Rice mill intelligent control method, device and equipment and storage medium

Abstract

The invention discloses an intelligent control method, device and equipment for a rice mill and a storage medium, and relates to the technical field of grain processing.The method comprises the steps that a digital twinning system of the rice mill is constructed, the digital twinning system comprises a physical layer, a virtual layer and an intelligent service layer, and the virtual layer comprises a simulation model based on abrasive belt-rice grain contact dynamics; acquiring operation data of the rice mill through a physical layer, and performing space-time alignment on the operation data and the simulation model to obtain target data; predicting the change trend of the broken rice rate in the rice milling process based on the target data and the simulation model, and generating an initial parameter combination of the abrasive belt pressure, the rotating speed and the gap of the whitening chamber; iteratively updating the initial parameter combination through a reinforcement learning algorithm of an intelligent service layer to obtain a target parameter combination; and adjusting an execution mechanism of the rice mill according to the target parameter combination to complete intelligent control of the rice mill. According to the rice milling device, the broken rice rate in the rice milling process can be reduced, and intelligent control is achieved.

Description

Technical Field

[0001] This application relates to the technical field of grain processing, and particularly relates to an intelligent control method, device, equipment and storage medium for a rice milling machine. Background Art

[0002] With the continuous development of grain processing technology, the rice milling machine, as the main equipment for paddy processing, plays a key role in ensuring grain supply and maintaining food security. Food security is an important foundation for the national economy and people's livelihood, and reducing grain loss is an important way to ensure food security.

[0003] Currently, traditional rice milling machines mainly remove the surface bran layer through the rigid collision between brown rice and the rice milling roller and rice sieve, as well as the collision and friction between brown rice grains. Although this processing method can achieve the basic rice milling function, there are the following main problems in the processing process: First, the broken rice rate is high. Especially for indica rice varieties with an aspect ratio greater than 3, such as Jianghan rice, the broken rice rate is as high as 20% - 30%, resulting in a large amount of grain waste; Second, the nutrient loss is serious during the processing. The rice grains are prone to nutrient loss during the violent collision; Third, the energy consumption is high and the efficiency is low. The processing efficiency of traditional rice milling machines is difficult to meet the needs of large-scale production; Fourth, there is an environmental pollution problem. Traditional rice milling machines will generate more pollutants such as dust during the processing; Fifth, there is a lack of intelligent control. The parameter adjustment of traditional rice milling machines depends on manual experience and cannot achieve precise control and real-time optimization. Therefore, how to reduce the broken rice rate during the rice milling process and achieve intelligent control has become an urgent problem to be solved.

[0004] The above content is only used to assist in understanding the technical solution of this application, and does not represent an admission that the above content is prior art. Summary of the Invention

[0005] The purpose of this application is to provide an invention name, aiming to solve the technical problem of how to reduce the broken rice rate during the rice milling process and achieve intelligent control.

[0006] To achieve the above purpose, this application proposes an intelligent control method for a rice milling machine, and the method includes:

[0007] Construct a digital twin system of the rice milling machine, where the digital twin system includes a physical layer, a virtual layer, and an intelligent service layer, and the virtual layer includes a simulation model based on the contact dynamics of the abrasive belt - rice grains;

[0008] Collect the operation data of the rice milling machine through the physical layer, and perform spatio-temporal alignment on the operation data and the simulation model to obtain target data;

[0009] Based on the target data and the simulation model, predict the change trend of the broken rice rate during the rice milling process, and generate an initial parameter combination of the abrasive belt pressure, rotation speed, and milling chamber clearance;

[0010] Iteratively update the initial parameter combination through the reinforcement learning algorithm of the intelligent service layer to obtain the target parameter combination;

[0011] Adjust the actuator of the rice milling machine according to the target parameter combination to complete the intelligent control of the rice milling machine.

[0012] In addition, to achieve the above object, the present application also proposes an intelligent control device for a rice milling machine, the device includes:

[0013] A digital twin construction module for constructing a digital twin system of the rice milling machine, the digital twin system includes a physical layer, a virtual layer and an intelligent service layer, and the virtual layer includes a simulation model based on the contact dynamics of the abrasive belt-rice grains;

[0014] A data fusion module for collecting the operation data of the rice milling machine through the physical layer, and performing spatio-temporal alignment on the operation data and the simulation model to obtain target data;

[0015] A prediction module for predicting the change trend of the broken rice rate during the rice milling process based on the target data and the simulation model, and generating an initial parameter combination of the abrasive belt pressure, rotational speed and the gap of the whitening chamber;

[0016] A parameter optimization module for iteratively updating the initial parameter combination through the reinforcement learning algorithm of the intelligent service layer to obtain the target parameter combination;

[0017] A control module for adjusting the actuator of the rice milling machine according to the target parameter combination to complete the intelligent control of the rice milling machine.

[0018] In addition, to achieve the above object, the present application also proposes an intelligent control device for a rice milling machine, the device includes: a memory, a processor, and a computer program stored on the memory and executable on the processor, and the computer program is configured to implement the steps of the rice milling machine intelligent control method as described above.

[0019] In addition, to achieve the above object, the present application also proposes a storage medium, the storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium, and when the computer program is executed by a processor, it implements the steps of the rice milling machine intelligent control method as described above.

[0020] One or more technical solutions proposed by the present application have at least the following technical effects:

[0021] First, build a digital twin system for the rice milling machine. This system includes a physical layer, a virtual layer, and an intelligent service layer. The virtual layer contains a simulation model based on the contact dynamics between the abrasive belt and the rice grains. In this way, the operating state of the rice milling machine can be monitored in real time, providing a theoretical basis and infrastructure for subsequent optimization. Next, use the sensors in the physical layer to collect the operating data of the rice milling machine, and align these data with the simulation model in space and time to obtain target data. The space-time alignment ensures the accuracy and reliability of the data, providing high-quality input for subsequent prediction and optimization. Then, based on the target data and the simulation model, predict the change trend of the broken rice rate during the rice milling process, and generate an initial parameter combination of the abrasive belt pressure, rotation speed, and gap of the whitening chamber accordingly. The initial parameter combination provides a reasonable starting point for optimization, reduces the number of blind adjustments, and optimizes the parameters in advance to avoid the generation of broken rice. After that, use the reinforcement learning algorithm in the intelligent service layer to iteratively update the initial parameter combination to obtain the target parameter combination. The reinforcement learning algorithm can automatically optimize the parameter combination to adapt to different operating conditions, significantly reduce the broken rice rate, and improve the whitening uniformity and production efficiency. Finally, adjust the actuator of the rice milling machine according to the target parameter combination to complete the intelligent control. The precise adjustment of the actuator ensures that the operating parameters of the rice milling machine reach the optimal state, and the closed-loop feedback mechanism adjusts the parameters in real time to adapt to the dynamic environment, ultimately achieving the optimization goals of low broken rice rate, high whitening uniformity, and low energy consumption, and improving the performance and economic benefits of the rice milling machine. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on the structures shown in these drawings.

[0023] Figure 1 It is a schematic flowchart provided in Embodiment 1 of the intelligent control method for the rice milling machine of the present application;

[0024] Figure 2 It is a schematic flowchart provided in Embodiment 2 of the intelligent control method for the rice milling machine of the present application;

[0025] Figure 3 and Figure 4 It is a schematic structural diagram of an embodiment of the rice milling machine provided by the present application;

[0026] Figure 5 is Figure 3 a schematic structural diagram of the adjusting mechanism and the rice milling mechanism in

[0027] Figure 6 is Figure 3 a schematic structural diagram of the frame and the rice milling mechanism in

[0028] Figure 7 is Figure 3 a schematic structural diagram of another embodiment of the rice milling component in

[0029] Figure 8 is Figure 3 a schematic structural diagram of the feeding device in

[0030] Figure 9 is Figure 3 a schematic structural diagram of the blowing device in

[0031] Figure 10 is Figure 3 a schematic structural diagram of the detection mechanism in

[0032] Figure 11 is Figure 3 a schematic structural diagram of the transfer material mechanism in the first material guiding position;

[0033] Figure 12 is Figure 3 a schematic structural diagram of the transfer material mechanism in the second material guiding position.

[0034] Explanation of the reference numerals in the drawings:

[0035] 100, rice milling machine; 1, frame; 2, feeding device; 21, feeding hopper; 211, side wall; 22, feeding rotating shaft; 23, adjusting plate; 3, rice milling component; 3a, feeding end; 3b, discharging end; 30, rice milling mechanism; 31, rice milling part; 31a, rice groove; 32, driving device; 33, first driving wheel; 34, second driving wheel; 35, transmission belt; 36, abrasive belt; 4, bran removing mechanism; 41, blowing device; 42, suction device; 5, adjusting mechanism; 51, sliding table module; 52, driving mechanism; 521, hand wheel; 6, detection mechanism; 61, conveying component; 61a, first end; 61b, second end; a, evacuation station; b, detection station; 62, evacuation teeth; 63, image acquisition device; 64, light shield; 7, transfer material mechanism; 71, guiding plate; 72, transfer hopper; 73, material guiding plate.

