Device for hermetically conveying glass melting raw materials based on industrial vision

Abstract

The invention relates to the technical field of glass melting raw material conveying devices, in particular to a glass melting raw material closed conveying device based on industrial vision, which comprises a conveying pipeline, a centrifugal blower, a stirring mixer, a pulse bag filter and a control unit, a discharge port of the stirring mixer is converged and connected to an inlet of the centrifugal blower through a rotary valve and an air suction pipe, and an outlet of the centrifugal blower is connected to an inlet of the pulse bag filter through an air output pipe. According to the device for hermetically conveying the glass melting raw materials based on the industrial vision, the conveying pipeline is adopted for hermetically conveying, and the problems that in the prior art, belt conveying is large in size, dust raising cannot be avoided, the conveying angle is limited, the mechanical structure is complex, and the failure rate is high are solved.

Description

Technical Field

[0001] This invention relates to the field of glass melting raw material conveying devices, specifically to a device for the closed conveying of glass melting raw materials based on industrial vision. Background Technology

[0002] Glass melting raw materials typically include the following main components: Silica (SiO2): This is one of the main components of glass, constituting the majority of its composition and determining its basic properties and characteristics; Oxides (such as CaO, Na2O, K2O, etc.): Oxides are commonly used as fluxes in glass, helping to lower the melting point and improve its processability and transparency; Fluorides (such as NaF, CaF2, etc.): Fluorides can reduce the viscosity of glass and promote melting and homogeneity; Oxides (such as Al2O3, MgO, etc.): Oxides are commonly used as stabilizers and reinforcing agents in glass, helping to improve its heat resistance and mechanical properties; Other additives (such as pigments, metal oxides, etc.): To impart specific colors, optical properties, or special functions to glass, small amounts of other elements or compounds can be added. These raw materials are mixed according to a certain formula ratio, melted at high temperature to form glass, and then processed through steps such as forming and cooling to obtain the final glass product. Different types of glass products may require different raw material compositions and different formula ratios.

[0003] Chinese patent CN215209107U discloses a raw material batching and conveying system for ultra-white glass, including a raw material storage tank, a raw material feeder, an electronic scale, a weighing instrument, a distribution and transfer auger, a batching conveyor, and a central control cabinet. The electronic scale has a weighing sensor on its outer side and a discharge pipe at its bottom, connected to the inlet of the distribution and transfer auger, which conveys the raw materials to the batching conveyor. The weighing sensor acquires the weight information of the raw materials and uploads it to the weighing instrument. The weighing instrument compares the weight information with set information and generates a PID control signal based on the difference. The central control cabinet is connected to the feeder and controls the feeder's feeding speed according to the weighing instrument. By updating the batching control method, the original PLC-based batching control is transferred to instrument-based control, reducing time wastage caused by communication delays and improving batching accuracy, thus significantly improving batching efficiency and accuracy.

[0004] As mentioned in the above application, existing glass melting raw materials are mostly transported using belt conveyor devices. Existing belt conveyor devices have many disadvantages, such as large size, dust problems caused by the inability to be completely sealed, limited conveying angle, and high failure rate due to their complex mechanical structure. Summary of the Invention

[0005] To address the aforementioned problems, this invention provides a closed-loop conveying device for glass melting raw materials based on industrial vision.

[0006] The present invention adopts the following technical solution: a closed conveying device for glass melting raw materials based on industrial vision, including a conveying pipeline, a centrifugal blower, a mixing machine, a pulse bag filter and a control unit. The conveying pipeline consists of an air intake pipe and an air output pipe. The discharge port of the mixing machine is connected to the inlet of the centrifugal blower through a rotary valve and the air intake pipe. The outlet of the centrifugal blower is connected to the inlet of the pulse bag filter through the air output pipe.

[0007] The control unit includes a data acquisition module, a model training module, and a controller;

[0008] The data acquisition module is used to collect historical raw material conveying parameters. The historical raw material conveying parameters are collected under normal and stable conveying conditions. The historical raw material conveying parameters include comprehensive raw material conveying parameters and raw material feeding speed. The comprehensive raw material conveying parameters include raw material conveying coefficient, pipeline conveying coefficient, and air flow data.

