Material loading control method

Abstract

The invention relates to a material loading control method, which comprises the following steps of: acquiring volume data of loading materials in a time period, calculating volume according to the acquired volume data, and calculating a mapping relation by utilizing a fitting algorithm; installing a camera device, capturing a material flow image in real time, and calculating a fluidity parameter in a flow velocity vector field by using an algorithm; substituting the calculated fluidity parameters into the mapping relation, correcting the mapping relation, and establishing a corrected volume mapping relation model; the prediction result is compared with the volume of the material monitored in real time, the control parameters are dynamically adjusted according to the comparison result, and the deviation in the loading process can be found and corrected in time by monitoring the volume and fluidity parameters of the material in real time and dynamically adjusting the control parameters according to the corrected volume mapping relation model; and the control precision of material loading is improved.

Description

Technical Field

[0001] The present invention relates to the field of automation technology, and in particular to a material loading control method. Background Art

[0002] In modern industrial production, material loading is a key link, which directly affects the efficiency and accuracy of the production line. With the continuous development of automation and intelligent technology, precise control of the material loading process has become an urgent need in the industry. Traditional material loading methods often rely on manual operation and experience judgment, and have problems such as low efficiency and large errors. Therefore, it is particularly important to develop an efficient and accurate material loading control method. The core of the material loading control method lies in the real-time monitoring and accurate prediction of material volume and flow state. In the past, due to technical limitations, the measurement of material volume was often estimated or intermittently measured. This method not only has limited accuracy, but also cannot reflect the changes in material volume in real time. With the advancement of sensor technology and data processing algorithms, continuous and real-time material volume data collection can now be achieved. By installing suitable sensors, such as laser scanners, stereo vision systems, etc., the volume changes of materials during loading can be accurately measured, providing an accurate data basis for subsequent control.

[0003] For example, Chinese patent application No. 202411237495.X discloses a material loading control method and device, which is as follows: determine the loading material weight and loading material volume corresponding to multiple sampling times in a first time period; determine the total time length of the second time period based on the loading material weight corresponding to multiple sampling times; determine the first mapping relationship between the target time and the loading material volume based on the loading material volume corresponding to multiple sampling times; determine the target loading material volume corresponding to the target sampling time based on the first mapping relationship and the total time length; determine the parameter value of the control parameter in the second time period based on the target loading material volume. This method uses the target sampling time as the time point when the material warehouse is full, and can accurately measure the material loading volume under the full warehouse condition, so as to adjust the parameter value of the control parameter in the second time period according to the target loading material volume, so as to achieve accurate material loading under the full warehouse condition, and effectively improve the material loading accuracy.

[0004] In existing patent documents, relying solely on volume data is not enough to achieve accurate material loading control. The flow of materials will be affected by many factors, such as the speed of the conveyor belt, the physical properties of the materials (such as density, shape, friction coefficient, etc.), and possible collisions and adhesions. These factors will lead to the instability of the material flow state, thereby affecting the accuracy and efficiency of loading. Therefore, it is necessary to monitor and identify the flow state of materials in real time. Summary of the invention

[0005] In view of the technical problems existing in the prior art, the present invention provides a material loading control method, which can timely discover and correct deviations in the loading process and improve the control accuracy of material loading by real-time monitoring of the volume and fluidity parameters of the material and dynamically adjusting the control parameters according to a revised volume mapping relationship model.

[0006] The technical solution of the present invention to solve the above technical problems is as follows: A material loading control method, comprising: S101, collecting volume data of loaded materials within a time period, calculating the volume according to the collected volume data, and using a fitting algorithm to calculate a mapping relationship between the target time and the volume of loaded materials; S102, installing a camera device, capturing a material flow image in real time using the camera device, identifying a state change of the material during the flow process, and calculating a fluidity parameter in a flow velocity vector field using an algorithm; S103, substituting the calculated fluidity parameter into the mapping relationship in step S101, correcting the mapping relationship, and establishing a corrected volume mapping relationship model; S104, according to the prediction result of the modified volume mapping relationship model, the prediction result is compared with the volume of the material monitored in real time, and the control parameters are dynamically adjusted according to the comparison result. Preferably, the mapping relationship between the target time and the volume of loaded materials is: V=f(t;θ)=at+b, wherein V is the volume of loaded materials, indicating the volume occupied by the loaded materials at a certain time t; t is the target time, indicating the time interval from the start of loading to a certain sampling time, used to indicate the time progress in the loading process; a is the slope parameter, indicating the rate at which the volume of loaded materials changes with time, and determines how fast the volume increases with time. If a is positive, it indicates that the volume increases with time; if a is negative, it indicates that the volume decreases with time; b is the intercept parameter, indicating the initial volume of the loaded materials at time t=0; θ is a parameter set, which includes all parameters used to describe the mapping relationship, θ={a,b}.