[0036] The realization of the purpose, functional features and advantages of this application will be further described with reference to the embodiments and the accompanying drawings. Specific embodiments

[0037] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of this application and are not used to limit this application.

[0038] In order to better understand the technical solutions of this application, the following will be described in detail with reference to the drawings in the specification and specific embodiments.

[0039] It should be noted that the execution subject of the embodiments of the present application can be a computing service device with data processing, network communication, and program running functions, such as a tablet computer, a personal computer, a mobile phone, etc., or an electronic device capable of implementing the above functions, an intelligent control system of a sand belt rice mill, etc. Hereinafter, an intelligent control system of a sand belt rice mill will be taken as an example to illustrate this embodiment and the following embodiments.

[0040] Based on this, the embodiments of the present application provide an intelligent control method for a rice mill, referring to Figure 3 , Figure 3 which is a schematic flowchart of the first embodiment of the intelligent control method for the rice mill of the present application.

[0041] In this embodiment, the intelligent control method for the rice mill includes steps S10 to S50:

[0042] Step S10, constructing a digital twin system for the rice mill, where the digital twin system includes a physical layer, a virtual layer, and an intelligent service layer, and the virtual layer includes a simulation model based on the contact dynamics between the sand belt and the rice grains.

[0043] It should be noted that the digital twin system is a technical architecture that uses data such as physical models, sensor updates, and operation history to integrate a simulation process of multiple disciplines, multiple physical quantities, multiple scales, and multiple probabilities, and completes mapping in the virtual space. It can reflect the entire life cycle process of the corresponding physical entity and achieve real-time monitoring, prediction, and optimization of the physical entity.

[0044] The physical layer is the physical part of the digital twin system, mainly responsible for data acquisition and the actual operation of the device, including the following: (1) Multimodal sensors: Deployed on the rice mill to monitor the operation status of the device in real time, including: force sensors, used to monitor mechanical parameters such as sand belt tension and pressure distribution in the whitening chamber; vibration sensors, used to detect the wear status of the main shaft bearing, and the sampling frequency needs to reach more than 10 kHz; vision sensors (industrial cameras), integrated with a machine vision system, used to obtain the morphological data of the rice grains in real time, such as the integrity of the rice grains, crack area, whiteness degree, etc. (2) Data acquisition and transmission: Using an industrial Internet of Things gateway that supports the OPC UA protocol to achieve synchronous acquisition and transmission of multi-source data. The timestamp alignment accuracy of data acquisition needs to be less than 1 ms to ensure the accuracy and real-time nature of the data.

[0045] The virtual layer is the virtual part of the digital twin system, mainly responsible for the modeling and simulation of physical entities, including the following: (1) High-precision parametric model: A virtual / simulation model of the abrasive belt rice mill constructed based on SolidWorks / ANSYS, including a mechanical structure model and a process model. (2) Mechanical structure model: Used to simulate the contact dynamics between the abrasive belt and rice grains, and analyze the movement and force conditions of rice grains in the rice milling chamber. (3) Process model: Used to describe the relationship between the degree of rice whitening, pressure, and rotational speed, and predict the rice milling effect through an equation model. (4) Bidirectional data drive: Realize the bidirectional data drive between the virtual model and the physical device using Unity3D / TwinCAT. The simulation step size needs to be controlled within 10 ms to ensure that the virtual model can reflect the operating state of the physical device in real time.

[0046] The intelligent service layer is the intelligent part of the digital twin system, mainly responsible for data analysis, decision support, and optimization control, including the following: (1) Digital twin database: Used to store all-dimensional data of the device, including operating parameters, maintenance records, fault cases, etc. Supports SQL / NoSQL hybrid queries for quick retrieval and analysis of data. (2) Intelligent decision-making module: Parameter optimization, based on the reinforcement learning (RL) algorithm, dynamically adjusts the abrasive belt pressure and rotational speed to optimize the rice milling effect. Fault prediction, uses the LSTM neural network to predict the remaining useful life (RUL) of the bearing and gives early warnings of equipment failures. Energy consumption management, constructs a multi-objective optimization model to balance processing efficiency and unit energy consumption and achieve energy-saving optimization.

[0047] The simulation model is the core part of the virtual layer, used to simulate the behavior and performance of physical entities. The simulation model is constructed based on the contact dynamics between the abrasive belt and rice grains, specifically including the following: (1) Abrasive belt-rice grain contact dynamics model: Through mechanical analysis, simulate the contact, friction, and collision behaviors between the abrasive belt and rice grains. This model can predict the force conditions and movement trajectories of rice grains during the rice milling process, providing theoretical support for optimizing the rice milling process. (2) Process model: Describes the relationship between the degree of rice whitening, pressure, and rotational speed, and predicts the rice milling effect through an equation model. This model can predict indicators such as the whiteness and broken rice rate after rice milling based on the input process parameters (such as abrasive belt pressure, rotational speed, etc.), providing data support for the intelligent decision-making module.

[0048] It is understandable that, first of all, the intelligent control system of the abrasive belt rice mill will use software such as SolidWorks and ANSYS to build a high-precision virtual model of the rice mill, including mechanical structure and process models, and realize two-way data interaction between the virtual model and the physical device through Unity3D and TwinCAT, ensuring that the virtual model can reflect the operating state of the physical device in real time, and optimizing the rice milling process parameters through simulation analysis. Then, the intelligent control system will establish a digital twin database to store all-dimensional data of the device operation, and realize functions such as parameter optimization, fault prediction and energy consumption management through reinforcement learning algorithms and LSTM neural networks to complete the construction of the digital twin system.

[0049] Step S20, collect the operation data of the rice mill through the physical layer, and perform spatio-temporal alignment on the operation data and the simulation model to obtain target data.

[0050] It should be noted that the operation data refers to various data related to the device operation state collected by sensors in the physical layer during the actual operation of the abrasive belt rice mill, including but not limited to: (1) Force sensor data: used to monitor the abrasive belt tension and the pressure distribution in the whitening chamber. These data can reflect the contact force between the abrasive belt and the brown rice during the rice milling process, as well as the pressure change in the whitening chamber. (2) Vibration sensor data: used to detect the vibration of the main shaft bearing. The vibration data can reflect the wear state of the bearing and the operating stability of the device. The sampling frequency needs to reach more than 10 kHz to capture high-frequency vibration signals. (3) Visual sensor data: The rice grain morphology data collected by the machine vision system, including the integrity of the rice grains, the crack area, the degree of whitening, etc. These data can intuitively reflect the rice milling effect and help evaluate the broken rice rate and the uniformity of whitening. (4) Other operating parameters: such as the rotation speed of the abrasive belt, the filling amount of brown rice, the gap of the rice milling chamber, etc. These parameters are crucial for understanding the physical behavior during the rice milling process and optimizing the process parameters.

[0051] Target data refers to the processed and aligned operation data, which is completely consistent with the virtual data in the simulation model in terms of time and space and can be directly used for the calibration, verification, and optimization of the simulation model. Target data includes: (1) Sensor data aligned in time and space: Through the time-space alignment algorithm, the operation data is matched with the timestamps and spatial positions in the simulation model. For example, the sand belt tension data collected by the force sensor is aligned with the sand belt tension at the corresponding time and position in the simulation model to ensure their consistency in time and space. (2) Multi-source data after fusion: Algorithms such as the Kalman filter are used to fuse the data from different sensors, eliminate the image jitter error caused by equipment vibration, and improve the accuracy and reliability of the data. For example, the rice grain morphology data collected by the vision sensor is fused with the force sensor data to obtain more comprehensive information on the rice milling process. (3) Low-dimensional feature data: Dimensionality reduction algorithms such as t-SNE are applied to map the high-dimensional operation data (such as vibration spectra) to a low-dimensional space and extract key fault features. These low-dimensional feature data can more intuitively reflect the operating state of the equipment, facilitating subsequent fault diagnosis and prediction.

[0052] It can be understood that, first, the intelligent control system of the sand belt rice mill collects real-time data during the operation of the rice mill through sensors in the physical layer, including key parameters such as sand belt tension, pressure in the whitening chamber, and rice grain morphology. Secondly, the system aligns the collected operation data with the simulation model in the virtual layer through the matching of timestamps and spatial positions to ensure the consistency of the operation data and the simulation model in time and space.

