Virtual simulation method for cut tobacco drying process

Abstract

The invention discloses a virtual simulation method for a cut tobacco drying process, and the method comprises the steps: arranging a plurality of high-speed camera devices at a plurality of positions of a cut tobacco drying cavity, so as to synchronously capture an image sequence of flowing cut tobacco from different angles; velocity vectors of the tobacco shreds at multiple angles are calculated according to the image sequence; performing data fusion on the velocity vectors of the multiple angles to obtain three-dimensional flow velocity distribution data of the tobacco shreds; based on the three-dimensional flow velocity distribution data, establishing a cut tobacco motion model, and combining the temperature data and the humidity data to construct a heat transfer and moisture transmission model; different process parameters are input into the virtual simulation system, the virtual simulation system is operated, and tobacco shred state changes under the different process parameters are simulated. According to the virtual simulation method for the cut tobacco drying process, the cut tobacco drying effects under different process parameters are simulated and optimized by accurately measuring the three-dimensional flow velocity distribution of the cut tobacco in the cut tobacco drying cavity and combining heat transfer and moisture transfer models.

Description

Technical Field

[0001] The present invention relates to the technical field of tobacco processing, and more specifically, to a virtual simulation method for a tobacco drying process. Background Art

[0002] In the tobacco production process, the tobacco drying process is a crucial link, which directly affects the drying effect, aroma retention and final product quality of tobacco. In the traditional tobacco drying process, the adjustment of process parameters such as air flow velocity, temperature and humidity mainly relies on experience and manual control, which is difficult to accurately and in real time respond to the flow changes of tobacco in the cavity. This imprecise control can easily lead to uneven drying of tobacco, resulting in fluctuations in product quality and even waste of energy.

[0003] In recent years, with the development of production automation technology, non-contact measurement technologies such as optical flow method have gradually been applied to tobacco flow velocity monitoring. However, the traditional optical flow method can only obtain velocity information in a two-dimensional plane, and it is difficult to reflect the three-dimensional flow velocity distribution of tobacco in the drying chamber. At the same time, the influence of environmental parameters such as temperature and humidity is also difficult to be effectively integrated through simple measurements.

[0004] Therefore, there is an urgent need for a virtual simulation method for the silk drying process. Summary of the invention

[0005] The purpose of the present invention is to provide a virtual simulation method for the tobacco drying process to solve the problems in the above-mentioned prior art. It can simulate and optimize the tobacco drying effect under different process parameters by accurately measuring the three-dimensional flow velocity distribution of tobacco in the drying chamber and combining the heat transfer and moisture transfer models.

[0006] The present invention provides a virtual simulation method for a tofu drying process, which includes:

[0007] Arrange multiple high-speed camera devices at multiple positions of the tobacco drying chamber to synchronously capture image sequences of flowing tobacco from different angles;

[0008] Calculating the velocity vectors of the tobacco at multiple angles according to the image sequences captured by the high-speed camera devices;

[0009] Perform data fusion on velocity vectors at multiple angles to obtain three-dimensional velocity distribution data of tobacco;

[0010] Based on the three-dimensional flow velocity distribution data, a movement model of the tobacco in the drying chamber is established, and a heat transfer model and a moisture transfer model are constructed in combination with the temperature data and the humidity data to simulate the drying state of the tobacco under different process conditions;

[0011] Different process parameters are input into the virtual simulation system, and the virtual simulation system is run to simulate the state changes of tobacco under different process parameters, wherein the virtual simulation system integrates three-dimensional flow velocity distribution data of tobacco, dynamic changes of temperature field obtained based on the heat transfer model, and dynamic changes of humidity field obtained based on the moisture transfer model.

[0012] The virtual simulation method for the tobacco drying process as described above, wherein preferably, the velocity vectors of the tobacco at multiple angles are calculated according to the image sequences captured by each of the high-speed camera devices, specifically including:

[0013] A synchronous trigger system is used to trigger all high-speed cameras to start shooting at the same time and capture image sequences at the same frame rate;

[0014] The multi-angle image sequences acquired by each of the high-speed camera devices are input into a computing unit, wherein the computing unit is configured with an image processing module based on an optical flow method, so as to process the image sequences through the image processing module to calculate the velocity vectors of the tobacco at multiple angles.

[0015] The virtual simulation method for the tobacco drying process as described above, wherein preferably, the multi-angle image sequence acquired by each high-speed camera device is input into a calculation unit, wherein the calculation unit is configured with an image processing module based on an optical flow method, so as to process the image sequence through the image processing module to calculate the velocity vector of the tobacco at multiple angles, specifically comprising:

[0016] Preprocessing the image series, wherein the preprocessing includes at least one of grayscale processing, denoising processing and enhancement processing;

[0017] Feature point detection and tracking: Through the feature point detection algorithm, the key feature points of the tobacco are detected in the preprocessed image;

[0018] The key feature points are tracked in the image of each perspective, and the position changes of the key feature points in adjacent frames are calculated to obtain the velocity information of the key feature points.