[0009] The model training module trains a machine learning model to predict raw material feeding speed data based on historical raw material conveying parameters, collects real-time comprehensive raw material conveying parameters, and predicts raw material feeding speed data based on the trained machine learning model.

[0010] The controller controls the material feeding speed of the rotary valve based on the predicted material feeding speed.

[0011] As a further description of the above technical solution: the parameters affecting the raw material conveying coefficient include raw material particle size data and raw material mass data per unit volume;

[0012] The parameters that affect the pipeline transport coefficient include the transport height difference, the pipeline friction coefficient, and the pipeline length.

[0013] As a further description of the above technical solution: the method for acquiring raw material particle size data includes:

[0014] A laser particle size analyzer is used to analyze the particle size of glass raw materials. By measuring the intensity of scattered light from particles in a liquid based on the principle of laser scattering, the particle size distribution of the particles is obtained, and the raw material particle size data of the glass raw materials is obtained.

[0015] The method for obtaining the mass data per unit volume of the raw material includes:

[0016] Take a predetermined volume of molten raw material and place it in a container. Use a weighing device to weigh the total weight of the container and the raw material. Subtract the weight of the container itself to obtain the net weight of the raw material. Calculate the mass per unit volume based on the net weight of the raw material and the volume of the container used.

[0017] As a further description of the above technical solution: the expression for the raw material conveying coefficient is:

[0018]

[0019] In the formula, YX is the raw material conveying coefficient, Ld is the raw material particle size data, and Z1 is the mass data per unit volume of the raw material. and As a weighting factor, and All are greater than 0.

[0020] As a further description of the above technical solution: the method for acquiring the transport height difference data includes:

[0021] Obtain the transport path of the molten raw material in the air intake pipe and air output pipe;

[0022] Using the ground as a reference, obtain the height values ​​of the lowest and highest points along the transport path;

[0023] The expression for transmitting height difference data is: Hc = H1 - H2;

[0024] In the formula, Hc represents the transmission height difference data, H1 represents the height value of the highest point, and H2 represents the height value of the lowest point.

[0025] The method for obtaining the pipeline friction coefficient includes:

[0026] Pressure sensors are deployed at the connection between the air intake pipe and the rotary valve and at the outlet of the air output pipe. When the centrifugal blower is turned on, the pressure sensors collect the first pressure value at the connection between the air intake pipe and the rotary valve and the second pressure value at the outlet of the air output pipe.

[0027] The expression for the coefficient of friction of a pipeline is:

[0028] In the formula, F mx Where p1 is the pressure at the connection between the air intake pipe and the rotary valve, p2 is the pressure at the outlet of the air output pipe, and D is the coefficient of friction of the pipe. gd L is the inner diameter of the conveying pipe. gd The pipe length data, i.e., the length from the connection between the air intake pipe and the rotary valve to the air outlet of the air output pipe, is obtained directly through measurement. ρ gd To determine the density of a dissolving material, it is calculated by placing the material into a container of known volume, measuring its mass, and then dividing the mass by the volume. (V) gd The flow rate of the raw materials being melted is directly measured by installing a flow rate sensor in the conveying pipeline.

[0029] As a further description of the above technical solution: the expression for the pipeline transport coefficient is:

[0030]

[0031] In the formula, Gd xs This is the pipeline transport coefficient. and As a weighting factor, and All are greater than 0.

[0032] As a further description of the above technical solution: the training method for the machine learning model for predicting raw material feeding speed data includes:

[0033] The collected historical raw material transportation parameters are converted into a corresponding set of feature vectors;

[0034] The collected comprehensive parameters of raw material transportation are used as input to the machine learning model. The machine learning model takes the raw material feeding speed corresponding to each set of comprehensive parameters as output, the actual raw material feeding speed corresponding to each set of comprehensive parameters as prediction target, and minimizing the loss function value of the machine learning model as training target. Training stops when the loss function value of the machine learning model is less than or equal to the preset target loss value.