[0053] As an example, the physical layer is equipped with a three-axis force sensor and an industrial camera. The steps of aligning the operation data with the simulation model in time and space to obtain target data include: performing Butterworth low-pass filtering on the tension data collected by the three-axis force sensor to obtain the filtered tension data; aligning the rice grain images collected by the industrial camera with the filtered tension data in terms of timestamps to obtain a synchronized data packet; and mapping the rice grain position coordinates in the synchronized data packet to the three-dimensional coordinate system of the simulation model through a feature point matching algorithm to obtain target data.

[0054] Tension data refers to the numerical information about the tension exerted on the sand belt during operation collected by the three-axis force sensor, which reflects the force condition when the sand belt contacts the paddy during the rice milling process.

[0055] The filtered tension data refers to the tension data after Butterworth low-pass filtering. The noise is removed through low-pass filtering to make the data smoother, thereby more accurately reflecting the actual force condition of the sand belt.

[0056] A synchronous data packet refers to a data set obtained by aligning the timestamp of the rice grain image collected by an industrial camera with the filtered tension data. Timestamp alignment means ensuring the temporal consistency between the image data and the tension data, so that the shape of the rice grains in the image can be associated with the corresponding tension data.

[0057] The rice grain position coordinates refer to the two-dimensional coordinate information of the position where the rice grains are located in the rice grain image collected by the industrial camera, which can be extracted through image processing algorithms and used to describe the specific position of the rice grains in the image.

[0058] A three-dimensional coordinate system refers to the coordinate system used to describe the spatial position of an object in the simulation model, and is used in this embodiment to represent the actual spatial position of the rice grains in the rice milling machine. Through the feature point matching algorithm, the two-dimensional position coordinates in the rice grain image are mapped into this three-dimensional coordinate system, and the accurate position of the rice grains in the simulation model can be obtained, thereby realizing the spatio-temporal alignment of the image data and the simulation model.

[0059] First, perform Butterworth low-pass filtering on the abrasive belt tension data collected by the triaxial force sensor to remove the high-frequency noise in the data and retain the stable low-frequency signal, so as to obtain smoother and more accurate filtered tension data; then, accurately align the rice grain image collected by the industrial camera with the filtered tension data according to the timestamp to ensure the temporal consistency between the image data and the tension data, and form a synchronous data packet containing both pieces of information; finally, use the feature point matching algorithm to match the two-dimensional position coordinates of the rice grain image in the synchronous data packet with the three-dimensional coordinate system of the simulation model, map the actual position of the rice grains into the three-dimensional space of the simulation model, and obtain the target data that can be used for subsequent analysis and optimization.

[0060] [[ID=X]]Step S30, based on the target data and the simulation model, predict the changing trend of the broken rice rate during the rice milling process, and generate an initial parameter combination of the abrasive belt pressure, rotational speed, and milling chamber clearance.

[0061] It should be noted that the changing trend of the broken rice rate refers to the law of how the broken rice rate (i.e., the ratio of the broken rice mass to the mass of the intact rice grains before milling) changes during the rice milling process as process parameters such as the abrasive belt pressure, rotational speed, and milling chamber clearance change. Through the analysis of the target data and the simulation model, the rising or falling trend of the broken rice rate under different parameter settings can be predicted. The abrasive belt pressure refers to the magnitude of the force applied when the abrasive belt contacts the paddy during the rice milling process. It directly affects the frictional force between the paddy and the abrasive belt, and thus affects the effect of removing the bran layer and the breaking condition of the rice grains.

[0062] The rotational speed refers to the number of revolutions per minute of the abrasive belt during the rice milling process, usually expressed in meters per minute (m / min). The rotational speed determines the movement speed of the brown rice in the rice milling chamber and the contact frequency with the abrasive belt. A higher rotational speed may increase the risk of rice grain breakage, but can also improve the efficiency of bran layer removal; while a lower rotational speed may reduce the breakage rate, but may require a longer time to complete the whitening process.

[0063] The whitening chamber gap refers to the distance between the abrasive belt and the paddy turning device (i.e., the area where the brown rice passes through). The size of this gap directly affects the movement trajectory and force condition of the brown rice in the rice milling chamber. A larger gap may cause the brown rice to not be in close contact with the abrasive belt, resulting in incomplete bran layer removal; while a smaller gap may increase the risk of rice grain breakage. The initial parameter combination refers to a set of initial setting values of the abrasive belt pressure, rotational speed, and whitening chamber gap generated based on the target data and the results predicted by the simulation model before optimizing the rice milling process. This set of parameters is obtained based on the analysis and prediction of the change trend of the broken rice rate, aiming to provide a reasonable starting point for subsequent process optimization.

[0064] It can be understood that, first, the intelligent control system of the abrasive belt rice milling machine inputs the target data into the simulation model. The simulation model analyzes and simulates based on the target data, and predicts how the broken rice rate will change under different conditions, that is, the change trend of the broken rice rate. Then, the system selects a set of values of the abrasive belt pressure, rotational speed, and whitening chamber gap that can make the broken rice rate relatively low and meet the production requirements based on the predicted change trend of the broken rice rate and the requirements for the broken rice rate in production, as the initial parameter combination.

[0065] Step S40, iteratively update the initial parameter combination through the reinforcement learning algorithm of the intelligent service layer to obtain the target parameter combination.

[0066] It should be noted that the target parameter combination refers to the parameter combination of the optimal abrasive belt pressure, rotational speed, and whitening chamber gap obtained through optimization by the reinforcement learning algorithm. This set of parameters can minimize the broken rice rate, improve the whitening uniformity, and reduce energy consumption on the premise of meeting the production requirements.

[0067] It can be understood that, first, the intelligent control system of the abrasive belt rice mill applies the initial parameter combination to the actual rice milling process, and collects operation data and rice grain quality data (such as broken rice rate, whiteness uniformity, etc.) through sensors. Then, based on these data, the reinforcement learning algorithm in the intelligent service layer evaluates the performance of the current parameter combination according to the preset reward function (such as the comprehensive score of goals like reducing the broken rice rate, improving whiteness uniformity, and reducing energy consumption). According to the evaluation results, the algorithm gradually adjusts parameters such as abrasive belt pressure, rotation speed, and whitening chamber gap by exploring new parameter combinations and exploiting known better parameter combinations to optimize the value of the reward function. This process is continuously repeated, and the parameter combination is updated according to the new evaluation results in each iteration until an optimal parameter combination, that is, the target parameter combination, is found.

[0068] Step S50, adjust the actuator of the rice mill according to the target parameter combination to complete the intelligent control of the rice mill.

[0069] It should be noted that the actuator refers to those mechanical or electrical devices that can directly adjust and control the physical parameters of the rice mill (such as abrasive belt pressure, rotation speed, whitening chamber gap, etc.) according to the instructions issued by the intelligent control system.

[0070] It can be understood that, first, the intelligent control system of the abrasive belt rice mill sends execution instructions to each actuator according to the target parameter combination obtained by the reinforcement learning algorithm. Then, these actuators work together according to the execution instructions to adjust the operating parameters of the rice mill to the target state, thereby achieving precise intelligent control of the rice milling process, effectively reducing the broken rice rate and improving the rice milling quality.

[0071] As an example, the actuator includes a whitening chamber gap adjustment mechanism, an abrasive belt drive motor, and a ball screw slide. The step of adjusting the actuator of the rice mill according to the target parameter combination to complete the intelligent control of the rice mill includes: converting the target abrasive belt pressure value in the target parameter combination into a pulse width modulation signal; controlling the whitening chamber gap adjustment mechanism according to the pulse width modulation signal; adjusting the input voltage of the abrasive belt drive motor based on the target rotation speed value in the target parameter combination through a fuzzy proportional-integral-derivative controller; adjusting the ball screw slide according to the target gap value in the target parameter combination to make the whitening chamber gap reach the target gap value, and completing the intelligent control of the rice mill.

[0072] The whitening chamber gap adjustment mechanism refers to a mechanical device used to adjust the distance between the abrasive belt and the material turning device. By adjusting the whitening chamber gap, the movement trajectory and force condition of the brown rice in the rice milling chamber can be optimized, thereby improving the whiteness uniformity and reducing the broken rice rate.

[0073] The abrasive belt drive motor refers to the electric motor used to drive the operation of the abrasive belt. By adjusting the input voltage of the abrasive belt drive motor, the running speed of the abrasive belt can be changed. The ball screw slide is a high-precision mechanical transmission device used to achieve precise control of linear motion. In an abrasive belt rice mill, the ball screw slide is usually used to adjust the clearance of the whitening chamber. The target abrasive belt pressure value refers to the ideal value of the contact pressure between the abrasive belt and paddy rice determined according to the target parameter combination.

[0074] The pulse width modulation signal (PWM) is a form of signal that controls the power output by changing the pulse width of the signal. By adjusting the pulse width of the PWM signal, the output power of the motor or other actuators can be precisely controlled, thereby achieving the adjustment of the clearance of the whitening chamber.