[0007] Preferably, the fluidity parameters include the average flow rate and the standard deviation of the flow rate. The average flow rate is calculated using the formula: ,in, represents the average flow velocity, represents the modulus of the i-th velocity vector, that is, the size of the vector, and N represents the total number of vectors in the velocity vector field.

[0008] Preferably, the standard deviation of the flow rate is calculated using the formula: ,in, represents the standard deviation of flow velocity, represents the difference between the modulus of the i-th velocity vector and the average velocity, represents the modulus of the i-th velocity vector, that is, the size of the vector, and N represents the total number of vectors in the velocity vector field.

[0009] Preferably, the fluidity parameter is substituted into the mapping relationship between the target time and the loading material volume, and the result is: a′=a+ × + × , where a′ is the slope parameter corrected according to the liquidity parameter, and is the correction coefficient, which indicates the influence of the average velocity and the standard deviation of the velocity distribution on the parameter a. The correction coefficient is determined by experiments or data fitting. Substituting the corrected slope parameter into the mapping relationship, the formula is obtained: V=f(t; ′)=a′t+b, where ′ represents the modified parameter set, ′={a′,b}.

[0010] Preferably, the specific steps for identifying the state change of the material during the flow process are: S201, setting a plurality of static recognition units on a conveyor belt for conveying materials according to the width and length of the conveyor belt and the speed of material flow; S202, identifying the changing state of the material according to the set static identification unit to determine whether the material is adhered. If the material particles are connected together and the area of ​​the connected part exceeds 15% of the total area of ​​the particles, the material is in an adhered state. Otherwise, the material is not adhered.

[0011] Preferably, the conveyor belt is divided into an entrance pre-inspection area, a dynamic balancing area and a loading positioning area, wherein the entrance pre-inspection area is located at the silo, the dynamic balancing area is located at the geometric center of the longest horizontal section of the conveyor belt, and the loading positioning area is located at the material unloading location.

[0012] Preferably, the method for identifying the change state of a material further comprises: S301, collecting material data, and building a material movement trajectory prediction model based on the collected material data; S302, according to the image recognized by the static recognition unit, when there is an undetected material contour in the recognition result, the image is input into the material motion trajectory prediction model to predict the undetected material contour; S303, integrating the material movement trajectory prediction model with the loading control system.

[0013] Preferably, the method for constructing a material movement trajectory prediction model further includes: S401, setting a dynamic recognition unit on the conveyor belt according to the sensor data of the material; S402, establishing a physical model based on the collected sensor data, identifying information of historical collision events based on the dynamic recognition unit and the physical model, establishing a collision event library, and extracting the law of collision events based on the established collision event library; S403, classifying the bounce patterns according to the data in the collision event library, and establishing a material flow prediction system according to the rules of the bounce patterns and collision events.

[0014] Preferably, the bouncing mode includes linear bouncing and rotational bouncing, wherein the linear bouncing is that the material bounces along a straight line after the collision, and the rotational bouncing is that the material rotates simultaneously after the collision.

[0015] The beneficial effects of the present invention are as follows: by real-time monitoring of the volume and fluidity parameters of the material, dynamically adjusting the control parameters according to the revised volume mapping relationship model, the deviation in the loading process can be discovered and corrected in time, the control accuracy of material loading can be improved, and the loading failure caused by the change of material flow characteristics can be reduced. The high-speed camera device and fluidity parameters are introduced, and the flow characteristics of the material are monitored in real time, and the control parameters are dynamically adjusted, which can better adapt to the change of the material flow characteristics and improve the stability and accuracy of the loading process; By setting up several static identification units on the conveyor belt, the state changes of materials in high-speed flow can be accurately identified. The loading control system dynamically adjusts the control parameters according to the changes in the material state, reduces the residue of sticky materials on the conveyor belt, and improves the loading accuracy and stability. The improvement of loading accuracy reduces the error of loading quantity, improves the accuracy of loading, reduces the material waste and cost increase caused by errors, and improves the loading stability. Reduce fluctuations and failures in the loading process, improve the reliability and safety of loading, and reduce downtime and maintenance costs caused by failures; A material motion trajectory prediction model was established and verified, the complete contour information of the obscured material was inferred, and the control parameters of the loading control system were optimized. The material motion trajectory prediction model makes the identification more accurate, improves the adaptability of the loading system, reduces the residue of sticky materials on the conveyor belt, and improves the loading accuracy and stability; The dynamic recognition unit can accurately identify the collision and bounce phenomena of materials, restore the dynamic process at the moment of collision, extract rules from historical collision records, provide early warning and prediction support for material flow under similar conditions, and extract valuable collision rules from historical data, providing a scientific basis for early warning and prediction of material flow. The enhanced function of the dynamic recognition unit can more accurately identify the collision and bounce phenomena of materials, improve the comprehensiveness and accuracy of material status identification, and the collision event library further improves the efficiency and stability of material flow. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a flow chart of a material loading control method of the present invention; Figure 2 A schematic diagram of a process for identifying state changes of materials during flow in the present invention; Figure 3 A schematic diagram of a process for identifying a change state of a material according to the present invention; Figure 4 A schematic diagram of the process of constructing a material motion trajectory prediction model according to the present invention. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0018] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, "plurality" means two or more, unless otherwise clearly and specifically defined.