[0035] As a further description of the above technical solution: the mixing machine includes a mixing box, a frame is installed at the top center of the mixing box, a mixing motor is bolted to the frame, the output shaft of the mixing motor is fixed to a mixing shaft through a coupling, the mixing shaft extends into the mixing box, a mixing rod is welded to the mixing shaft inside the mixing box, and a feeding pipe is provided on the top of the mixing box outside the frame.

[0036] As a further description of the above technical solution: the rotary valve includes a valve body, a valve shaft is rotatably connected inside the valve body, and both ends of the valve shaft extend to the outside of the valve body. Several blades are welded on the valve shaft inside the valve body. A mounting bracket is welded on the outer wall of the valve body. A drive motor for driving the valve shaft to rotate is bolted on the mounting bracket. Connecting parts are welded at the openings at the bottom and top of the valve body, and bolt holes are opened on the connecting parts.

[0037] As a further description of the above technical solution: an air filter and a V-cone flow meter are installed at the air inlet end of the air intake pipe, and the V-cone flow meter is used to measure the air flow during the delivery process.

[0038] Beneficial effects:

[0039] In the above technical solution, the device for closed conveying of glass melting raw materials based on industrial vision provided by the present invention mainly consists of a conveying pipeline, a centrifugal blower, a pulse bag filter and a control unit. It adopts closed conveying through a conveying pipeline, which overcomes the problems of large volume, unavoidable dust, limited conveying angle and complex mechanical structure and high failure rate of the prior art using belt conveying.

[0040] Furthermore, when collecting historical raw material conveying parameters, this invention comprehensively considers the raw material conveying coefficient and the pipeline conveying coefficient, and comprehensively considers various factors affecting the conveying stability of the molten raw materials. Then, by collecting historical raw material conveying parameters, the machine learning model is trained, and the real-time raw material feeding speed is predicted based on the trained machine learning model, thereby improving the accuracy of the machine learning model's prediction. Then, the controller controls the raw material feeding speed of the rotary valve based on the predicted raw material feeding speed, thereby ensuring the stability of the molten raw materials conveyed in the conveying pipeline. Attached Figure Description

[0041] The present invention will be further explained below with reference to the accompanying drawings and embodiments:

[0042] Figure 1 A schematic diagram of the structure of the closed conveying device for glass melting raw materials based on industrial vision provided by the present invention;

[0043] Figure 2 This is a schematic diagram of the structure of the mixing machine provided by the present invention;

[0044] Figure 3 This is a schematic diagram of the rotary valve provided by the present invention;

[0045] Figure 4 A flowchart of the control unit provided by the present invention.

[0046] In the diagram: 1. Air intake pipe; 11. Conveying pipe; 2. Air output pipe; 3. Centrifugal blower; 4. Mixer; 41. Mixing box; 42. Frame; 43. Mixing motor; 44. Feeding pipe; 45. Mixing shaft; 46. Mixing rod; 401. Rotary valve; 411. Valve body; 412. Valve shaft; 413. Blade; 414. Mounting bracket; 415. Drive motor; 416. Connecting parts; 5. Pulse bag filter; 6. Control unit; 7. Air filter; 8. V-cone flow meter. Detailed Implementation

[0047] To make the technical means, creative features, objectives, and effects of this invention readily understandable, the invention is further described below with reference to specific illustrations. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0048] Example 1

[0049] Please see Figures 1-4 This invention provides a technical solution: a closed conveying device for glass melting raw materials based on industrial vision, comprising a conveying pipe 11, a centrifugal blower 3, a mixer 4, a pulse bag filter 5, and a control unit 6. The conveying pipe 11 consists of an air intake pipe 1 and an air output pipe 2. The discharge port of the mixer 4 is connected to the inlet of the centrifugal blower 3 via a rotary valve 401 and the air intake pipe 1. The outlet of the centrifugal blower 3 is connected to the inlet of the pulse bag filter 5 via the air output pipe 2. An air filter 7 and a V-cone flow meter 8 are installed at the air inlet end of the air intake pipe 1. The V-cone flow meter 8 is used to measure the air flow during the conveying process, and the air filter 7 is used to filter the air sent into the conveying pipe 11 to ensure air quality and prevent dust pollution.