[0075] The target rotational speed value refers to the ideal rotational speed of the abrasive belt drive motor determined according to the target parameter combination. The rotational speed of the abrasive belt directly affects the movement speed of paddy rice in the rice milling chamber and the contact frequency with the abrasive belt. An appropriate rotational speed can improve the removal efficiency of the bran layer while reducing the breakage of rice grains.

[0076] The target clearance value refers to the ideal clearance size of the whitening chamber determined according to the target parameter combination. The size of the clearance of the whitening chamber directly affects the movement trajectory and force condition of paddy rice in the rice milling chamber.

[0077] First, the intelligent control system of the abrasive belt rice mill converts the target abrasive belt pressure value in the target parameter combination into a corresponding pulse width modulation signal, and controls the clearance adjustment mechanism of the whitening chamber through this signal, so that the contact pressure between the abrasive belt and paddy rice reaches the target value, thereby optimizing the rice milling effect. Second, based on the target rotational speed value in the target parameter combination, the system uses a fuzzy proportional-integral-derivative controller to adjust the input voltage of the abrasive belt drive motor to ensure that the running speed of the abrasive belt precisely matches the target rotational speed to improve the whitening efficiency. Finally, the system adjusts the ball screw slide according to the target clearance value in the target parameter combination, and drives the ball screw to rotate through the motor, so that the clearance of the whitening chamber reaches the target value, further optimizing the rice milling process and finally completing the intelligent control of the abrasive belt rice mill.

[0078] This embodiment provides an intelligent control method for a rice milling machine. First, a digital twin system of the rice milling machine is constructed, which includes a physical layer, a virtual layer, and an intelligent service layer. The virtual layer includes a simulation model based on the contact dynamics between the abrasive belt and rice grains. In this way, the operating state of the rice milling machine can be monitored in real time, providing a theoretical basis and infrastructure for subsequent optimization. Next, the operating data of the rice milling machine is collected by sensors in the physical layer, and these data are aligned in space and time with the simulation model to obtain target data. The space-time alignment ensures the accuracy and reliability of the data, providing high-quality input for subsequent prediction and optimization. Then, based on the target data and the simulation model, the changing trend of the broken rice rate during the rice milling process is predicted, and an initial parameter combination of the abrasive belt pressure, rotation speed, and gap of the milling chamber is generated accordingly. The initial parameter combination provides a reasonable starting point for optimization, reduces the number of blind adjustments, and optimizes the parameters in advance to avoid the generation of broken rice. After that, the initial parameter combination is iteratively updated through the reinforcement learning algorithm in the intelligent service layer to obtain the target parameter combination. The reinforcement learning algorithm can automatically optimize the parameter combination to adapt to different operating conditions, significantly reduce the broken rice rate, and improve the milling uniformity and production efficiency. Finally, the actuator of the rice milling machine is adjusted according to the target parameter combination to complete the intelligent control. The precise adjustment of the actuator ensures that the operating parameters of the rice milling machine reach the optimal state, and the closed-loop feedback mechanism adjusts the parameters in real time to adapt to the dynamic environment, ultimately achieving the optimization goals of low broken rice rate, high milling uniformity, and low energy consumption, and improving the performance and economic benefits of the rice milling machine.

[0079] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar content as that in the above-mentioned first embodiment can be referred to the above introduction and will not be repeated hereinafter. On this basis, please refer to Figure 4 , Figure 4 which is a schematic flowchart of the second embodiment of the intelligent control method for the rice milling machine of the present application. The step S30 of the intelligent control method for the rice milling machine includes steps S31 to S34:

[0080] Step S31: Input the target data into the simulation model to obtain the stress distribution of the rice grain group in the milling chamber.

[0081] It should be noted that the stress distribution refers to the spatial distribution of the internal stress generated by the rice grain group under the action of various forces in the milling chamber, reflecting the changes in the magnitude and direction of the stress generated by the contact, friction, and collision of the rice grains with the abrasive belt, the turning device, and other rice grains during the rice milling process.

[0082] It can be understood that, first, the target data (including information such as the position coordinates of rice grains and the tension of the abrasive belt) after spatio-temporal alignment and feature point matching is input into the simulation model of the abrasive belt rice milling machine. The simulation model is based on the contact dynamics between the abrasive belt and rice grains and can simulate the movement trajectory of rice grains in the whitening chamber and their interactions with the abrasive belt, the turning device, and other rice grains. By calculating the forces generated by these interactions, the simulation model can accurately analyze the stress distribution of the rice grain group in the whitening chamber, including the magnitudes of normal stress and shear stress and their spatial distributions inside and on the surface of the rice grains.

[0083] Step S32: According to the stress distribution, count the number of rice grains with stress greater than the preset rice grain compressive strength threshold in each time window to obtain the broken rice probability data.

[0084] It should be noted that a time window refers to dividing the entire rice milling process into several continuous time periods with a certain length during the analysis process. Its setting is to decompose the complex continuous process into multiple shorter time periods to more carefully analyze the stress conditions of rice grains in different time periods.

[0085] The preset rice grain compressive strength threshold is a stress value preset according to the physical properties of rice grains and is used to determine whether a rice grain is likely to break. The compressive strength of a rice grain refers to the maximum stress that a rice grain can withstand. When the stress on the rice grain exceeds this threshold, the rice grain may break, generally 18 MPa.

[0086] The broken rice probability data refers to the proportion of the number of rice grains with stress greater than the preset compressive strength threshold in each time window to the total number of rice grains. This proportion reflects the possibility of rice grain breakage in that time window. By statistically analyzing the broken rice probability data in each time window, the change trend of the broken rice rate during the entire rice milling process can be obtained.

[0087] It can be understood that, first, the system divides the entire rice milling process into multiple continuous time periods according to the set time interval, and each time period is a time window. Secondly, the system respectively collects and analyzes the stress data of the rice grain group in each time window and counts the situation where the stress on the rice grains exceeds the preset compressive strength threshold. Finally, in this way, the system can dynamically monitor the stress conditions of rice grains in different time periods, thereby more accurately identifying specific time periods with a high broken rice rate and providing data support for subsequent optimization of rice milling process parameters to reduce the broken rice rate and improve the rice milling quality.

[0088] Step S33: Based on the broken rice probability data, construct a broken rice rate prediction surface, and obtain the initial safe parameter ranges of the abrasive belt pressure, rotational speed, and whitening chamber clearance according to the broken rice rate prediction surface.

[0089] It should be noted that the broken rice rate prediction surface refers to a mathematical model or graph constructed based on broken rice probability data, which can intuitively show the relationship between the broken rice rate and process parameters such as abrasive belt pressure, rotational speed, and whitening chamber clearance. Specifically, the broken rice rate prediction surface is a three-dimensional or multi-dimensional surface, where the abscissa and ordinate respectively represent different process parameters (such as abrasive belt pressure, rotational speed, whitening chamber clearance), and the points on the surface represent the corresponding predicted broken rice rate values under these parameter combinations.

[0090] It can be understood that, first, the system classifies and organizes the collected broken rice probability data according to different combinations of abrasive belt pressure, rotational speed, and whitening chamber clearance to form a data set. Second, the system uses this data set to construct a broken rice rate prediction surface through mathematical modeling or machine learning algorithms, which can intuitively show the relationship between the broken rice rate and each process parameter. Finally, based on the broken rice rate prediction surface, the system analyzes under which parameter combinations the broken rice rate is lower, thereby determining the initial safe parameter range of abrasive belt pressure, rotational speed, and whitening chamber clearance, providing a reliable starting point for subsequent process optimization, and ensuring that excessive broken rice will not be generated due to improper parameter settings during the optimization process.

[0091] As an example, the steps of constructing a broken rice rate prediction surface based on the broken rice probability data and obtaining the initial safe parameter range of abrasive belt pressure, rotational speed, and whitening chamber clearance according to the broken rice rate prediction surface include: fitting the non-linear relationship between abrasive belt pressure, rotational speed, and whitening chamber clearance and the broken rice rate through the Gaussian process regression algorithm according to the broken rice probability data to generate a broken rice rate prediction surface; delimiting a safety threshold boundary on the broken rice rate prediction surface according to a preset broken rice rate threshold and a preset standard deviation of milling uniformity to obtain the initial safe parameter range.

[0092] The non-linear relationship means that there is no simple linear proportional relationship between abrasive belt pressure, rotational speed, and whitening chamber clearance and the broken rice rate. In the actual rice milling process, the influence of these process parameters on the broken rice rate is complex, and there may be interactions and non-linear changes. For example, increasing the abrasive belt pressure may reduce the broken rice rate to a certain extent, but after exceeding a certain critical value, the broken rice rate may rise sharply.

[0093] The preset broken rice rate threshold refers to the upper limit value of the broken rice rate preset in advance according to production requirements and quality standards.