[0019] In the description of the present application, the term "for example" is used to mean "used as an example, illustration or description". Any embodiment described as "for example" in the present application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any technician in the field to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes will not be elaborated in detail to avoid unnecessary details to obscure the description of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in the present application.

[0020] Embodiment 1: Figure 1 1 is a flow chart of a material loading control method according to the first embodiment of the present invention, comprising the following steps: S101, collecting volume data of loaded materials within a time period, calculating the volume according to the collected volume data, and using a fitting algorithm to calculate a mapping relationship between the target time and the volume of loaded materials; Specifically, the time period is the time period in which the volume of the loaded materials changes from a first value to a second value, and a preset laser radar is used to collect point cloud data of the loaded materials. The preset laser radar is installed in the silo and is used to obtain the three-dimensional shape information of the materials. The laser radar forms a series of three-dimensional coordinate points by emitting lasers and receiving reflected signals. The three-dimensional coordinate points constitute the point cloud data of the materials. The point cloud data includes the shape, spatial position, size and other geometric features of the material surface. The bottom area of ​​the material truck is obtained by measurement. Based on the point cloud data collected by the preset laser radar, the Delaunay triangulation 3D modeling technology can be used to construct a three-dimensional model of the material. The height information of the material is extracted from the constructed three-dimensional model. The height is the real-time vertical size of the material in the loading space. Based on the real-time height of the material, the bottom area of ​​the material and the formed shape of the material, a geometric calculation method is used to calculate the volume of the loaded materials. If the material If the shape of the material is approximately regular (such as a cuboid, cylinder, etc.), the corresponding volume formula can be used directly for calculation; if the shape of the material is irregular, the space of the silo can be divided into multiple small volume units (such as small cuboids or triangular prisms, etc.), and then the volume of each small volume unit is calculated. Finally, the volumes of all small volume units are added up to get the total volume. The fitting algorithm GBRT is used to fit the data of the loading material volume corresponding to multiple sampling times within the time period to obtain the mapping relationship between the target time and the loading material volume: V=f(t;θ)=at+b, where V is the loading material volume, which indicates the volume occupied by the loading material at a certain time t; t is the target time, which indicates the time interval from the start of loading to a certain sampling time, and is used to indicate the time progress of the loading process; a is the slope parameter, which indicates the rate at which the loading material volume changes over time and determines how fast the volume increases over time. If a is positive, it means that the volume increases with time; if a is negative (unlikely in actual loading), it means that the volume decreases with time; b is the intercept parameter, which represents the initial volume of the loaded material at time t=0. In practical applications, it represents the volume of the empty car at the beginning of loading (if the net increase in volume is considered) or a certain baseline volume value; θ (in general form) is a parameter set that includes all parameters used to describe the mapping relationship, θ={a,b}.

[0021] S102, installing a camera device, capturing a material flow image in real time according to the camera device, and using an algorithm to calculate a fluidity parameter in a flow velocity vector field; Furthermore, according to the environmental conditions of the loading site, the HS1000 camera device is selected and debugged, including adjusting the parameters such as focal length, aperture and exposure time to ensure that the camera can capture clear images and that the color and brightness of the images meet the actual requirements. The image captured by the camera is observed to be stable and to see if there is any jitter or blur. Calibration is performed to ensure that the captured image is consistent with the actual material flow. Start the high-speed camera device to capture the image of material flow in real time, convert the real-time captured image into a grayscale image, use a Gaussian filter to remove noise in the image, improve image quality, and enhance the contrast of the image to more clearly display the details of material flow. Use the optical flow algorithm Lucas-Kanade to calculate the motion vector of pixels between adjacent frames. The motion vector represents the flow velocity of the material on the image plane. The calculated motion vectors are combined into a flow velocity vector field to represent the flow of the material in the entire image area. The fluidity parameters include the average flow velocity and standard deviation. The average value of all vectors in the flow velocity vector field is calculated. The formula is: ,in, represents the average flow velocity, represents the modulus of the i-th velocity vector, that is, the magnitude of the vector, regardless of direction, and N represents the total number of vectors in the velocity vector field. In image processing, this usually corresponds to the number of all flow vectors detected in the image. The standard deviation of the modulus of the vectors in the velocity vector field is calculated as: ,in, represents the standard deviation of flow velocity, Represents the difference between the modulus of the i-th flow velocity vector and the average flow velocity.