[0019] The virtual simulation method for the tobacco drying process as described above, wherein preferably, the data fusion of the velocity vectors at multiple angles to obtain the three-dimensional flow velocity distribution data of the tobacco shreds specifically includes:

[0020] The three-dimensional flow velocity distribution data is obtained by calculating the optical flow algorithm of multi-angle images, wherein the optical flow algorithm uses the Lucas-Kanade method to track the characteristic points of the tobacco;

[0021] The three-dimensional flow velocity distribution data is fused from multiple perspectives.

[0022] The virtual simulation method for the tobacco drying process as described above, wherein preferably, the three-dimensional flow velocity distribution data is obtained by calculating the optical flow algorithm of the multi-angle image, wherein the optical flow algorithm uses the Lucas-Kanade method to track the characteristic points of the tobacco, specifically including:

[0023] The optical flow method is used to calculate the velocity vector of the key feature points on the two-dimensional plane by analyzing the brightness changes of the image. The optical flow calculation is based on the following basic optical flow equation:

[0024] I(x,yy,t)=I(x+Δx,y+Δy,t+Δt)

[0025] After linearization, the optical flow constraint equation is obtained:

[0026]

[0027] Among them, u represents the component of the optical flow velocity vector in the x direction, and v represents the component of the optical flow velocity vector in the y direction;

[0028] The Lucas-Kanade algorithm is as follows: In a local window of the image, the optical flow constraint equations of multiple pixels are used to solve the velocity vector

[0029]

[0030] Among them, A represents the image gradient matrix:

[0031]

[0032] b represents the negative vector of the time gradient:

[0033]

[0034] The multi-view fusing of the three-dimensional velocity distribution data specifically includes: performing multi-view fusing by weighted averaging method or Kalman filtering method,

[0035] Among them, the weighted average method for multi-view fusion specifically includes:

[0036] Triangulation: Reconstruct the position of a 3D point P = (X, Y, Z) through velocity vectors from multiple perspectives, using the projection matrix P of two or more perspectives i and P j , solve the following minimization problem to reconstruct the 3D points:

[0037]

[0038] Among them, P i,3 and P j,3Respectively represent the third row of the projection matrix;

[0039] Calculation of three-dimensional velocity vector: According to the three-dimensional position change ΔP of the three-dimensional point P in consecutive frames, the velocity vector is calculated by the following formula:

[0040]

[0041] Multi-view velocity fusion: Assume that the velocity vectors obtained from multiple viewpoints are and These vectors are combined to get the three-dimensional flow velocity:

[0042]

[0043] Among them, w i and w j Represents the weight of the perspective.

[0044] The virtual simulation method for the tobacco drying process as described above, wherein preferably, based on the three-dimensional flow velocity distribution data, a movement model of tobacco in the tobacco drying chamber is established, and in combination with the temperature data and the humidity data, a heat transfer model and a moisture transfer model are constructed to simulate the drying state of tobacco under different process conditions, specifically including:

[0045] Arrange temperature sensors and humidity sensors in the drying chamber;

[0046] The temperature data and humidity data are collected in real time through the temperature sensor and humidity sensor respectively;

[0047] According to the collected temperature data, humidity data and three-dimensional flow velocity distribution data, a heat transfer model of the tofu drying chamber is established to simulate the heat transfer in the chamber. The heat transfer model follows Fourier's heat conduction law:

[0048]

[0049] Where q represents the heat flux, k represents the thermal conductivity, represents the temperature gradient;

[0050] Based on the collected temperature data, humidity data and three-dimensional flow velocity distribution data, a moisture transfer model is established to evaluate the influence of humidity and air flow velocity on the drying rate of tobacco. The moisture transfer model follows Fick's diffusion law and is used to simulate the moisture transfer and evaporation process in tobacco:

[0051]

[0052] Where J represents the water diffusion flux, D represents the diffusion coefficient, and C represents the water concentration.

[0053] The virtual simulation method for the tobacco drying process as described above, wherein preferably, inputting different process parameters into the virtual simulation system, running the virtual simulation system, and simulating the tobacco state changes under different process parameters specifically includes:

[0054] Different combinations of process parameters are input into the virtual simulation system, and the virtual simulation system outputs simulation results to demonstrate the tobacco drying effects under different conditions.

[0055] As described above, in the virtual simulation method for tobacco drying process, preferably, the process parameter combination includes at least one of air flow velocity, temperature and humidity, fan speed, and heater power, and the simulation results include tobacco flow velocity distribution diagrams, tobacco drying rate curves, and dynamic changes of temperature field and humidity field corresponding to different process parameter combinations.