[0094] The preset standard deviation of milling uniformity refers to the upper limit value that the variation degree of milling uniformity should not exceed during the production process. The level of milling uniformity directly affects the appearance and quality of rice. The smaller the standard deviation, the more uniform the milling effect and the higher the quality of rice.

[0095] The safety threshold boundary refers to the parameter range boundary defined on the broken rice rate prediction surface according to the preset broken rice rate threshold and the preset standard deviation of milling uniformity. Within this boundary, the combination of abrasive belt pressure, rotational speed, and milling chamber clearance can ensure that the broken rice rate does not exceed 12%, and the standard deviation of milling uniformity is less than 0.5.

[0096] First, the system uses the Gaussian process regression algorithm to analyze the broken rice probability data, fits the non-linear relationship between the abrasive belt pressure, rotational speed, and milling chamber clearance and the broken rice rate, and generates the broken rice rate prediction surface. This step is to more accurately predict the broken rice rate under different parameter combinations. Then, on the prediction surface, the system defines the safety threshold boundary according to the preset broken rice rate threshold (such as 12%) and the preset standard deviation of milling uniformity (such as less than 0.5), that is, finds the parameter combination area that satisfies the broken rice rate not exceeding 12% and the standard deviation of milling uniformity less than 0.5, so as to determine the initial safety parameter range. This step is to ensure that when operating the rice milling machine within these parameter ranges, it can not only effectively reduce the broken rice rate but also ensure that the milling uniformity meets the quality standards. Finally, the system uses this initial safety parameter range as the starting point for subsequent optimization to ensure that excessive broken rice will not occur due to improper parameter settings during the optimization process.

[0097] In step S34, combined with the fault countermeasure knowledge graph, screen the parameter combinations that meet the milling uniformity requirements within the initial safety parameter range as the initial parameter combinations. The fault countermeasure knowledge graph is generated based on historical tension data, historical movement trajectory data, and historical fault case data.

[0098] It should be noted that the fault countermeasure knowledge graph is an intelligent decision-making support system constructed based on historical data. It integrates historical tension data, historical movement trajectory data, and historical fault case data, and forms a structured knowledge network by analyzing the correlation relationships of these data. This knowledge graph can provide reference for the current process parameter optimization, help identify the parameter combinations that may cause faults or quality problems, and provide corresponding solutions. For example, if the historical data shows that under a certain combination of abrasive belt pressure and rotational speed, there have been problems such as too high broken rice rate or uneven milling, the knowledge graph will record this situation and avoid selecting similar parameter combinations during the subsequent optimization process.

[0099] The requirement for milling uniformity refers to the quality standard that requires the removal degree of the bran layer on the surface of rice to be uniform during the rice milling process, and can be evaluated by measuring indicators such as the color and luster of the rice surface and the residual amount of the bran layer.

[0100] Historical tension data refers to the recorded data of the tension borne by the abrasive belt during past rice milling processes. These data reflect the force conditions of the abrasive belt under different operating conditions. By analyzing the historical tension data, the impact of the abrasive belt pressure on rice milling effects (such as the broken rice rate and milling uniformity) can be understood.

[0101] Historical movement trajectory data refers to the recorded data of the movement trajectories of rice grains in the whitening chamber. These data can be obtained by sensors (such as industrial cameras) collecting the positions and movement states of rice grains. By analyzing the historical movement trajectory data, the distribution, collision frequency, and movement paths of rice grains in the whitening chamber can be understood. This information is crucial for optimizing the design and process parameters of the whitening chamber. For example, if the historical data shows that rice grains frequently collide and break at a certain part of the whitening chamber, then the movement trajectory of rice grains can be improved and the collisions reduced by adjusting the gap of the whitening chamber or the rotation speed of the abrasive belt.

[0102] Historical fault case data refers to the case data of equipment failures and quality problems recorded during past rice milling processes. These data include the time of failure, failure types (such as abrasive belt breakage, excessive broken rice grains, uneven milling, etc.), the process parameters at that time, and the solutions taken. By analyzing the historical fault case data, the parameter combinations and operating conditions that may lead to failures can be identified, so as to avoid similar problems in subsequent process optimization.

[0103] It can be understood that, first, the system constructs a fault countermeasure knowledge graph using historical tension data, historical movement trajectory data, and historical fault case data. This knowledge graph can reflect the equipment operating states and quality problems under different parameter combinations. Then, within the initial safety parameter range, the system analyzes the knowledge graph and excludes those parameter combinations that have caused failures or quality problems. Finally, the system screens out the parameter combinations that meet the requirements of milling uniformity and uses them as the initial parameter combinations, providing a safe and efficient starting point for subsequent process optimization to ensure that the rice mill can produce high-quality rice when operating under these parameters.

[0104] As an example, the step of iteratively updating the initial parameter combination through the reinforcement learning algorithm of the intelligent service layer to obtain the target parameter combination includes: based on the broken rice rate data, the detected value of milling uniformity, and the feedback data of the energy consumption sensor in the broken rice rate prediction surface, defining a multi-objective reward function: exploring parameters in the initial parameter combination through the deep deterministic policy gradient algorithm to generate a Pareto optimal solution set of abrasive belt pressure and rotation speed; screening out candidate parameter combinations with a broken rice rate less than the preset broken rice rate threshold and unit mass energy consumption less than the preset energy consumption threshold from the Pareto optimal solution set; updating the candidate parameter combination based on the historical correction data in the fault countermeasure knowledge graph, and returning the step of exploring parameters in the initial parameter combination through the deep deterministic policy gradient algorithm to generate a Pareto optimal solution set of abrasive belt pressure and rotation speed until the multi-objective reward function converges to obtain the target parameter combination.

[0105] The broken rice rate data refers to the broken rice rate values calculated or actually detected through the broken rice rate prediction surface under different process parameter combinations. These data reflect the proportion of broken rice under specific parameter settings such as abrasive belt pressure, rotation speed, and milling chamber clearance.

[0106] The detected value of milling uniformity refers to the value of the uniformity of bran layer removal on the rice surface measured through machine vision or other detection means. Usually, the milling uniformity is evaluated by analyzing indicators such as the color and bran layer residue amount on the rice surface.

[0107] The feedback data of the energy consumption sensor refers to the energy consumption data collected in real time by the energy consumption sensor installed on the equipment during the rice milling process. These data reflect the energy consumption situation of the rice milling machine under different process parameter combinations.

[0108] The multi-objective reward function is an optimization function that comprehensively considers multiple objectives such as broken rice rate, milling uniformity, and energy consumption. It is used to evaluate the advantages and disadvantages of different parameter combinations during the reinforcement learning process. Specifically, the multi-objective reward function can be defined as:

[0109] R = α × (1 - broken rice rate) + β × milling uniformity - γ × unit mass energy consumption

[0110] Among them, α, β, and γ are preset weight coefficients used to balance the priorities between different objectives. For example, R = 0.7 × (1 - real-time broken rice rate) + 0.5 × milling uniformity - 0.2 × unit energy consumption.

[0111] The Pareto optimal solution set refers to the solution set in a multi-objective optimization problem where a set of parameter combinations achieve optimality in some objectives and do not get worse in other objectives. In this embodiment, the Pareto optimal solution set refers to the solution set where the combination of abrasive belt pressure and rotational speed achieves a balance between broken rice rate and energy consumption under the condition that the rice grain stress is less than the preset compressive strength threshold.

[0112] The energy consumption per unit mass refers to the electrical energy consumed per unit mass of rice during the rice milling process, which is usually expressed in units of kilowatt-hours per ton (kWh / ton).

[0113] The preset energy consumption threshold refers to the upper limit value of the energy consumption per unit mass preset according to production requirements and energy efficiency standards. In this embodiment, the preset energy consumption threshold is set to 0.3 kWh / ton.

[0114] The candidate parameter combination refers to the parameter combination in the Pareto optimal solution set that satisfies the broken rice rate being less than the preset broken rice rate threshold and the energy consumption per unit mass being less than the preset energy consumption threshold.

[0115] The historical rectification data refers to the rectification operation data recorded during past operations due to equipment failures or improper parameter settings. These data include the time of rectification, the reasons, the measures taken, and the final results. The historical rectification data is used to update the candidate parameter combination to avoid repeating similar problems in subsequent optimization processes.