[0022] S103, substituting the calculated fluidity parameter into the mapping relationship in step S101, correcting the mapping relationship, and establishing a corrected volume mapping relationship model; Specifically, the flow parameters include the average flow rate and the standard deviation of the flow rate distribution. Substituting the flow parameters into the mapping relationship between the target time and the loading material volume, we get: a′=a+ × + × , where a′ is the slope parameter corrected according to the liquidity parameter, and is the correction coefficient, which indicates the influence of the average velocity and the standard deviation of the velocity distribution on the parameter a. The correction coefficient is determined by experiments or data fitting. Substituting the corrected slope parameter into the mapping relationship, the formula is obtained: V=f(t; ′)=a′t+b, where ′ represents the modified parameter set, where ′={a′,b}; According to the corrected mapping relationship, a corrected volume mapping relationship model is established using a machine learning algorithm decision tree. During the modeling process, the influence of the fluidity parameters on the volume mapping relationship is considered and incorporated into the structure and parameters of the model. According to the selected modeling method, a preliminary structure of the model is constructed, including the fluidity parameters and other influencing factors as the input variables of the model, and the volume mapping relationship is determined as the output variable of the model. The model is fit using a training data set, and the training data is input into the model. The parameters of the model are adjusted through an optimization algorithm to minimize the prediction error. Methods such as cross-validation are used to evaluate the generalization ability of the model and prevent overfitting. After fitting the training data, the preliminary parameters and form of the model are obtained, and the model can predict the output variables based on the input variables.

[0023] S104, comparing the prediction result of the modified volume mapping relationship model with the volume of the material monitored in real time, and dynamically adjusting the control parameters according to the comparison result; Furthermore, a high-speed camera is used to collect the volume and fluidity parameters of the material in real time, and the collected volume and fluidity parameters are input into the revised volume mapping relationship model. The model calculates the predicted volume of the material based on the input data, and compares the predicted result of the model with the actual loading situation to monitor whether there is a deviation. By adjusting the material unloading speed and the telescopic length of the unloading chute, if the deviation is a positive value, the material unloading speed is slowed down to reduce the amount of material entering the loader per unit time, thereby reducing the deviation of the loading volume, and at the same time shortening the telescopic length of the unloading chute to reduce material accumulation, thereby reducing the deviation of the loading volume; if the deviation is a negative value, the material unloading speed is accelerated to increase the amount of material entering the loader per unit time, thereby increasing the loading volume and reducing the deviation, and at the same time extending the telescopic length of the unloading chute to increase material accumulation, thereby reducing the deviation of the loading volume, and adjusting the control parameters in real time to ensure a smooth transition during the adjustment process to avoid impact or interference on the loading process. After adjusting the control parameters, continue to monitor the loading situation of the material in real time.

[0024] The technical solutions in the above-mentioned embodiments of the present application have at least the following technical effects or advantages: improving the accuracy of volume prediction, improving the accuracy of material loading, and reducing loading failures. By real-time monitoring of the volume and fluidity parameters of the material, and dynamically adjusting the control parameters (such as the telescopic length of the unloading chute, the unloading speed, etc.) according to the revised volume mapping relationship model, deviations in the loading process can be discovered and corrected in time, the control accuracy of material loading can be improved, and loading failures caused by changes in material flow characteristics can be reduced. A high-speed camera device and fluidity parameters are introduced. By real-time monitoring of the flow characteristics of the material and dynamically adjusting the control parameters, the change of the material flow characteristics can be better adapted to the improvement of the stability and accuracy of the loading process.

[0025] Example 2: Based on the phenomenon in Example 1 that the material particles are broken or adhered to each other due to factors such as collision and friction during high-speed flow, this example identifies the state changes of the material during high-speed flow and dynamically adjusts the control parameters according to the state changes of the material.

[0026] like Figure 2 As shown in the figure, the specific steps for identifying the state changes of materials during the flow process are: S201, setting a plurality of static recognition units on a conveyor belt for conveying materials according to the width and length of the conveyor belt and the speed of material flow; Specifically, the static identification unit is a system arranged on the conveyor belt, which is used to statically (i.e., not move with the material) identify the state changes of the material, such as crushing or adhesion, when the material flows at high speed, and monitor the state of the material in the conveying process in real time. The laser rangefinder obtains the precise size of the conveyor belt, measures the total length, width and slope of the conveyor belt, and sets static identification units in the entrance pre-inspection area, dynamic balancing area and loading positioning area of ​​the conveyor belt. For the entrance pre-inspection area, a static identification unit is set in the entrance pre-inspection area. The entrance pre-inspection area is near the silo, and the entrance pre-inspection area covers the full width of the conveyor belt and extends a certain distance on both sides. The extension on both sides is to prevent the material from overflowing from the edge. For the dynamic The balancing area and the dynamic balancing area are provided with two static identification units. The dynamic balancing area is located at the geometric center of the longest horizontal section of the conveyor belt and avoids vibration sources (such as reducer and sorting mechanism). For the loading positioning area, a static identification unit is provided in the loading positioning area. The loading positioning area is near the material unloading point. The number of static identification units is determined according to the length of the conveyor belt and the material flow path to monitor the state changes of the materials during the loading process in real time. The distance between the static identification units on the conveyor belt is adjusted according to the frequency of material state changes to ensure that the materials are always in a monitored state during the flow process. The planned installation position of the static identification units is marked on the conveyor belt using a marking tool and spray paint.