[0056] As described above, in the virtual simulation method for the silk drying process, preferably, the virtual simulation system adopts a finite element analysis model, and the finite element analysis model performs fine grid division on the airflow, heat and humidity in different areas of the cavity; the virtual simulation system is optimized based on a machine learning algorithm, and the prediction model is trained by inputting historical production data and simulation results to estimate the optimal combination of process parameters.

[0057] In the virtual simulation method for tobacco drying process as described above, preferably, the simulation results output by the virtual simulation system are used to optimize the tobacco drying process in actual production; the simulation data generated by the virtual simulation system is associated with the production batch data and stored in a database for building a quality traceability system; the virtual simulation system displays the dynamic changes of flow rate, temperature and humidity of tobacco through a three-dimensional visualization interface.

[0058] The present invention provides a virtual simulation method for a tobacco drying process and a virtual simulation solution based on a multi-angle optical flow method. The method simulates and optimizes the tobacco drying effect under different process parameters by accurately measuring the three-dimensional flow velocity distribution of tobacco in a tobacco drying chamber and combining heat transfer and moisture transfer models. The method can not only simulate the flow and drying process of tobacco under different conditions, but also automatically optimize process parameters through a machine learning algorithm, thereby providing an efficient, energy-saving and intelligent production solution, and greatly improving the automation level of tobacco production and the consistency of product quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention will be further described below with reference to the accompanying drawings, in which:

[0060] Figure 1 The present invention is a flowchart of an embodiment of a virtual simulation method for a tofu drying process provided by the present invention. DETAILED DESCRIPTION

[0061] Various exemplary embodiments of the present disclosure will now be described in detail with reference to the accompanying drawings. The description of the exemplary embodiments is merely illustrative and is in no way intended to limit the present disclosure and its application or use. The present disclosure can be implemented in many different forms and is not limited to the embodiments described herein. These embodiments are provided to make the present disclosure thorough and complete and to fully express the scope of the present disclosure to those skilled in the art. It should be noted that unless otherwise specifically stated, the relative arrangement of the components and steps, the composition of the materials, the numerical expressions and the numerical values ​​set forth in these embodiments should be interpreted as being merely exemplary and not as limiting.

[0062] The words "first", "second" and similar words used in this disclosure do not indicate any order, quantity or importance, but are only used to distinguish different parts. The words "include" or "comprises" and similar words mean that the elements before the word include the elements listed after the word, and do not exclude the possibility of including other elements. "Up", "down" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0063] In the present disclosure, when a specific component is described as being located between a first component and a second component, there may or may not be an intermediate component between the specific component and the first component or the second component. When a specific component is described as being connected to other components, the specific component may be directly connected to the other components without an intermediate component, or may not be directly connected to the other components but have an intermediate component.

[0064] All terms (including technical terms or scientific terms) used in the present disclosure have the same meanings as those understood by ordinary technicians in the field to which the present disclosure belongs, unless otherwise specifically defined. It should also be understood that terms defined in general dictionaries, such as general dictionaries, should be interpreted as having meanings consistent with their meanings in the context of the relevant technology, and should not be interpreted in an idealized or extremely formal sense, unless explicitly defined herein.

[0065] Technologies, methods, and equipment known to ordinary technicians in the relevant art may not be discussed in detail, but where appropriate, the technologies, methods, and equipment should be considered part of the specification.

[0066] The present invention combines a multi-angle optical flow method with a virtual simulation method of process parameter simulation to accurately control and optimize the tofu drying process.

[0067] Virtual simulation technology can comprehensively analyze the impact of different process parameters on the tobacco drying process by establishing a model of tobacco flow, temperature distribution, and humidity changes in the cavity. This can not only help factories simulate the optimal combination of process parameters before production, but also feedback the actual production process through simulation results to achieve real-time optimization and control of the tobacco drying process. In addition, the use of finite element analysis (FEA) and machine learning technology can further improve simulation accuracy and computing efficiency, providing new technical means for companies to save production costs and improve product quality.

[0068] like Figure 1 As shown, the virtual simulation method for the tofu drying process provided in this embodiment includes the following steps in actual execution:

[0069] Step S1, arranging a plurality of high-speed camera devices at a plurality of positions of the tobacco drying chamber to synchronously capture an image sequence of flowing tobacco from different angles.

[0070] Specifically, high-speed cameras are installed at multiple locations (at least two different viewing angles) in the drying chamber to capture the flow images of tobacco from different angles. The frame rate of the high-speed camera should be no less than 300 frames per second to ensure that the rapid movement and subtle changes of tobacco can be captured. It should be noted that the high-speed camera should have high resolution and good imaging capabilities under low light conditions to clearly record the flow of tobacco in the drying chamber.

[0071] Step S2: Calculate the velocity vectors of the tobacco at multiple angles according to the image sequences captured by the high-speed camera devices.

[0072] In one embodiment of the virtual simulation method for the tofu drying process of the present invention, the step S2 may specifically include:

[0073] Step S21: trigger all high-speed camera devices to start shooting at the same time through a synchronous trigger system, and capture image sequences at the same frame rate.