[0116] First, the system collects the broken rice rate data in the broken rice rate prediction surface, the actually detected rice whitening uniformity value, and the energy consumption data fed back by the energy consumption sensor. Based on these data, a multi-objective reward function is defined, which comprehensively considers the broken rice rate, rice whitening uniformity, and energy consumption and is used to evaluate the advantages and disadvantages of different parameter combinations. Then, the deep deterministic policy gradient algorithm is used to explore parameters within the initial parameter combination range. By adjusting the values of the abrasive belt pressure and rotational speed, the Pareto optimal solution set that satisfies the rice grain stress being less than the preset compressive strength threshold is found, and these solution sets achieve a balance between the broken rice rate and energy consumption. Next, the candidate parameter combinations with a broken rice rate lower than 12% and an energy consumption per unit mass lower than 0.3 kWh / ton are screened out from the Pareto optimal solution set. After that, based on the historical rectification data in the fault countermeasure knowledge graph, it is checked whether the candidate parameter combination has ever caused faults or quality problems. If so, it is excluded; otherwise, it is retained, thereby updating the candidate parameter combination. Finally, the updated candidate parameter combination is re-input into the deep deterministic policy gradient algorithm for parameter exploration, and the above process is repeated until the multi-objective reward function converges, that is, the reward value no longer changes significantly, and finally the optimized target parameter combination is obtained, ensuring that when the rice milling machine operates under these parameters, it can effectively reduce the broken rice rate, ensure the rice whitening uniformity and energy consumption efficiency, and at the same time avoid the recurrence of historical faults.

[0117] As an example, the step of updating the candidate parameter combination based on the historical rectification data in the fault countermeasure knowledge graph includes: when the wear amount detected by the sand belt edge wear sensor exceeds a preset threshold, obtaining the compensation amount of the sand belt rectifying roller according to the historical rectification data in the fault countermeasure knowledge graph; converting the compensation amount of the sand belt rectifying roller into the adjustment amount of the whitening chamber gap according to a linear conversion relationship, where the linear conversion relationship is established based on the geometric relationship between the sand belt rectifying roller and the whitening chamber; and updating the candidate parameter combination based on the adjustment amount of the whitening chamber gap.

[0118] The sand belt edge wear sensor refers to a sensor installed on a sand belt rice milling machine for real-time monitoring of the wear condition of the sand belt edge. This sensor can detect the wear degree of the sand belt edge and convert the wear data into an electrical signal for output.

[0119] The wear amount refers to the reduction in thickness or width of the sand belt edge due to factors such as friction during use. The wear amount is a quantitative index for measuring the wear degree of the sand belt, usually in millimeters (mm).

[0120] The preset threshold refers to an upper limit value of the wear amount set in advance according to the design and usage requirements of the sand belt. When the wear amount detected by the sand belt edge wear sensor exceeds this preset threshold, the system will trigger corresponding rectification or maintenance operations.

[0121] The compensation amount of the sand belt rectifying roller refers to the amount that the sand belt rectifying roller needs to be adjusted according to the historical rectification data in the fault countermeasure knowledge graph. The sand belt rectifying roller is a component used to adjust the running direction of the sand belt to ensure the normal operation of the sand belt in the whitening chamber.

[0122] The linear conversion relationship refers to the mathematical relationship between the compensation amount of the sand belt rectifying roller and the adjustment amount of the whitening chamber gap. This relationship is established based on the geometric relationship between the sand belt rectifying roller and the whitening chamber and is usually a linear function:

[0123] Δg = k·Δd

[0124] Where, Δd is the displacement compensation amount of the rectifying roller, Δg is the adjustment amount of the whitening chamber gap, and k is the proportionality coefficient, generally 0.5, which is determined by the geometric parameters of the sand belt wrap angle (120°) and the roller diameter (200 mm).

[0125] The adjustment amount of the whitening chamber gap refers to the amount that the whitening chamber gap needs to be adjusted according to the linear conversion relationship. The whitening chamber gap is the distance between the sand belt and the material turning device, and the size of this distance directly affects the whitening effect of brown rice and the broken rice rate.

[0126] First, the abrasive belt edge wear sensor monitors the wear condition of the abrasive belt edge in real time. When the wear amount exceeds the preset threshold of 2 mm, the system records this state. Then, the system accesses the fault countermeasure knowledge graph, filters out historical rectification cases similar to the current abrasive belt wear situation, and extracts the compensation amount data of the abrasive belt rectification roller from them. These data are obtained based on past successful rectification experiences and can provide a reference for the current adjustment. Next, according to the geometric relationship between the abrasive belt rectification roller and the rice milling chamber, a linear conversion relationship is established. Specifically, the proportionality coefficient between the compensation amount of the abrasive belt rectification roller and the adjustment amount of the rice milling chamber gap is determined. For example, assuming that for every 1 mm compensation of the abrasive belt rectification roller, the rice milling chamber gap needs to be adjusted by 0.5 mm, then the proportionality coefficient is 0.5. Finally, according to the calculated adjustment amount of the rice milling chamber gap, the rice milling chamber gap value in the candidate parameter combination is updated, that is, the original gap value is added with the adjustment amount to obtain a new gap value, and this new value is applied to the subsequent rice milling process to optimize the rice milling effect, reduce rice grain breakage, and ensure the normal operation of the abrasive belt at the same time.

[0127] In this embodiment, first, the target data is input into the simulation model to obtain the stress distribution of the rice grain group in the rice milling chamber, which helps to understand the force conditions of the rice grains at different positions and time points. Then, according to the stress distribution data, the number of rice grains whose stress exceeds the preset rice grain compressive strength threshold in each time window is counted to obtain the broken rice probability data. This step can accurately identify which rice grains are more likely to break under specific process parameters, thereby providing a basis for optimizing the process parameters. Next, a broken rice rate prediction surface is constructed based on the broken rice probability data, and the initial safe parameter ranges of the abrasive belt pressure, rotation speed, and rice milling chamber gap are obtained according to the prediction surface. This helps to reduce the broken rice rate while ensuring the rice milling quality, improve production efficiency and economic benefits. Finally, in combination with the fault countermeasure knowledge graph, parameter combinations that meet the rice milling uniformity requirements are screened out within the initial safe parameter range as the initial parameter combination. This step uses historical data and experience to further ensure that the selected parameter combination can not only reduce the broken rice rate but also ensure the rice milling uniformity, improve product quality, avoid the recurrence of historical faults, and improve the reliability and stability of production.

[0128] This application also provides a rice milling machine intelligent control device, which includes:

[0129] A digital twin construction module for constructing a digital twin system of the rice milling machine. The digital twin system includes a physical layer, a virtual layer, and an intelligent service layer. The virtual layer includes a simulation model based on the contact dynamics of the abrasive belt - rice grain.

[0130] A data fusion module for collecting the operation data of the rice milling machine through the physical layer and performing spatio - temporal alignment of the operation data with the simulation model to obtain target data.

[0131] A prediction module, configured to predict the change trend of the broken rice rate during the rice milling process based on the target data and the simulation model, and generate an initial parameter combination of the abrasive belt pressure, rotation speed, and whitening chamber clearance;

[0132] A parameter optimization module, configured to iteratively update the initial parameter combination through the reinforcement learning algorithm of the intelligent service layer to obtain a target parameter combination;

[0133] A control module, configured to adjust the actuator of the rice milling machine according to the target parameter combination to complete the intelligent control of the rice milling machine.

[0134] The intelligent control device for a rice milling machine provided by this application adopts the intelligent control method for a rice milling machine in the above embodiment, and can solve the technical problem of how to reduce the broken rice rate during the rice milling process and achieve intelligent control. Compared with the prior art, the beneficial effects of the intelligent control device for a rice milling machine provided by this application are the same as those of the intelligent control method for a rice milling machine provided by the above embodiment, and other technical features in the intelligent control device for a rice milling machine are the same as the features disclosed in the above embodiment method, and will not be elaborated here.

[0135] This application provides an intelligent control device for a rice milling machine. The intelligent control device for a rice milling machine includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the intelligent control method for a rice milling machine in the first embodiment above.

[0136] The intelligent control device for a rice milling machine provided by this application adopts the intelligent control method for a rice milling machine in the above embodiment, and can solve the technical problem of how to reduce the broken rice rate during the rice milling process and achieve intelligent control. Compared with the prior art, the beneficial effects of the intelligent control device for a rice milling machine provided by this application are the same as those of the intelligent control method for a rice milling machine provided by the above embodiment, and other technical features in the intelligent control device for a rice milling machine are the same as the features disclosed in the previous embodiment method, and will not be elaborated here.

[0137] This application provides a computer-readable storage medium, having computer-readable program instructions (i.e., computer programs) stored thereon, and the computer-readable program instructions are used to execute the intelligent control method for a rice milling machine in the above embodiment. The above computer-readable storage medium may be included in the intelligent control device for a rice milling machine; or it may exist alone without being assembled into the intelligent control device for a rice milling machine.

[0138] The readable storage medium provided by this application is a computer-readable storage medium. The computer-readable storage medium stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned intelligent control method of the rice milling machine, and can solve the technical problem of how to reduce the broken rice rate during the rice milling process and achieve intelligent control. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided by this application are the same as those of the intelligent control method of the rice milling machine provided by the above embodiment, and will not be elaborated here.