[0027] Specific example: the conveyor belt width is 1.2 meters, the conveyor belt length is 30 meters, the material flow speed is 2 meters per second, the conveyor belt slope is an overall flat slope, but contains several horizontal sections and slightly inclined sections, the number of static identification units in the entrance pre-inspection area is 1, the entrance pre-inspection area is located at the entrance end of the conveyor belt, at least 5 meters away from the silo, the static identification unit in the entrance pre-inspection area is set at the center of the entrance end, and its monitoring range covers the full width of the conveyor belt horizontally plus 10 cm on both sides, that is, the total width is 1.4 meters, to prevent material from overflowing from the edge without being detected; the number of static identification units in the dynamic balancing area is 2, the dynamic balancing area is located at the geometric center of the longest horizontal section of the conveyor belt, and avoids vibration sources such as reduction The two static recognition units of the speed machine and sorting mechanism and the dynamic balance area are respectively arranged on the front and rear sides of the geometric center of the horizontal section, about 5 meters apart, ensuring the continuous monitoring of the dynamic balance state of the material during the flow process. The coordinated work of the two units can improve the accuracy and reliability of the monitoring; the number of static recognition units in the loading positioning area is 1, and the loading positioning area is 1.2 meters away from the material unloading point. The static recognition unit of the loading positioning area is precisely set 1.2 meters in front of the unloading point to ensure that the final state of the material before unloading can be accurately captured. The purpose is to accurately monitor the state of the material when it is about to be unloaded, so as to ensure the smooth progress of the loading process and avoid material overflow or uneven loading. The distance between the entrance pre-inspection area and the dynamic balance area is set to 10-15 meters according to the conveyor belt layout and monitoring requirements to cover the process from the material entering the conveyor belt to reaching the dynamic balance area. The distance within the dynamic balance area is 5 meters apart between the two static identification units to ensure continuous and intensive monitoring of the material status. The distance between the dynamic balance area and the loading positioning area is determined according to the remaining length of the conveyor belt and the material flow speed to ensure that the material is always under monitoring during the flow process.

[0028] S202, identifying the changing state of the material according to the set static identification unit to determine whether the material is adhered. If the material particles are connected together and the area of ​​the connected part exceeds 15% of the total area of ​​the particles, the material is in an adhered state. Otherwise, the material is not adhered. Furthermore, based on the physical properties of the material, an area ratio threshold of 15% is set. Experiments are conducted to verify whether the set recognition threshold is reasonable. Different types of materials and flow rates are used for testing, the recognition effect of the recognition unit is observed, and relevant data are recorded. Based on the test results, the recognition threshold is fine-tuned to ensure its accuracy in the actual production process. When two or more material particles are connected together and the area of ​​the connected part exceeds 15% of the total area of ​​the particles, the material is judged to be adhered. Otherwise, the material is not adhered. When the material flows through the static recognition unit, the static recognition unit captures the laser reflection signal of the material in real time. The signal contains characteristic information such as the surface shape, color, texture, etc. of the material. The static recognition unit compares the captured signal with the set recognition threshold to determine whether the material is in a state of adhesion. The static recognition unit is integrated with the loading control system. If the static recognition unit determines that the material is in a state of adhesion, it immediately sends a signal to the loading control system. This signal contains information such as the location and degree of the adhesion material. After receiving the signal, the loading control system will take corresponding measures according to the preset control logic to adjust the speed of the conveyor belt to change the flow state of the material; or change the loading position to avoid adhesion materials from entering the loading area.

[0029] The technical solutions in the above-mentioned embodiments of the present application have at least the following technical effects or advantages: by arranging a number of static identification units on the conveyor belt, the state changes of materials in high-speed flow can be accurately identified, and the loading control system dynamically adjusts the control parameters according to the changes in the material state, thereby reducing the residue of sticky materials on the conveyor belt and improving the loading accuracy and stability. The improvement in loading accuracy reduces the error in loading quantity, improves the accuracy of loading, reduces material waste and cost increase caused by errors, and improves loading stability. It reduces fluctuations and failures in the loading process, improves the reliability and safety of loading, and reduces downtime and maintenance costs caused by failures.

[0030] Example 3: Based on the fact that in Example 2, during the high-speed flow of materials, occlusion and overlap are easily generated between materials due to the high-speed flow, which affects the accuracy of material identification, this example establishes a material movement trajectory model to infer the complete contour information of the occluded material, thereby further improving the accuracy of material identification.