[0074] The synchronous trigger system ensures the temporal consistency of multi-angle images, laying the foundation for subsequent data processing and fusion.

[0075] Assume that high-speed cameras are placed at multiple locations in the drying chamber, and the viewing angle of each camera is represented by its projection matrix; the projection matrix P i Definition:

[0076] P i =K i [R i |t i ];

[0077] Among them, Ki represents the internal parameter matrix of the i-th camera device, which characterizes the internal optical characteristics of the camera device; R i represents the rotation matrix, t i Represents the translation vector, which represents the external parameter of the i-th camera device relative to the world coordinate system.

[0078] Step S22, inputting the multi-angle image sequence acquired by each of the high-speed camera devices into a calculation unit, wherein the calculation unit is configured with an image processing module based on the optical flow method, so as to process the image sequence through the image processing module to calculate the velocity vector of the tobacco at multiple angles.

[0079] In one embodiment of the virtual simulation method for the tofu drying process of the present invention, the step S22 may specifically include:

[0080] Step S221: pre-process the image series.

[0081] The preprocessing includes at least one of grayscale processing, denoising processing and enhancement processing. These preprocessings are helpful to improve the detection accuracy of feature points and the stability of optical flow calculation.

[0082] Step S222, feature point detection and tracking, through the feature point detection algorithm (such as SIFT, SURF or ORB, etc.), the key feature points of the tobacco are detected in the preprocessed image. These feature points should be consistent and stable in consecutive frames to facilitate tracking.

[0083] Step S223: Track key feature points in the image of each viewing angle, calculate the position change of the key feature points in adjacent frames, and thus obtain speed information of the key feature points.

[0084] The motion trajectory of the feature points is estimated by the optical flow algorithm. The SIFT algorithm is used as an example below. The SIFT (Scale-Invariant Feature Transform) algorithm is widely used in image processing and computer vision because it is invariant to image scaling, rotation, and illumination changes, and is suitable for feature point detection in complex environments. The SIFT algorithm includes the following steps:

[0085] 1) Gaussian pyramid construction:

[0086] In order to detect scale-invariant feature points, the image is first blurred at different scales to form a Gaussian pyramid. The Gaussian pyramid gradually blurs and shrinks the original image to form a series of images of different scales.

[0087] Each layer image L(x,y,σ) is obtained by convolving the original image I(x,y) with the Gaussian kernel G(x,y,σ):

[0088] L(x,y,σ)=G(x,y,σ)*I(x,y)

[0089] in: represents the Gaussian kernel function, and σ represents the scale parameter, which is used to control the degree of Gaussian blur.

[0090] 2) Difference of Gaussian (DoG) calculation

[0091] At each scale, a difference Gaussian pyramid is constructed by subtracting the Gaussian blurred images of adjacent scales. These difference images D(x, y, σ) enhance the edges and feature points in the image.

[0092] D(x,y,σ)=L(x,y,kσ)-L(x,y,σ)

[0093] Here, k represents a constant used to control the ratio between adjacent scales.

[0094] 3) Feature point detection:

[0095] In the differential Gaussian pyramid, for each pixel, its value is compared with the 26 points adjacent to it in space and scale (3x3x3 window). If the value of the point is the largest or smallest among these points, the point is considered to be a potential feature point.

[0096] 4) Interpolation method to accurately locate feature points

[0097] The potential feature points are accurately located using three-dimensional quadratic interpolation to obtain their precise positions in the scale space. Each feature point in the image is accurately located in three-dimensional space (x, y, scale σ) through Taylor expansion to remove low-contrast feature points and unstable edge responses.

[0098] For each candidate feature point, its exact position is determined by solving the following set of equations:

[0099]

[0100] in, Represents the displacement vector of the feature point; H represents the Hessian matrix, which is used to characterize the second-order derivative information around the feature point; Represents the gradient vector of the DoG image.

[0101] 5) Main direction calculation:

[0102] Calculate the main direction of each feature point to ensure the stability of the feature point under rotation invariance. The main direction is determined by calculating the gradient direction and gradient amplitude of the pixels around the feature point and statistically analyzing the histogram. The calculation formula for the main direction is:

[0103]

[0104] Among them, L x and L y Represent the image gradients in the x and y directions respectively.

[0105] 6) Generate feature descriptors

[0106] According to the main direction of the feature point, the pixel area around the feature point is divided to generate a feature descriptor with direction invariance. Usually, a 16x16 window is used, divided into 4x4 small areas, and an 8-dimensional direction histogram is generated for each area, forming a total of 128-dimensional feature vectors. These feature descriptors are used for subsequent feature point matching and can effectively identify the corresponding relationship of the same feature point under different viewing angles, different scales, and different rotations.