[0050] Among them, the centrifugal blower 3 is used to generate sufficient airflow to transport the material to the destination through the conveying pipe 11; the pulse bag filter 5 is used to separate air and material at the end of the material conveying, reduce dust and recover the material; the mixing machine 4 mixes the glass melting raw materials evenly and discharges them through the rotary valve 401, ready for conveying; the rotary valve 401 is used to control the speed of the material discharged from the mixing machine 4, that is, the raw material feeding speed.

[0051] Control unit 6 includes a data acquisition module, a model training module, and a controller;

[0052] The data acquisition module is used to collect historical raw material conveying parameters. These parameters are collected under normal and stable conveying conditions and include comprehensive raw material conveying parameters and raw material feeding speed. The comprehensive raw material conveying parameters include the raw material conveying coefficient, pipeline conveying coefficient, and air flow data.

[0053] It should be noted that the air flow data is obtained directly through the V-cone flow meter 8.

[0054] The model training module trains a machine learning model to predict raw material feeding speed data based on historical raw material conveying parameters, collects real-time comprehensive raw material conveying parameters, and predicts raw material feeding speed data based on the trained machine learning model.

[0055] The controller controls the raw material feeding speed of the rotary valve 401 based on the predicted raw material feeding speed. The controller controls the drive motor 415 to change the rotation speed of the blade 413 driven by the drive motor 415, thereby changing the raw material feeding speed of the rotary valve 401. Increasing the rotation speed of the blade 413 inside the rotary valve 401 can speed up the speed at which the molten raw material passes through the rotary valve 401, thereby increasing the raw material feeding speed. Conversely, decreasing the rotation speed will slow down the raw material feeding speed.

[0056] Example 2:

[0057] Reference Figures 1-3 This embodiment further discloses, based on Embodiment 1, that:

[0058] Parameters affecting the raw material conveying coefficient include raw material particle size data and raw material mass data per unit volume;

[0059] The parameters that affect the pipeline transport coefficient include the transport height difference, the pipeline friction coefficient, and the pipeline length.

[0060] It should be noted that different types of glass have different raw material ratios and types, resulting in different weight per unit volume and particle size of the raw materials. Because of the different particle size and weight per unit volume of the raw materials, the required airflow data for their transportation also differs.

[0061] Methods for obtaining raw material particle size data include:

[0062] A laser particle size analyzer is used to analyze the particle size of glass raw materials. By measuring the intensity of scattered light from particles in a liquid based on the principle of laser scattering, the particle size distribution of the particles is obtained, and the raw material particle size data of the glass raw materials is obtained.

[0063] Methods for obtaining mass data per unit volume of raw materials include:

[0064] Take a predetermined volume of molten raw material and place it in a container. Use a weighing device to weigh the total weight of the container and the raw material. Subtract the weight of the container itself to obtain the net weight of the raw material. Calculate the mass per unit volume based on the net weight of the raw material and the volume of the container used.

[0065] The expression for the raw material conveying coefficient is:

[0066]

[0067] In the formula, YX is the raw material conveying coefficient, Ld is the raw material particle size data, and Zl is the raw material mass per unit volume. and As a weighting factor, and All are greater than 0.

[0068] It should be noted that the weighting coefficient is a specific value obtained by quantifying each data point to facilitate subsequent comparison. The size of the weighting coefficient depends on the number of comprehensive parameters and the weighting coefficient initially set by those skilled in the art for each set of comprehensive parameters.

[0069] Secondly, the larger the raw material particle size data and the larger the quality data submitted by the raw material unit, the larger the raw material conveying coefficient. Under the same air flow data, the larger the raw material conveying coefficient, the slower the raw material feeding speed.

[0070] Methods for acquiring transmission height difference data include:

[0071] Obtain the transport path of the molten raw material in the air intake pipe 1 and the air output pipe 2;

[0072] Using the ground as a reference, obtain the height values ​​of the lowest and highest points along the transport path;

[0073] The expression for transmitting height difference data is: Hc = H1 - H2;

[0074] In the formula, Hc represents the transmission height difference data, H1 represents the height value of the highest point, and H2 represents the height value of the lowest point.

[0075] It should be noted that during the transportation of the raw materials for melting, they are lifted so that they can be directly discharged into the raw material silo of the glass melting furnace for melting and processing.