[0031] like Figure 3 As shown, the method for identifying the change state of the material also includes S301, collecting material data, and building a material movement trajectory prediction model based on the collected material data; Specifically, the material data includes physical property data and flow characteristic data. The physical property data includes the density, shape and friction coefficient of the material. The density data is used to understand the mass distribution of the material. The shape data reflects the force applied to the material during the flow process. The friction coefficient is an important parameter for the interaction between the material and the contact surface. The data is directly obtained through experimental measurement, and the collected physical property data is recorded. The fluidity data includes flow velocity and pressure distribution. The flow velocity data is used to reflect the speed and direction of material flow, and the pressure distribution reveals the mechanical effects applied to the material during the flow process. The flow velocity and pressure distribution are measured experimentally, and the collected physical property data and fluidity data are preprocessed. Data cleaning is to remove erroneous or abnormal data points, denoising is to smooth the data curve and reduce fluctuations, and normalization is to scale the data to a uniform range.

[0032] The neural network in the machine learning algorithm is used to construct a material motion trajectory prediction model. The neural network includes an input layer, a hidden layer, and an output layer. The input layer receives the preprocessed material physical properties and flow characteristics data, and the output layer outputs the predicted material motion trajectory. The number of hidden layers and the number of neurons in each layer need to be adjusted and optimized according to the specific problem. The ReLU activation function is selected to increase the nonlinear expression ability of the model. The preprocessed data is divided into a training set and a test set. The training set is used to train the model, and the test set is used to evaluate the prediction performance of the model. The neural network model is trained using the training set data. The model parameters are adjusted to minimize the loss function through optimization methods such as the back propagation algorithm and the gradient descent algorithm. During the model training process, special attention is paid to the influence of the occluded area, and a specific network structure is designed to deal with the problem of missing information in the occluded area. The trained neural network model is tested using the test set data, and the model's accuracy, error rate and other indicators are recorded to evaluate the prediction performance of the model. The model prediction results are compared with the actual data to evaluate the prediction accuracy of the model.

[0033] S302, according to the image recognized by the static recognition unit, when there is an undetected material contour in the recognition result, the image is input into the material motion trajectory prediction model to predict the undetected material contour; Furthermore, the image recognized by the static recognition unit is processed by an image enhancement algorithm, and the preprocessed image is processed by CLAHE to improve the brightness, contrast and clarity of the image. The edge detection algorithm Canny edge detection is used to enhance the edges in the image and improve the contour clarity of the image. The image recognized by the static recognition unit is analyzed to find out the part of the material contour that is missing due to occlusion. By comparing the difference between the complete material contour and the recognition result, the position and shape of the missing contour are determined. By using image processing and analysis technology, the position and range of the occluded area are determined by analyzing the brightness, color, texture and other features in the image, and the occluded area is marked. The material motion trajectory prediction model is used to infer the material contour of the occluded area in combination with the material motion trajectory and contour information before and after the occluded area. The material motion trajectory prediction model outputs the inferred material contour, and the inferred contour is verified. By comparing with the actual contour, the accuracy of the inference result is evaluated. If there is a large difference between the inference result and the actual contour, the material motion trajectory prediction model is corrected and adjusted.

[0034] S303, integrating the constructed material movement trajectory prediction model with the loading control system; Specifically, the material movement trajectory prediction model is integrated into the loading control system. The loading control system is connected to the material movement trajectory prediction model through an interface. The prediction results provided by the material movement trajectory prediction model are used to more accurately predict the movement trajectory of the material during the loading process. According to the prediction results, the relevant control parameters of the loading control system are adjusted, such as loading speed and position, to ensure that the material can be accurately loaded into the vehicle.

[0035] The technical solutions in the above-mentioned embodiments of the present application have at least the following technical effects or advantages: a material movement trajectory prediction model is established and verified, an effective image enhancement algorithm is adopted, the complete contour information of the obscured material is inferred, and the control parameters of the loading control system are optimized. The material movement trajectory prediction model enables higher accuracy in identification, provides more reliable input for the loading control system, can handle more types of materials and more complex flow conditions, improves the adaptability of the loading system, reduces the residue of sticky materials on the conveyor belt, and improves the loading accuracy and stability.

[0036] Example 4: Based on Example 2 and Example 3, when the material moves at high speed, the material collides and bounces, causing unpredictable changes in the material's movement trajectory, affecting the accuracy of the prediction model. This example identifies the material state through a dynamic recognition unit, and restores the dynamic process of the collision moment by counting the amount of data and data type in the collision event library, thereby reducing the impact of the collision on the identification of the material and improving the accuracy of material identification.

