gaussian noise distribution q D (μ0, σ), wherein μ0 is the distribution center and σ is the variance of the noise, equal to 1, to obtain a noisy data matrix

[0063]

[0064] Among them: represents the normalized voltage signal at the jth point of the ith sampling point after adding noise, m represents the total number of sampling points, and n represents the point.

[0065] Step S3: According to the formula , the input noisy data is encoded to obtain high-level features h, wherein f θ is a function relationship used in the encoding process, s is a Sigmoid activation function, W and b are the weight matrix and bias vector of the encoder respectively.

[0066] Step S4: According to the formula Decode the high-level features h of the noisy data to obtain the reconstructed output of the original data. Where g a Let s represent the functional relationship used in the decoding process, where s is the sigmoid activation function, and C′ and d′ are the weight matrix and bias vector of the encoder, respectively. Typically, C′ = W. T .

[0067] Step S5: Through the objective function For noisy data matrices and reconstructing the data matrix To minimize the error between them, we can obtain the optimal parameter r, r = {W, C′, b, d′}.

[0068] Step S6: Reconstruct the data matrix As the filtered signal, further extraction The peak value, mean, and standard deviation are used as input features. After concatenating the features, the feature matrix is ​​obtained. Among them, fz n,k jz represents the k-th peak feature at the n-th point. n,k bzc represents the k-th mean feature of the n-th point. n,k Let C represent the k-th standard deviation feature of the n-th point. Matrix C is divided into a test set and a training set according to a 20% and 80% ratio, respectively. This training set serves as the input layer of a backpropagation neural network (BPNN) with 3 neurons. The hidden layers of the BPNN have 14 neurons, and the activation function is ReLU. The output layer of the BPNN has 15 neurons according to pre-defined point state labels, namely [0000; 1000; 0100; 0010; 0001; 1100; 1010; 1001; 0110; 0011; 1110; 1011; 1 [101; 0111; 1111], the labels are marked as no material present, material present at point 1, material present at point 2, material present at point 3, material present at point 4, material present at point 5, material present at point 6, material present at point 7, material present at point 8, material present at point 9, material present at point 10, material present at point 11, material present at point 12, material present at point 13, material present at point 14, and material present at point 15. BPNN training is performed, with accuracy as the indicator, and the training is repeated multiple times to obtain a material detection model for slide rails and chutes using DAE filtering and BPNN diagnosis.

[0069] like Figure 11As shown, the display alarm system of the display alarm unit can realize real-time collection of voltage signals of 14 points and diagnosis and alarm display of voltage signals of each second, and can display the real-time state of points 1-14. When any one of the display lamps in the combination of points 1 and 2, 3 and 4, 5 and 6, and 7 and 8 is always on in the point where the chute has material, the duck-shaped air jet nozzle between the two points blows high-speed airflow to the point until the display lamp is not on. When any one of the display lamps of points 9, 10, 11, 12, 13 and 14 is always on in the point where the chute has material, the duck-shaped air jet nozzle below the point blows high-speed airflow to the point. At the same time, the data can be saved in a fixed path, and the saved data can be normalized, added with noise, and subjected to other signal processing, DAE filtering, BPNN model continuous training and other operations. Through LabView 2018b, a display area of voltage peak value is established, and the voltage peak value changes of points 1-14 can be displayed in real time. Through LabView 2018b, a parameter setting area is established, and the sampling rate, sampling channel and the like can be set.

[0070] A detection system of a gas flow cleaning device for cleaning materials in a cleaning screen chute, comprising a storage medium; the storage medium stores a program written by the detection method of the gas flow cleaning device for cleaning materials in the cleaning screen chute. The storage medium includes a hard disk or a CD-ROM or an optical storage device or a magnetic storage device or a combination thereof. Those skilled in the art should understand that various features described herein can be implemented by a method, a data processing system or a computer program product. Therefore, these features can be expressed in the form of hardware, software or a combination of hardware and software. In addition, the above features can also be expressed in the form of a computer program product stored on one or more computer readable storage media, which contains computer readable program code segments or instructions stored in the storage medium. Any computer readable storage medium can be used, including hard disks, CD-ROMs, optical storage devices, magnetic storage devices and / or combinations thereof.