[0076] Methods for obtaining the friction coefficient of pipelines include:

[0077] Pressure sensors are deployed at the connection between the air intake pipe 1 and the rotary valve 401 and at the outlet of the air output pipe 2. When the centrifugal blower 3 is turned on, the pressure sensors collect the first pressure value at the connection between the air intake pipe 1 and the rotary valve 401 and the second pressure value at the outlet of the air output pipe 2.

[0078] The expression for the coefficient of friction of a pipeline is:

[0079] In the formula, F mx Where p1 is the coefficient of friction of the pipe, p2 is the pressure value at the connection between the air intake pipe 1 and the rotary valve 401, and p2 is the pressure value at the outlet of the air output pipe 2. gd L is the inner diameter of the conveying pipe 11. gd The pipe length data, i.e., the length from the connection point of the air intake pipe 1 and the rotary valve 401 to the air outlet of the air output pipe 2, is obtained directly through measurement. ρ gd To determine the density of a dissolving material, it is calculated by placing the material into a container of known volume, measuring its mass, and then dividing the mass by the volume. (V) gd The flow rate of the raw materials being melted is directly measured by installing a flow rate sensor in the conveying pipe 11.

[0080] The expression for the pipeline transport coefficient is:

[0081]

[0082] In the formula, Gd xs This is the pipeline transport coefficient. and As a weighting factor, and All are greater than 0.

[0083] It should be noted that the weighting coefficient is a specific value obtained by quantifying each data point to facilitate subsequent comparison. The size of the weighting coefficient depends on the number of comprehensive parameters and the weighting coefficient initially set by those skilled in the art for each set of comprehensive parameters.

[0084] Secondly, the larger the height difference in the conveying height, the greater the height that the raw materials need to be conveyed and lifted, and the greater the required airflow. Conversely, the smaller the height difference, the greater the airflow. The larger the pipe friction coefficient, the greater the friction between the two objects. Even under the same external force, the relative motion between the two objects will be more difficult. That is, the larger the pipe friction coefficient, the greater the resistance that the raw materials need to be conveyed, and the greater the required airflow. Conversely, the smaller the pipe length, the greater the length that the raw materials need to be conveyed, and the greater the required airflow. Conversely, the smaller the pipe length, the greater the length that the raw materials need to be conveyed, and the greater the required airflow.

[0085] Training methods for machine learning models that predict raw material feeding speed data include:

[0086] The collected historical raw material transportation parameters are converted into a corresponding set of feature vectors;

[0087] The collected comprehensive parameters of raw material transportation are used as input to the machine learning model. The machine learning model takes the raw material feeding speed corresponding to each set of comprehensive parameters as output, the actual raw material feeding speed corresponding to each set of comprehensive parameters as prediction target, and minimizing the loss function value of the machine learning model as training target. Training stops when the loss function value of the machine learning model is less than or equal to the preset target loss value.

[0088] The machine learning model can be either a deep neural network model or a deep belief network model, and the loss function of the machine learning model is the mean squared error.

[0089] Mean squared error is one of the commonly used loss functions. It is obtained by... By training the model with minimization as the objective, the machine learning model can better fit the data, thereby improving the model's performance and accuracy.

[0090] In the loss function, MSE1 represents the loss function value of the machine learning model, x is the feature vector group number, m is the number of feature vector groups, and yx represents the raw material feeding speed data corresponding to the x-th feature vector group. The x-th feature vector corresponds to the real-time raw material feeding speed data;

[0091] Other model parameters of the machine learning model, such as the target loss value, optimization algorithm, ratio of training set to test set to validation set, and optimization of the loss function, are all obtained through actual engineering implementation and continuous experimental tuning.

[0092] Example 3

[0093] Reference Figure 1 and Figure 2 This embodiment further discloses a mixing machine 4, which includes a mixing box 41. A frame 42 is installed at the top center of the mixing box 41. A mixing motor 43 is bolted to the frame 42. The output shaft of the mixing motor 43 is fixed to a mixing shaft 45 through a coupling. The mixing shaft 45 extends into the mixing box 41. A mixing rod 46 is welded to the mixing shaft 45 inside the mixing box 41. A feeding pipe 44 is provided on the top of the mixing box 41 outside the frame 42.