[0107] Step S3, fusing the velocity vectors at multiple angles to obtain three-dimensional flow velocity distribution data of the tobacco.

[0108] In one embodiment of the virtual simulation method for the tofu drying process of the present invention, the step S3 may specifically include:

[0109] Step S31, obtaining the three-dimensional flow velocity distribution data by calculating the optical flow algorithm of the multi-angle image, wherein the optical flow algorithm uses the Lucas-Kanade method to track the characteristic points of the tobacco.

[0110] Specifically, the optical flow method is used to calculate the velocity vector of the key feature points on the two-dimensional plane by analyzing the brightness changes of the image. The optical flow calculation is based on the following basic optical flow equation:

[0111] I(x,y,t)=I(x+Δx,y+Δy,t+Δt)

[0112] After linearization, the optical flow constraint equation is obtained:

[0113]

[0114] Among them, u represents the component of the optical flow velocity vector in the x direction, and v represents the component of the optical flow velocity vector in the y direction;

[0115] The Lucas-Kanade algorithm is as follows: In a local window of the image, the optical flow constraint equations of multiple pixels are used to solve the velocity vector

[0116]

[0117] Among them, A represents the image gradient matrix:

[0118]

[0119] b represents the negative vector of the time gradient:

[0120]

[0121] Step S32: performing multi-view fusion on the three-dimensional flow velocity distribution data.

[0122] The accuracy of the flow velocity data can be improved by multi-perspective fusion, specifically, by weighted average method or Kalman filter method.

[0123] Among them, in one embodiment of the virtual simulation method for the tofu drying process of the present invention, the weighted average method for multi-view fusion specifically includes:

[0124] Step S321, triangulation: reconstruct the position of the three-dimensional point P = (X, Y, Z) through the velocity vectors of multiple perspectives, using the projection matrix P of two or more perspectives i and P j , solve the following minimization problem to reconstruct the 3D points:

[0125]

[0126] Among them, P i,3 and P j,3 They represent the third row of the projection matrix, which is used for normalization.

[0127] Step S322, 3D velocity vector calculation: according to the 3D position change ΔP of the 3D point P in the continuous frames, the velocity vector is calculated by the following formula:

[0128] Step S323, multi-view velocity fusion: Assume that the velocity vectors obtained from multiple viewpoints are and These vectors are combined to get the three-dimensional flow velocity:

[0129]

[0130] Among them, w i and w j The weight representing the viewpoint can be adjusted based on image quality, confidence of feature point tracking, or other parameters.

[0131] Furthermore, in other embodiments of the present invention, an optimal estimation algorithm such as Kalman filtering may be used to dynamically fuse velocity information from multiple perspectives to improve noise resistance and measurement accuracy.

[0132] The present invention adopts the Lucas-Kanade algorithm and combines it with multi-frame difference technology to improve the accuracy and robustness of tobacco flow velocity measurement. The present invention uses multiple high-speed camera devices to synchronously capture the image sequence of tobacco flowing in the drying chamber from different angles, and combines the optical flow algorithm to track feature points and calculate velocity vectors, so as to accurately obtain the flow velocity distribution of tobacco in three-dimensional space. Compared with the traditional two-dimensional flow velocity measurement method, the present invention can fully grasp the three-dimensional motion state of tobacco, avoid measurement blind spots and errors, improve the accuracy of flow velocity measurement, and realize accurate three-dimensional flow velocity measurement.

[0133] Step S4: Based on the three-dimensional flow velocity distribution data, a movement model of the tobacco in the drying chamber is established, and a heat transfer model and a moisture transfer model are constructed in combination with the temperature data and the humidity data to simulate the drying state of the tobacco under different process conditions.

[0134] In one embodiment of the virtual simulation method for the tofu drying process of the present invention, the step S4 may specifically include:

[0135] Step S41, arranging a temperature sensor and a humidity sensor in the wire drying chamber.

[0136] Step S42: collecting temperature data and humidity data in real time through the temperature sensor and the humidity sensor respectively.

[0137] The temperature data and humidity data, together with the three-dimensional flow velocity distribution of the tobacco, are used as input parameters of the virtual simulation model.

[0138] Step S43: Based on the collected temperature data, humidity data and three-dimensional flow velocity distribution data, a heat transfer model of the tofu drying chamber is established to simulate the heat transfer in the chamber. The heat transfer model follows Fourier's heat conduction law:

[0139]

[0140] Where q represents the heat flux, k represents the thermal conductivity, Represents the temperature gradient.

[0141] Step S44: Based on the collected temperature data, humidity data and three-dimensional flow velocity distribution data, a moisture transfer model is established to evaluate the influence of humidity and air flow velocity on the drying rate of the tobacco shreds. The moisture transfer model follows Fick's diffusion law and is used to simulate the moisture transfer and evaporation process in the tobacco shreds:

[0142]

[0143] Where J represents the water diffusion flux, D represents the diffusion coefficient, and C represents the water concentration.