[0037] like Figure 4As shown, the method for constructing a material movement trajectory prediction model also includes: S401, setting a dynamic recognition unit on the conveyor belt according to the sensor data of the material; Furthermore, sensors are used to obtain real-time data when a collision event occurs. The sensor data of the material includes the speed and frequency of the material flow. The dynamic identification unit is an intelligent device system that integrates advanced sensor technology, data processing algorithms and real-time monitoring functions. It is specially designed to monitor and analyze the movement state of the material during the material transportation process. It collects the dynamic parameters of the material and uses advanced data processing algorithms to analyze these parameters in real time, thereby realizing accurate identification and judgment of the movement state of the material. By setting dynamic identification units in the high-collision area, the bounce monitoring belt and the trajectory correction area, for the high-collision area, the high-collision area is within the range before and after the connection point of the conveyor belt (horizontal to inclined section), and the dynamic identification unit set in the high-collision area The number of units is determined according to the width, length and monitoring requirements of the conveyor belt, but considering that this is a high-incidence area, relatively dense dynamic identification units are usually set up. For the bounce monitoring belt, the bounce monitoring belt is located in the blanking buffer section. The bounce monitoring belt is a high-incidence area for material bouncing behavior, and relatively dense dynamic identification units are set up. For the trajectory correction area, the trajectory correction area is the calibration section before loading. The trajectory correction area is mainly used to correct and verify the material trajectory. Usually, one dynamic identification unit is set. The adjacent dynamic identification units in the same area in the collision high-incidence area, the bounce monitoring belt and the trajectory correction area overlap and cover 1% of the length of the dynamic unit to ensure the continuity and accuracy of the monitoring. The setting of the dynamic unit is separated from the setting of the static unit by a distance to avoid signal interference.

[0038] A specific example is: the width of a conveyor belt is 1.2m, the length is 30m, the flow rate is 2m / s, the high-incidence area of ​​collision is in the range before and after the connection point of the conveyor belt (horizontal to inclined section), and this range is 2m before and after the connection point. Two dynamic identification units are set in the high-incidence area of ​​collision, the bounce monitoring belt is in the material falling buffer section, and the bounce monitoring belt is 3-5m away from the silo outlet. The bounce monitoring belt is a high-incidence area for material bouncing behavior, and two dynamic identification units are set in the bounce monitoring belt. The trajectory correction area is the calibration section before loading, and the trajectory correction area is 0.5-2m away from the unloading point. This area is mainly used to correct and verify the material trajectory, so a dynamic identification unit is set.

[0039] S402, establishing a physical model based on the collected sensor data, identifying information of historical collision events based on the dynamic recognition unit and the physical model, establishing a collision event library, and extracting the law of collision events based on the established collision event library; Specifically, a physical model for accurately restoring the dynamic process of the collision moment is established based on the collected sensor data and physical principles. The set dynamic recognition unit and physical model are used to identify and record the information of the collision event, which includes the time and location of the collision (i.e., the specific location on the conveyor belt), the type of materials involved, and the intensity of the collision. The collision event information collected by the dynamic recognition unit is summarized to establish a structured collision event library. The number of collision events occurring in different time periods (such as hours, days, weeks, and months) is counted, and the time distribution pattern of the collision events is analyzed. The number of collision events occurring at different locations (such as the connection points, turns, and drop-off points of the conveyor belt) is counted, the location distribution pattern of the collision events is analyzed, the number of collision events involving different material types is counted, and the relationship between different material types and collision events is analyzed. The pattern of collision events is as follows: during the morning and afternoon peak hours on weekdays, at specific locations such as the connection points, turns, and drop-off points of the conveyor belt, materials with irregular shapes, large size differences, or heavy weights are more likely to collide, and the collision intensity is greater.

[0040] S403, classifying the bounce patterns according to the data in the collision event library, and establishing a material flow prediction system according to the rules of the bounce patterns and collision events; Furthermore, based on the data in the collision event library, the bouncing phenomenon of the material after the collision is analyzed, and the linear bounce of the material along a straight trajectory after the collision and the rotational bounce of the material simultaneously rotating after the collision are obtained. A material flow prediction system is constructed according to the classified bouncing patterns and the law of collision events. When the system detects that the material flow state conforms to certain characteristics of the collision law, a warning signal is immediately issued. The system can predict possible collision events in the future based on the current material flow state and provide corresponding prediction support.

[0041] The technical solutions in the above-mentioned embodiments of the present application have at least the following technical effects or advantages: accurately identifying the collision and bounce phenomena of materials through the dynamic recognition unit, restoring the dynamic process at the moment of collision, extracting rules from historical collision records, and providing early warning and prediction support for material flow under similar conditions. The collision event library and the classification of bounce patterns provide rich data and model support for the control and optimization of material flow, and can extract valuable collision rules from historical data, providing a scientific basis for early warning and prediction of material flow. Through the enhanced function of the dynamic recognition unit, the collision and bounce phenomena of materials can be identified more accurately, which improves the comprehensiveness and accuracy of material state identification, and the collision event library further improves the efficiency and stability of material flow.