[0071] It should be understood that although the present specification is described in terms of various embodiments, not every embodiment contains only one independent technical solution, and the description of the specification is only for the sake of clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be combined to form other embodiments that those skilled in the art can understand.

[0072] The above series of detailed descriptions are only specific descriptions of feasible embodiments of the present application, and are not intended to limit the protection scope of the present application. Any equivalent embodiments or changes made without departing from the spirit of the present application should be included in the protection scope of the present application.

Claims

1. A pneumatic cleaning device for cleaning material in a cleaning sieve chute, characterized in that The machine frame (1), the cleaning screen (2), the chute slide rail (3), the canard jet head (4), the gas control valve (5), the gas source, the signal acquisition module (6) and the diagnostic alarm module (7) are included. The chute slide rail (3) is arranged on the both sides of the machine frame (1), and the cleaning screen (2) is movably arranged in the chute slide rail (3); a plurality of canard jet heads (4) are arranged on the side wall of the machine frame (1) above the chute slide rail (3), and the plurality of canard jet heads (4) are communicated with the gas source through the gas control valve (5) and used for cleaning the materials in the chute slide rail (3). The signal acquisition module (6) includes infrared sensors (6-1) and strain gauges (6-2), a plurality of infrared sensors (6-1) are arranged on the side of the chute slide rail (3) and used for detecting whether there is material in the chute slide rail (3), and a plurality of strain gauges (6-2) are arranged on the bottom of the chute slide rail (3) and used for detecting the deformation of the bottom of the chute slide rail (3); the canard jet head (4) is arranged near the detection point of the infrared sensor (6-1) and the strain gauge (6-2); the diagnostic alarm module (7) acquires the signals detected by the infrared sensor (6-1) and the strain gauge (6-2), analyzes the acquired signals, determines the point where the material in the chute slide rail (3) is located, and controls the gas control valve (5) to make the canard jet head (4) near the point work and clean the materials in the chute slide rail (3).

2. The pneumatic cleaning apparatus for material in an inner trough of a cleaning screen according to claim 1, characterized in that, The canard jet head (4) includes an air inlet (4-1), a rectifier chamber (4-2), a grid strip (4-3) and a canard nozzle (4-4); the rectifier chamber (4-2) is provided with the air inlet (4-1) which is gradually expanded along the flow direction on one side, the air inlet (4-1) is communicated with the gas control valve (5), the rectifier chamber (4-2) is provided with a plurality of grid strips (4-3) which are distributed at intervals, and the rectifier chamber (4-2) is provided with the canard nozzle (4-4) on the other side, and the canard nozzle (4-4) is gradually tapered along the flow direction.

3. The pneumatic cleaning apparatus for cleaning material in an inner chute of a cleaning screen according to claim 2, characterized in that, One end of the canard nozzle (4-4) is embedded in the internal space of the chute slide rail (3), and the distance from the bottom of the chute slide rail (3) is not more than 5mm.

4. The pneumatic cleaning apparatus for material in an inner trough of a cleaning screen according to claim 1, characterized in that, The gas control valve (5) includes a gas distribution head (5-1), a valve (5-2), a gas distribution chamber (5-3) and a gas delivery head (5-4); a plurality of gas distribution heads (5-1) are uniformly distributed on the gas delivery head (5-4), each gas distribution head (5-1) is provided with a valve (5-2), and the inlet of the gas delivery head (5-4) is communicated with the gas source; each gas distribution head (5-1) is communicated with the corresponding canard jet head (4), and the diagnostic alarm module (7) selectively controls the valve (5-2) to open, so that the canard jet head (4) near the point works.