[0094] Specifically, during use, the glass melting raw material is added into the mixing tank 41 through the feeding pipe 44. The stirring motor 43 is turned on, which drives the stirring shaft 45 to rotate. The stirring shaft 45 drives the stirring rod 46 to rotate, thereby stirring and mixing the glass melting raw material in the mixing tank 41. After the mixture is evenly mixed, the stirring motor 43 is turned off.

[0095] Example 4

[0096] Reference Figure 1 and Figure 3 This embodiment further discloses a rotary valve 401, which includes a valve body 411. A valve shaft 412 is rotatably connected inside the valve body 411, and both ends of the valve shaft 412 extend to the outside of the valve body 411. Several blades 413 are welded on the valve shaft 412 inside the valve body 411. A mounting bracket 414 is welded on the outer wall of the valve body 411. A drive motor 415 for driving the valve shaft 412 to rotate is bolted on the mounting bracket 414. Connecting parts 416 are welded at the openings at the bottom and top of the valve body 411, and bolt holes are provided on the connecting parts 416.

[0097] Specifically, during use, when the raw materials in the mixing tank 41 are mixed evenly, the drive motor 415 is turned on for conveying. The rotation of the drive motor 415 drives several blades 413 to rotate, thereby conveying the molten raw materials from the mixing tank 41 to the air intake pipe 1. By changing the rotation speed of the blades 413, the feeding speed of the raw materials can be adjusted.

[0098] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A closed-loop conveying device for glass melting raw materials based on industrial vision, comprising a conveying pipe (11), a centrifugal blower (3), a mixing machine (4), a pulse bag filter (5), and a control unit (6), wherein the conveying pipe (11) consists of an air intake pipe (1) and an air output pipe (2), characterized in that, The discharge port of the mixer (4) is connected to the inlet of the centrifugal blower (3) through a rotary valve (401) and an air intake pipe (1), and the outlet of the centrifugal blower (3) is connected to the inlet of the pulse bag filter (5) through an air output pipe (2). The control unit (6) includes a data acquisition module, a model training module, and a controller; The data acquisition module is used to collect historical raw material conveying parameters. The historical raw material conveying parameters are collected under normal and stable conveying conditions. The historical raw material conveying parameters include comprehensive raw material conveying parameters and raw material feeding speed. The comprehensive raw material conveying parameters include raw material conveying coefficient, pipeline conveying coefficient, and air flow data. The model training module trains a machine learning model to predict raw material feeding speed data based on historical raw material conveying parameters, collects real-time comprehensive raw material conveying parameters, and predicts raw material feeding speed data based on the trained machine learning model. The controller controls the material feeding speed of the rotary valve (401) based on the predicted material feeding speed.

2. The closed-loop conveying device for glass melting raw materials based on industrial vision according to claim 1, characterized in that, Parameters affecting the raw material conveying coefficient include raw material particle size data and raw material mass data per unit volume; The parameters that affect the pipeline transport coefficient include the transport height difference, the pipeline friction coefficient, and the pipeline length.

3. The closed-loop conveying device for glass melting raw materials based on industrial vision according to claim 1, characterized in that, The method for obtaining raw material particle size data includes: A laser particle size analyzer is used to analyze the particle size of glass raw materials. By measuring the intensity of scattered light from particles in a liquid based on the principle of laser scattering, the particle size distribution of the particles is obtained, and the raw material particle size data of the glass raw materials is obtained. The method for obtaining the mass data per unit volume of the raw material includes: Take a predetermined volume of molten raw material and place it in a container. Use a weighing device to weigh the total weight of the container and the raw material. Subtract the weight of the container itself to obtain the net weight of the raw material. Calculate the mass per unit volume based on the net weight of the raw material and the volume of the container used.

4. The closed-loop conveying device for glass melting raw materials based on industrial vision according to claim 1, characterized in that, The expression for the raw material conveying coefficient is: In the formula, YX is the raw material conveying coefficient, Ld is the raw material particle size data, and Z1 is the mass data per unit volume of the raw material. and As a weighting factor, and All are greater than 0.