[0042] It should be noted that in the above embodiments, the description of each embodiment has its own emphasis, and for parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0043] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0044] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0045] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0046] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0047] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0048] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A material loading control method, characterized in that: include: S101, collecting volume data of loaded materials within a time period, calculating the volume according to the collected volume data, and using a fitting algorithm to calculate a mapping relationship between the target time and the volume of loaded materials; S102, installing a camera device, capturing a material flow image in real time using the camera device, identifying a state change of the material during the flow process, and calculating a fluidity parameter in a flow velocity vector field using an algorithm; S103, substituting the calculated fluidity parameter into the mapping relationship in step S101, correcting the mapping relationship, and establishing a corrected volume mapping relationship model; S104, according to the prediction result of the modified volume mapping relationship model, the prediction result is compared with the volume of the material monitored in real time, and the control parameters are dynamically adjusted according to the comparison result.

2. A material loading control method according to claim 1, characterized in that: The mapping relationship between the target time and the volume of loaded materials is: V=f(t;θ)=at+b, where V is the volume of loaded materials, indicating the volume occupied by the loaded materials at a certain time t; t is the target time, indicating the time interval from the start of loading to a certain sampling time, which is used to indicate the time progress of the loading process; a is the slope parameter, indicating the rate at which the volume of loaded materials changes with time, which determines how fast the volume increases with time. If a is positive, it means that the volume increases with time; if a is negative, it means that the volume decreases with time; b is the intercept parameter, indicating the initial volume of the loaded materials at time t=0; θ is a parameter set, which includes all parameters used to describe the mapping relationship, θ={a,b}.

3. A material loading control method according to claim 1, characterized in that: The fluidity parameters include the average flow rate and the standard deviation of the flow rate. The average flow rate is calculated using the formula: ,in, represents the average flow velocity, represents the modulus of the i-th velocity vector, that is, the size of the vector, and N represents the total number of vectors in the velocity vector field.

4. A material loading control method according to claim 3, characterized in that: Calculate the standard deviation of the flow rate using the formula: ,in, represents the standard deviation of flow velocity, represents the difference between the modulus of the i-th velocity vector and the average velocity, represents the modulus of the i-th velocity vector, that is, the size of the vector, and N represents the total number of vectors in the velocity vector field.

5. A material loading control method according to claim 1, characterized in that: Substituting the liquidity parameter into the mapping relationship between the target time and the loading material volume, we get: Where a′ is the slope parameter corrected according to the liquidity parameter, and is the correction coefficient, which indicates the influence of the average velocity and the standard deviation of the velocity distribution on the parameter a. The correction coefficient is determined by experiments or data fitting. Substituting the corrected slope parameter into the mapping relationship, the formula is obtained: ,in, represents the modified parameter set, .

6. A material loading control method according to claim 1, characterized in that: The specific steps to identify the state changes of materials during the flow process are: S201, setting a plurality of static recognition units on a conveyor belt for conveying materials according to the width and length of the conveyor belt and the speed of material flow; S202, identifying the changing state of the material according to the set static identification unit to determine whether the material is adhered. If the material particles are connected together and the area of ​​the connected part exceeds 15% of the total area of ​​the particles, the material is in an adhered state. Otherwise, the material is not adhered.

7. A material loading control method according to claim 6, characterized in that: The conveyor belt is divided into an entrance pre-inspection area, a dynamic balancing area and a loading positioning area. The entrance pre-inspection area is located at the silo, the dynamic balancing area is located at the geometric center of the longest horizontal section of the conveyor belt, and the loading positioning area is located at the material unloading location.

8. A material loading control method according to claim 6, characterized in that: Methods for identifying the changing state of materials also include: S301, collecting material data, and building a material movement trajectory prediction model based on the collected material data; S302, according to the image recognized by the static recognition unit, when there is an undetected material contour in the recognition result, the image is input into the material motion trajectory prediction model to predict the undetected material contour; S303, integrating the material movement trajectory prediction model with the loading control system.

9. A material loading control method according to claim 8, characterized in that: The method of constructing a material movement trajectory prediction model also includes: S401, setting a dynamic recognition unit on the conveyor belt according to the sensor data of the material; S402, establishing a physical model based on the collected sensor data, identifying information of historical collision events based on the dynamic recognition unit and the physical model, establishing a collision event library, and extracting the law of collision events based on the established collision event library; S403, classifying the bounce patterns according to the data in the collision event library, and establishing a material flow prediction system according to the rules of the bounce patterns and collision events.

10. A material loading control method according to claim 9, characterized in that: The bouncing mode includes straight-line bouncing and rotational bouncing. The straight-line bouncing is that the material bounces along a straight line after collision, and the rotational bouncing is that the material rotates simultaneously after collision.

Images