5. The pneumatic cleaning apparatus for cleaning material in an inner chute of a cleaning screen according to claim 1, characterized in that, The sensing head of the infrared sensor (6-1) extends into the internal space of the slide rail (3-2), the infrared sensor (6-1) is in a linear array, is installed above the canard jet (4), and is used to detect whether the material exists at each point of the slide rail (3-2); a plurality of strain gauges (6-2) are installed at the bottom of the chute (3-1), and when the material falls into the chute (3-1), the displacement of the pulley changes, so that the strain gauge (6-2) deforms.

6. The pneumatic cleaning apparatus for cleaning material in an inner chute of a cleaning screen according to claim 1, characterized in that, The diagnosis alarm module (7) comprises a signal processing diagnosis unit and a display alarm unit; the signal processing diagnosis unit learns and diagnoses the signals collected by the infrared sensor (6-1) and the strain gauge (6-2), establishes a deep learning diagnosis model, and the deep learning diagnosis model outputs the point information where the material is located; the display alarm unit alarms according to the point where the material is located output by the deep learning diagnosis model, and the display alarm unit controls the canard jet (4) corresponding to the point to be in communication with the gas source for the point removal of the material.

7. A method of detecting the airflow cleaning device of the material in the sifting screen chute according to any one of claims 1-6, characterized in that, The method comprises the following steps: A plurality of infrared sensors (6-1) are distributed on the side of the chute slide rail (3) to detect whether there is material on the chute slide rail (3); a plurality of strain gauges (6-2) are distributed on the bottom of the chute slide rail (3) to detect the deformation of the bottom of the chute slide rail (3); the plurality of infrared sensors (6-1) and the plurality of strain gauges (6-2) constitute n point signal input diagnostic alarm modules (7), wherein the original data matrix Z of the n point signals is in [X1, X2,..., Xn], wherein Xj represents the signal of the jth point, j∈[1, n]; j n j ;​​ All the original data matrices Z are normalized to make the matrix elements between [0, 1], and a normalized data matrix M is obtained; The normalized data matrix M is subjected to noise addition, and Gaussian noise distribution q is added D (μ0, σ), where μ0 is the distribution center and σ is the variance of the noise, to obtain the data matrix with noise wherein: Vij represents the normalized voltage signal at the jth point under the ith sampling point after adding noise, m represents the total number of sampling points, and n represents the point. According to the formula The input noisy data Encoding, get high-level features h, f θ The function relationship used in the encoding process, where s is the Sigmoid activation function, W is the weight matrix of the encoder, and b is the bias vector of the encoder. According to the formula The high-level features h of the noisy data are decoded to obtain the reconstructed output of the original data Where g a is a function relationship used in the decoding process, s is a Sigmoid activation function, C' is a weight matrix of the encoder, and d' is a bias vector of the encoder; C' = W T ; minimizing the error between the target function a noisy data matrix and a reconstructed data matrix to obtain the optimal parameters r, r = {W, C', b, d'}; reconstructed data matrix The peak value, mean value and standard deviation of the filtered input signal are extracted as input features The feature matrix is obtained after splicing the features Wherein, fz n，k represents the kth peak value feature of the nth point, jz n，k represents the kth mean value feature of the nth point, and bzc n，k represents the kth standard deviation feature of the nth point. The feature matrix C is divided into a test set and a training set and used as an input layer of the BP neural network, 3 neurons are set, the number of neurons in the hidden layer is (2 / 3)*(n+4), the activation function is ReLU, the output layer of the BP neural network is set to n+1 according to the point state label, the point state label includes no material and the jth point where the material exists, the BP neural network is performed, and a slide rail and chute material deep learning diagnosis model is obtained; The point where the material is located is determined by using the slide rail and chute material deep learning diagnosis model, and the canard jet (4) corresponding to the point is controlled to be in communication with the gas source for the point removal of the material.

8. A detection system for an air flow cleaning device for cleaning material in an inner material channel of a cleaning sieve, characterized in that The storage medium comprises a program written by the detection method of the air flow removal device for the material in the cleaning screen chute according to claim 7.

Images