[0144] The moisture transport model was used to evaluate the effects of humidity and airflow velocity on the drying rate of cut tobacco.

[0145] Step S5, input different process parameters into the virtual simulation system, run the virtual simulation system, and simulate the state changes of tobacco under different process parameters, wherein the virtual simulation system integrates three-dimensional flow velocity distribution data of tobacco, dynamic changes of temperature field obtained based on the heat transfer model, and dynamic changes of humidity field obtained based on the moisture transfer model.

[0146] The virtual simulation system can simulate the changes in tobacco flow rate, drying rate and moisture content under different conditions based on the constructed model. Specifically, different combinations of process parameters are input into the virtual simulation system, and the virtual simulation system outputs simulation results to show the tobacco drying effect under different conditions. In one embodiment of the present invention, the process parameter combination includes at least one of air flow velocity, temperature and humidity, fan speed, and heater power, and the simulation results include the tobacco flow velocity distribution diagram, tobacco drying rate curve, and dynamic changes of temperature field and humidity field corresponding to different process parameter combinations. These results are used to predict the tobacco drying effect under different process parameters and help production personnel find the best process parameter combination. The present invention can simulate the changes in tobacco flow rate, drying rate and moisture content under different process parameters through a virtual simulation system.

[0147] The present invention establishes a tobacco shred movement model based on three-dimensional flow velocity data, and combines Fourier's heat conduction law and Fick's diffusion law to establish a temperature and humidity transfer model to achieve a comprehensive simulation of the tobacco shred drying process. By inputting different process parameters (such as air flow velocity, temperature and humidity), the virtual simulation system can simulate the flow, drying rate and moisture content of tobacco shreds under various conditions, helping production personnel find the best process parameter combination, thereby optimizing the drying effect and achieving comprehensive process parameter simulation and optimization.

[0148] The present invention can predict and analyze the energy utilization efficiency through virtual simulation technology, and reduce unnecessary energy waste. By accurately controlling parameters such as air flow velocity, temperature and humidity, it ensures that the tobacco is dried under optimal conditions, avoids excessive heating or humidification, and reduces energy consumption. At the same time, simulation technology reduces the need for testing and adjustment in production, reduces the wear cost of raw materials and equipment, and achieves energy conservation, environmental protection and cost optimization.

[0149] In one embodiment of the present invention, the virtual simulation system adopts a finite element analysis (FEA) model to improve the accuracy and computational efficiency of the simulation results, and the finite element analysis model performs fine grid division on the airflow, heat and humidity in different areas of the cavity. Further, the virtual simulation system is optimized based on a machine learning algorithm, and by inputting historical production data and simulation results, training a prediction model, and automatically learning the optimal process parameter combination to estimate the optimal process parameter combination, thereby reducing the cost of testing and adjustment, and improving production efficiency and product quality. For example, in future production processes, the virtual simulation system of the present invention can compare the real-time data with the historical model, predict the optimal process adjustment, and automatically apply it to the production control system, further improving the intelligent level of process optimization. The present invention combines the finite element analysis (FEA) and machine learning technology through a virtual simulation system, which can not only simulate the production effect under different process parameters, but also optimize the simulation model using historical production data to improve prediction accuracy and optimization effect.

[0150] Furthermore, the virtual simulation system of the present invention can be combined with actual production equipment to achieve feedback control. The simulation results output by the virtual simulation system are used to optimize the tobacco drying process in actual production to adjust the fan speed, heater power and humidifier workload in real time. The present invention automatically learns the optimal process parameter combination through a machine learning algorithm, can provide intelligent production suggestions for operators, reduce errors caused by manual adjustments, realize intelligent production control, and realize simulation and data-driven intelligent process optimization. The virtual simulation system of the present invention can be combined with the actual production process. Through the three-dimensional flow rate data and environmental parameters (temperature, humidity) collected in real time, the simulation system can generate optimized feedback signals, adjust the fan speed, heater power and humidification equipment workload in real time, ensure that the tobacco drying effect in the actual production process is optimal, and realize the automated optimization and adjustment of the production process. This feedback control greatly improves the stability and efficiency of the tobacco drying process, ensures the uniformity and quality consistency of tobacco drying, and realizes real-time feedback and production control optimization. For example, when it is detected that the tobacco flow rate in a certain area is too fast or the temperature is too high, the virtual simulation system can automatically adjust the parameters of the fan or heater to balance the airflow and heat distribution to ensure a more uniform drying process for the tobacco.

[0151] Furthermore, the simulation data generated by the virtual simulation system is associated with the production batch data and stored in a database for building a quality traceability system to achieve continuous monitoring and improvement of process quality.