[0139] During the processes of hulling and whitening paddy rice, the rice grains will come into contact with and collide with the rice milling roller, the sieve mesh, and other rice grains. This interaction prompts the rice grains to tumble and gradually remove the outer layer to achieve the effect of whitening. However, if the collision force received by the rice grains exceeds the limit of their structural strength, the rice grains will break and form broken rice.

[0140] This application also proposes a rice milling machine 100, aiming to solve the problem of high broken rice rate of traditional sand roller rice milling machines.

[0141] Please refer to Figure 3 and Figure 4 , in an embodiment of this application, the rice milling machine 100 includes a frame 1, a feeding device 2, and a rice milling assembly 3. The feeding device 2 is installed on the frame 1. The feeding device 2 includes a feeding hopper 21. The rice milling assembly 3 is installed on the frame 1. The rice milling assembly 3 has a feeding end 3a and a discharging end 3b. The feeding end 3a of the rice milling assembly 3 is arranged below the outlet of the feeding hopper 21. The rice milling assembly 3 includes two spaced-apart rice milling mechanisms 30 to form a whitening chamber gap between the two rice milling mechanisms 30. Each rice milling mechanism 30 includes a rice milling part 31 facing the other rice milling mechanism 30. The rice milling part 31 is elastically arranged and has a concavo-convex structure thereon. Among them, the two rice milling parts 31 can move relative to each other in the first horizontal direction.

[0142] It should be noted that the frame 1 is the support frame of the entire device, ensuring that each component can be stably installed.

[0143] The feeding device 2 includes a feeding hopper 21, which is directly installed on the upper part of the frame 1 and is used to evenly distribute the paddy rice to be processed into the rice milling assembly 3.

[0144] The rice milling assembly 3 is composed of two spaced-apart rice milling mechanisms 30, forming a whitening chamber gap. Each rice milling mechanism 30 includes the rice milling part 31 with the concavo-convex structure. By setting the concavo-convex structure, the friction force can be increased, promoting the rotational movement of the brown rice rather than simply linear movement.

[0145] The two rice milling parts 31 can move relative to each other in the first horizontal direction. That is to say, it can be that both of the two rice milling parts 31 can move in the first horizontal direction, but they move at different speeds. Thus, a relative movement can be generated between the two rice milling parts 31; it can also be that only one of the rice milling parts 31 can move in the first horizontal direction while the other is fixed, and the same effect of promoting the rotation of paddy around its own axis can be achieved; it can also be that both of the two rice milling parts 31 can move in the first horizontal direction, but they accelerate and decelerate alternately according to a certain periodic pattern (such as a sine wave form).

[0146] The two rice milling parts 31 can move relative to each other in the first horizontal direction. In this way, long-grain paddy rotates around its own long axis within the gap of the rice milling chamber. Compared with the complex and irregular movements that rice grains may experience in the traditional rice milling process, the bending moment acting on the paddy in the direction of its long axis can be reduced, thereby reducing the risk of breakage caused by excessive stress.

[0147] In addition, while the rice grains are rotating, they are also moving in the direction of the linear speed difference. The paddy enters from the feeding end 3a, is milled through the gap of the rice milling chamber, and then flows out from the discharging end 3b.

[0148] It should be noted that the rice milling part 31 is a roller with an adjustable concave-convex structure, and grooves and protrusions with specific shapes and sizes are designed on the surface of the roller to increase the frictional force between the paddy and the rice milling part 31 and promote the rotation of the paddy around its long axis.

[0149] The rice milling part 31 can also be set as a vibrating sieve plate. Two sieve plates are installed opposite to each other and can generate relative vibrations in the first horizontal direction. The surface of the sieve plate is covered with small holes and irregular protrusions. When the paddy passes through, it will be affected by the vibration of the sieve plate and rotate.

[0150] The preferred material of the rice milling part 31 is made of a material that can undergo slight deformation or has a certain elasticity, such as rubber or abrasive belt 36, etc.

[0151] In the technical solution of the present application, by enabling the two rice milling parts 31 to be relatively movable in the first horizontal direction, when long-grain paddy enters the gap of the rice milling chamber from the feed hopper 21, it is subjected to forces between the two rice milling parts 31 with the concave-convex structure. Since the rice milling parts 31 can move relative to each other in the first horizontal direction, the paddy is guided to rotate around its long axis, making the pressure distribution on the paddy during the entire rice milling process more uniform, effectively avoiding the problem of breakage caused by excessive local pressure. At the same time, the contact between the surface of the paddy and the rice milling part 31 is more sufficient and smooth, improving the hulling and rice milling efficiency and solving the problem of high broken rice rate of the traditional sand roller rice milling machine 100.

[0152] Further, in this embodiment, both of the two rice milling parts 31 are movably arranged along the first horizontal direction, and the linear speed differences of the two rice milling parts 31 along the first horizontal direction are set so that a linear speed difference is formed between the two rice milling parts 31 in the first horizontal direction.

[0153] Specifically, to achieve the setting of the linear speed differences of the two rice milling parts 31 moving along the first horizontal direction, the two rice milling parts 31 are respectively connected to independent drive systems (such as servo motors) to precisely control their respective rotational speeds and linear speeds. By adjusting the speed of one or both of the rice milling parts 31, the required linear speed difference can be formed between them.

[0154] When there is a linear speed difference between the two rice milling parts 31, the rice bran adhering to the rice milling area will be "swept away" by the relative movement, reducing the possibility of rice bran accumulation. For the areas that do not participate in the direct rice milling work, the existence of the linear speed difference also promotes air circulation and helps to remove the possible remaining impurities.

[0155] Specifically, please refer to Figure 7 , in the first embodiment, the two rice milling mechanisms 30 are arranged at intervals in the second horizontal direction, and each rice milling part 31 is provided with a rice groove 31a extending in the up-down direction.

[0156] It can be understood that the two rice milling mechanisms 30 are arranged along the direction perpendicular to the first horizontal direction (i.e., the second horizontal direction) to ensure that there is enough space for the paddy rice to freely fall between the two mechanisms.

[0157] Each rice milling part 31 is provided with a rice groove 31a extending from the top to the bottom, and the paddy rice naturally slides down through the rice groove 31a under the action of gravity.

[0158] The rice milling part 31 is made of a material with a certain elasticity, which can be adaptively adjusted according to the different shapes and sizes of the paddy rice to form an optimal contact state, and at the same time allows the shape of the rice groove 31a to be changed to increase the path length of the paddy rice falling.

[0159] The rice milling part 31 moves along the first direction at a high speed and contacts the surface of the paddy rice to achieve the grinding process. The paddy rice enters the rice groove 31a from the upper end of the rice milling part 31 and starts to slide down under the action of gravity. Due to the setting of the rice groove 31a, the paddy rice not only moves downward but also rotates around its own axis due to the frictional force of the rice milling part 31, performing the grinding process on the paddy rice and removing the bran on the outer layer of the paddy rice to gradually achieve the refinement of the rice.

[0160] The rice milling part 31 arranged elastically can automatically adjust the contact pressure according to the size of paddy rice, ensuring sufficient grinding effect without damaging the paddy rice. The elastic contact can also effectively reduce the collision between paddy rice grains, further reducing the broken rice rate.

[0161] By changing the shape of the rice groove 31a, the falling travel of paddy rice can be increased, thereby prolonging the contact time between the paddy rice and the abrasive belt 36, ensuring a more thorough grinding effect. The longer contact time enables the rice milling machine 100 to have better hulling and polishing effects, contributing to improving the quality and consistency of the final product.

[0162] It should be noted that the number of the rice grooves 31a determines the quality of paddy rice that can be processed per unit time, and it cannot be adjusted arbitrarily, and the maintenance is relatively cumbersome. In order to solve this problem, please refer to Figures 3 to 6 , in the second embodiment, the two rice milling mechanisms 30 are arranged at intervals in the vertical direction.

[0163] By adjusting the rotation speed of each rice milling part 31, the linear speed of paddy rice when moving on its surface can be directly controlled. Different linear speeds will affect the contact time and force between the paddy rice and the surface of the rice milling part 31, thereby affecting the hulling and whitening effects.

[0164] Similarly, changing the rotation frequency of the rice milling part 31 can adjust the angular velocity of paddy rice, making it more inclined to spin during the rice milling process, which helps to reduce local stress concentration and lower the broken rice rate.

[0165] Compared with the complex multi-directional motion design, the rice milling mechanism 30 arranged at intervals in the vertical direction has a relatively simple structure, is easy to manufacture and maintain. It reduces the demand for precision parts and lowers the production cost.