5. The closed-loop conveying device for glass melting raw materials based on industrial vision according to claim 1, characterized in that, The method for acquiring the transport height difference data includes: Obtain the transport path of the molten raw material in the air intake pipe (1) and the air output pipe (2); Using the ground as a reference, obtain the height values ​​of the lowest and highest points along the transport path; The expression for transmitting height difference data is: Hc = H1 - H2; In the formula, Hc represents the transmission height difference data, H1 represents the height value of the highest point, and H2 represents the height value of the lowest point. The method for obtaining the pipeline friction coefficient includes: Air pressure sensors are deployed at the connection between the air intake pipe (1) and the rotary valve (401) and at the outlet of the air output pipe (2). The centrifugal blower (3) is turned on, and the first air pressure value at the connection between the air intake pipe (1) and the rotary valve (401) and the second air pressure value at the outlet of the air output pipe (2) are collected by the air pressure sensors. The expression for the coefficient of friction of a pipeline is: In the formula, F mx D is the coefficient of friction of the pipeline, p1 is the pressure value at the connection between the air intake pipe (1) and the rotary valve (401), p2 is the pressure value at the outlet of the air output pipe (2), and D is the pressure value at the outlet of the air output pipe (2). gd L is the inner diameter of the conveying pipe (11). gd The pipe length data, i.e., the length from the connection between the air intake pipe (1) and the rotary valve (401) to the air outlet of the air output pipe (2), is obtained directly through measurement. ρ gd To determine the density of a dissolving material, it is calculated by placing the material into a container of known volume, measuring its mass, and then dividing the mass by the volume. (V) gd The flow rate of the raw materials to be melted is directly measured by installing a flow rate sensor in the conveying pipe (11).

6. The closed-loop conveying device for glass melting raw materials based on industrial vision according to claim 1, characterized in that, The expression for the pipeline transport coefficient is: In the formula, Gd xs This is the pipeline transport coefficient. and As a weighting factor, and All are greater than 0.

7. The closed-loop conveying device for glass melting raw materials based on industrial vision according to claim 1, characterized in that, Training methods for machine learning models that predict raw material feeding speed data include: The collected historical raw material transportation parameters are converted into a corresponding set of feature vectors; The collected comprehensive parameters of raw material transportation are used as input to the machine learning model. The machine learning model takes the raw material feeding speed corresponding to each set of comprehensive parameters as output, the actual raw material feeding speed corresponding to each set of comprehensive parameters as prediction target, and minimizing the loss function value of the machine learning model as training target. Training stops when the loss function value of the machine learning model is less than or equal to the preset target loss value.

8. The closed-loop conveying device for glass melting raw materials based on industrial vision according to claim 1, characterized in that, The mixing machine (4) includes a mixing box (41), a frame (42) is installed at the top center of the mixing box (41), a mixing motor (43) is bolted on the frame (42), the output shaft of the mixing motor (43) is fixed to a mixing shaft (45) through a coupling, the mixing shaft (45) extends into the mixing box (41), a mixing rod (46) is welded on the mixing shaft (45) inside the mixing box (41), and a feeding pipe (44) is provided on the top of the mixing box (41) outside the frame (42).

9. The closed-loop conveying device for glass melting raw materials based on industrial vision according to claim 1, characterized in that, The rotary valve (401) includes a valve body (411), a valve shaft (412) is rotatably connected inside the valve body (411), and both ends of the valve shaft (412) extend to the outside of the valve body (411). Several blades (413) are welded on the valve shaft (412) inside the valve body (411). A mounting bracket (414) is welded on the outer wall of the valve body (411). A drive motor (415) for driving the valve shaft (412) to rotate is bolted on the mounting bracket (414). Connecting parts (416) are welded at the openings at the bottom and top of the valve body (411), and bolt holes are provided on the connecting parts (416).

10. The closed-loop conveying device for glass melting raw materials based on industrial vision according to claim 1, characterized in that, An air filter (7) and a V-cone flow meter (8) are installed at the air inlet end of the air intake pipe (1). The V-cone flow meter (8) is used to measure the air flow during the conveying process.

Images