[0152] Furthermore, the virtual simulation system displays the dynamic changes of the flow rate, temperature and humidity of the tobacco shreds through a three-dimensional visualization interface, and has a real-time visualization function, so that the operator can intuitively observe the influence of different process parameters on the drying process, and provide decision support. The virtual simulation system of the present invention has a real-time three-dimensional visualization function, which can intuitively display the dynamic changes of the flow rate, temperature and humidity of the tobacco shreds in a virtual environment. The user can intuitively observe the dynamic changes of the flow rate, temperature and humidity of the tobacco shreds through the visualization interface, helping the operator to monitor the drying process in real time, grasp the changes of process parameters in the drying process in real time, and provide decision support, make decision adjustments in time, and realize visualization and decision support: This visualization effect enables the operator to quickly identify the problem area, adjust the process parameters in time, and ensure the stability and efficiency of the production process. The decision support function of the virtual simulation system can generate operation suggestions based on the simulation data, and provide optimization solutions to help operators optimize the key parameters in the production process and achieve precise control.

[0153] The present invention significantly improves the automation and intelligence level of the tobacco drying process. It not only realizes the accurate measurement of the three-dimensional flow rate of tobacco, but also optimizes the process control through virtual simulation and machine learning technology, achieving the purpose of energy saving and consumption reduction, improving production efficiency and stabilizing product quality. At the same time, the virtual simulation system provides real-time feedback control and visualization functions, providing important technical support for the intelligent production and process optimization of the tobacco industry.

[0154] The virtual simulation method for the tobacco drying process provided in the embodiment of the present invention provides a virtual simulation solution based on the multi-angle optical flow method, which simulates and optimizes the tobacco drying effect under different process parameters by accurately measuring the three-dimensional flow velocity distribution of tobacco in the tobacco drying chamber and combining the heat transfer and moisture transfer models. It can not only simulate the tobacco flow and drying process under different conditions, but also automatically optimize the process parameters through machine learning algorithms, provide an efficient, energy-saving and intelligent production solution, and greatly improve the automation level of tobacco production and the consistency of product quality.

[0155] So far, various embodiments of the present disclosure have been described in detail. In order to avoid obscuring the concept of the present disclosure, some details known in the art are not described. Based on the above description, those skilled in the art can fully understand how to implement the technical solution disclosed here.

[0156] Although some specific embodiments of the present disclosure have been described in detail by way of examples, it should be understood by those skilled in the art that the above examples are for illustration only and are not intended to limit the scope of the present disclosure. It should be understood by those skilled in the art that the above embodiments may be modified or some technical features may be replaced by equivalents without departing from the scope and spirit of the present disclosure. The scope of the present disclosure is defined by the appended claims.

Claims

1. A virtual simulation method for a torrefied drying process, characterized in that: include: Arrange multiple high-speed camera devices at multiple positions of the tobacco drying chamber to synchronously capture image sequences of flowing tobacco from different angles; Calculating the velocity vectors of the tobacco at multiple angles according to the image sequences captured by the high-speed camera devices; Perform data fusion on velocity vectors at multiple angles to obtain three-dimensional velocity distribution data of tobacco; Based on the three-dimensional flow velocity distribution data, a movement model of the tobacco in the drying chamber is established, and a heat transfer model and a moisture transfer model are constructed in combination with the temperature data and the humidity data to simulate the drying state of the tobacco under different process conditions; Different process parameters are input into the virtual simulation system, and the virtual simulation system is run to simulate the state changes of tobacco under different process parameters, wherein the virtual simulation system integrates three-dimensional flow velocity distribution data of tobacco, dynamic changes of temperature field obtained based on the heat transfer model, and dynamic changes of humidity field obtained based on the moisture transfer model.

2. The virtual simulation method for the tofu drying process according to claim 1, characterized in that: The calculating the velocity vector of the tobacco at multiple angles according to the image sequences captured by each high-speed camera device specifically includes: A synchronous trigger system is used to trigger all high-speed cameras to start shooting at the same time and capture image sequences at the same frame rate; The multi-angle image sequences acquired by each of the high-speed camera devices are input into a computing unit, wherein the computing unit is configured with an image processing module based on an optical flow method, so as to process the image sequences through the image processing module to calculate the velocity vectors of the tobacco at multiple angles.

3. The virtual simulation method for the tofu drying process according to claim 2 is characterized in that: The multi-angle image sequence acquired by each high-speed camera device is input into a calculation unit, wherein the calculation unit is configured with an image processing module based on an optical flow method, so as to process the image sequence through the image processing module to calculate the velocity vector of the tobacco at multiple angles, specifically including: Preprocessing the image series, wherein the preprocessing includes at least one of grayscale processing, denoising processing and enhancement processing; Feature point detection and tracking: Through the feature point detection algorithm, the key feature points of the tobacco are detected in the preprocessed image; The key feature points are tracked in the image of each perspective, and the position changes of the key feature points in adjacent frames are calculated to obtain the velocity information of the key feature points.

4. The virtual simulation method for the tofu drying process according to claim 1, characterized in that: The data fusion of the velocity vectors at multiple angles to obtain the three-dimensional flow velocity distribution data of the tobacco shreds specifically includes: The three-dimensional flow velocity distribution data is obtained by calculating the optical flow algorithm of multi-angle images, wherein the optical flow algorithm uses the Lucas-Kanade method to track the characteristic points of the tobacco; The three-dimensional flow velocity distribution data is fused from multiple perspectives.

5. The virtual simulation method for the tofu drying process according to claim 4 is characterized in that: The three-dimensional velocity distribution data is obtained by calculating the optical flow algorithm of the multi-angle image, wherein the optical flow algorithm uses the Lucas-Kanade method to track the characteristic points of the tobacco, specifically including: The optical flow method is used to analyze the brightness changes of the image and calculate the velocity vector of the key feature points on the two-dimensional plane. The optical flow calculation is based on the following basic optical flow equation: I(x,y,t)=I(x+Δx,y+Δy,t+Δt) After linearization, the optical flow constraint equation is obtained: Among them, u represents the component of the optical flow velocity vector in the x direction, and v represents the component of the optical flow velocity vector in the y direction; The Lucas-Kanade algorithm is as follows: In a local window of the image, the optical flow constraint equations of multiple pixels are used to solve the velocity vector Among them, A represents the image gradient matrix: b represents the negative vector of the time gradient: The multi-view fusing of the three-dimensional velocity distribution data specifically includes: performing multi-view fusing by weighted averaging method or Kalman filtering method, Among them, the weighted average method for multi-view fusion specifically includes: Triangulation: Reconstruct the position of a 3D point P = (X, Y, Z) through velocity vectors from multiple perspectives, using the projection matrix P of two or more perspectives i and P j , solve the following minimization problem to reconstruct the 3D points: Among them, P i,3 and P j,3 Respectively represent the third row of the projection matrix; Calculation of three-dimensional velocity vector: According to the three-dimensional position change ΔP of the three-dimensional point P in consecutive frames, the velocity vector is calculated by the following formula: Multi-view velocity fusion: Assume that the velocity vectors obtained from multiple viewpoints are and These vectors are combined to get the three-dimensional flow velocity: Among them, w i and w j Represents the weight of the perspective.

6. The virtual simulation method for the tofu drying process according to claim 1, characterized in that: Based on the three-dimensional flow velocity distribution data, a movement model of tobacco in the drying chamber is established, and a heat transfer model and a moisture transfer model are constructed in combination with the temperature data and the humidity data to simulate the drying state of tobacco under different process conditions, specifically including: Arrange temperature sensors and humidity sensors in the drying chamber; The temperature data and humidity data are collected in real time through the temperature sensor and humidity sensor respectively; According to the collected temperature data, humidity data and three-dimensional flow velocity distribution data, a heat transfer model of the tofu drying chamber is established to simulate the heat transfer in the chamber. The heat transfer model follows Fourier's heat conduction law: Where q represents the heat flux, k represents the thermal conductivity, represents the temperature gradient; Based on the collected temperature data, humidity data and three-dimensional flow velocity distribution data, a moisture transfer model is established to evaluate the influence of humidity and air flow velocity on the drying rate of tobacco. The moisture transfer model follows Fick's diffusion law and is used to simulate the moisture transfer and evaporation process in tobacco: Where J represents the water diffusion flux, D represents the diffusion coefficient, and C represents the water concentration.

7. The virtual simulation method for the tofu drying process according to claim 1, characterized in that: The step of inputting different process parameters into the virtual simulation system, running the virtual simulation system, and simulating the state change of tobacco under different process parameters specifically includes: Different combinations of process parameters are input into the virtual simulation system, and the virtual simulation system outputs simulation results to demonstrate the tobacco drying effects under different conditions.

8. The virtual simulation method for the tofu drying process according to claim 7, characterized in that: The process parameter combination includes at least one of air flow velocity, temperature and humidity, fan speed, and heater power, and the simulation results include tobacco flow velocity distribution diagrams, tobacco drying rate curves, and dynamic changes of temperature fields and humidity fields corresponding to different process parameter combinations.

9. The virtual simulation method for the tofu drying process according to claim 1, characterized in that: The virtual simulation system adopts a finite element analysis model, which performs fine grid division on the airflow, heat and humidity in different areas of the cavity; the virtual simulation system is optimized based on a machine learning algorithm, and the prediction model is trained by inputting historical production data and simulation results to estimate the optimal combination of process parameters.

10. The virtual simulation method for the tofu drying process according to claim 1, characterized in that: The simulation results output by the virtual simulation system are used to optimize the tofu drying process in actual production; the simulation data generated by the virtual simulation system is associated with the production batch data and stored in a database for building a quality traceability system; The virtual simulation system displays the dynamic changes of flow rate, temperature and humidity of the tobacco shreds through a three-dimensional visualization interface.

Images