[0166] Specifically, please refer to Figure 5 and Figure 6 , in one embodiment, each rice milling mechanism 30 includes a driving device 32, a first driving wheel 33, a second driving wheel 34, a transmission belt 35 and an abrasive belt 36. The driving device 32 has a driving part that rotates around the second horizontal direction; the first driving wheel 33 and the second driving wheel 34 are arranged at intervals in the first horizontal direction, and the first driving wheel 33 is connected to the driving part to be driven to rotate by the driving part; the transmission belt 35 is wound around the peripheries of the two driving wheels to drive the second driving wheel 34 to rotate when the first driving wheel 33 rotates; the abrasive belt 36 is connected to the transmission belt 35 and is wound around the peripheries of the first driving wheel 33 and the second driving wheel 34 to be driven to move by the transmission belt 35; wherein, among the two abrasive belts 36, one side of each abrasive belt 36 facing the other abrasive belt 36 forms the rice milling part 31.

[0167] The drive device 32 has a drive part that can rotate around the second horizontal direction. The drive device 32 provides a power source for the rice milling mechanism 30.

[0168] The first drive wheel 33 and the second drive wheel 34 are spaced apart in the first horizontal direction. The first drive wheel 33 is directly connected to the drive part of the drive device 32, and thus is driven to rotate by the drive part, forming a driving wheel. The second drive wheel 34 is connected to the first drive wheel 33 through a transmission belt 35 and rotates synchronously with the rotation of the first drive wheel 33, forming a driven wheel.

[0169] The transmission belt 35 surrounds the peripheries of the first drive wheel 33 and the second drive wheel 34, and transmits the power of the first drive wheel 33 to the second drive wheel 34, ensuring that the two drive wheels work in coordination.

[0170] The abrasive belt 36 is the part that actually contacts and processes the paddy rice. It is connected to the transmission belt 35 and also surrounds the peripheries of the first drive wheel 33 and the second drive wheel 34. The abrasive belt 36 is driven by the transmission belt 35 to move synchronously, forming an effective grinding surface. Two abrasive belts 36 are arranged face to face, and the sides facing each other constitute the rice milling part 31, that is, the area where the paddy rice is milled when passing between the two abrasive belts 36.

[0171] After the drive device 32 is started, the drive part begins to rotate around the second horizontal direction, driving the connected first drive wheel 33 to rotate. The first drive wheel 33 transmits the power to the second drive wheel 34 through the transmission belt 35, causing the two drive wheels to rotate synchronously. Since the abrasive belt 36 is connected to the transmission belt 35 and surrounds the peripheries of the two drive wheels, when the drive wheels rotate, the abrasive belt 36 will also move accordingly. The moving direction of the abrasive belt 36 is the same as the rotation direction of the drive wheels, ensuring a continuous and stable grinding process.

[0172] The paddy rice enters the gap of the whitening chamber between the two abrasive belts 36 from above. While moving downward under the action of gravity, it is subjected to the frictional force of the high-speed rotating abrasive belt 36. As the paddy rice moves forward along the abrasive belt 36, the outer bran is gradually removed, and finally polished rice is obtained. Since the two abrasive belts 36 are arranged face to face, the formed rice milling part 31 provides sufficient normal pressure to ensure a good whitening effect.

[0173] By adjusting the speed of the drive device 32, the linear speed of the abrasive belt 36 can be controlled, thereby affecting the degree of whitening of the paddy rice.

[0174] Please refer to Figure 6, in one embodiment, one side of the rice milling mechanism 30 in the second horizontal direction is rotatably mounted on the frame 1 around the rotation axis in the first horizontal direction.

[0175] In this way, when it is necessary to replace the abrasive belt 36, the rice milling mechanism 30 can be rotated from the horizontal direction to the vertical direction.

[0176] It can be understood that when the two rice milling mechanisms 30 are rotatably connected to the frame 1, they can be correspondingly rotatably connected to both sides of the frame 1 in the second horizontal direction. In this way, when it is necessary to replace the abrasive belt 36 of the two rice milling mechanisms 30, each rice milling mechanism 30 can be flipped to the opposite sides of the frame 1, which is convenient for operation. In addition, when it is only necessary to maintain the rice milling mechanism 30 located at the lower layer, the upper rice milling mechanism 30 can be flipped to provide clearance.

[0177] It should be noted that during the brown rice milling process, the bran layer on the surface of the brown rice is milled to form bran powder. Part of the bran powder adheres to the milled brown rice and flows out of the gap of the whitening chamber, while most of the bran powder will adhere to the abrasive belt 36. If not processed in time, the surface of the abrasive belt 36 will be covered with bran powder, and the milling efficiency will be greatly reduced.

[0178] Furthermore, in order to improve the milling efficiency, please refer to Figure 4 and Figure 9 , in this embodiment, the rice mill 100 further includes a bran removing mechanism 4. The bran removing mechanism 4 is provided corresponding to the abrasive belt 36, and the bran removing mechanism 4 includes a blowing device 41 and / or a suction device 42.

[0179] The blowing device 41 can be installed at a position close to the abrasive belt 36, especially those places where bran is likely to accumulate. The blowing device 41 can be fixed or can adjust the angle and position as needed to ensure that the air flow can cover all key areas.

[0180] By generating a directional strong air flow, the blowing device 41 can blow the bran adhering to the surface of the brown rice and the abrasive belt 36 away, preventing it from re-mixing into the processed rice grains.

[0181] Specifically, the blowing device 41 can be set as an air knife. After compressed air enters the air knife, it is blown out at a high speed in the form of an air flow thin sheet with a thickness of only 0.05 mm on one side. Through the Coanda effect principle and the special geometric shape of the air knife, this thin air curtain can reach up to 30 - 40 times the ambient air, and form a thin, high-strength and large-airflow impact air curtain.

[0182] The suction device 42 is mainly used to collect the chaff and fine impurities loosened by the blowing device 41 or other means. The suction port of the suction device 42 can be arranged to cooperate with the blowing device 41. Preferably, it is arranged downstream of the action range of the blowing device 41 to effectively capture the blown-up impurities.

[0183] It should be noted that during the milling process of paddy rice, the particle size distributions of different varieties of paddy rice are different. Even for the same variety of paddy rice, the dimensions of each axis will vary slightly due to differences in its own geometric shape.

[0184] Furthermore, in order to enable the rice milling machine 100 to meet the processing requirements of different types of paddy rice, the clearance height of the whitening chamber gap can be adjusted. Please refer to Figures 3 to 5 , in this embodiment, the two rice milling mechanisms 30 are arranged at intervals in the vertical direction, and the two rice milling mechanisms 30 are movably arranged in the direction of approaching and separating from each other.

[0185] By adjusting the distance between the two rice milling parts 31, the normal pressure applied to the paddy rice can be changed. A larger distance enables a smaller normal pressure to be generated, which is suitable for a gentler processing; while reducing the distance will increase the normal pressure, which is applicable to situations where stronger frictional force is required.

[0186] With such an arrangement, customized processing can be carried out for paddy rice with different hardnesses and shapes, improving the scope of application and processing quality.

[0187] Specifically, please refer to Figure 5 , the rice milling machine 100 further includes an adjusting mechanism 5. The adjusting mechanism 5 is used to adjust the movement of the movably arranged rice milling mechanism 30. The adjusting mechanism 5 includes a slide table module 51 and a driving mechanism 52. The slide table module 51 includes a linear guide rail, a slide seat and a ball screw structure. The linear guide rail extends in the vertical direction; the slide seat is movably arranged on the linear guide rail in the vertical direction; the ball screw structure includes a screw rod and a driving nut sleeved on the periphery of the screw rod, and the nut is connected to the slide seat; wherein, the driving mechanism 52 includes a handwheel 521 or a driving motor for driving the screw rod to rotate.

[0188] The slide table module 51 realizes the sliding of the slide seat through the ball screw structure to accurately adjust the up and down movement of the rice milling mechanism 30.

[0189] The driving mechanism 52 is used to drive the screw rod to rotate, thereby driving the driving nut and the slide seat to move. According to specific application requirements, the handwheel 521 or the driving motor can be selected as the power source.

[0190] Manually rotate the handwheel 521 to move the drive nut. This method is suitable for application scenarios that require precise positioning but infrequent adjustment.

[0191] When driven by a drive motor, the motor serves as the power source, and the rotation angle and speed of the lead screw can be automatically controlled through the control system. It is suitable for occasions with high automation, rapid response, and continuous adjustment.

[0192] When the position of the rice milling mechanism 30 needs to be adjusted, start the drive mechanism 52. If it is driven by the handwheel 521, the operator manually rotates the handwheel 521; if it is driven by a drive motor, the control system sends an instruction to the motor to make it start working. The lead screw starts to rotate. Since the drive nut is restricted to move only along the axial direction of the lead screw and cannot rotate itself, it will move up and down along the lead screw.

[0193] As the drive nut moves, the slide seat connected to it will also move up and down along the linear guide rail, thereby driving the entire rice milling mechanism 30 to change its position in the vertical direction.

[0194] The slide table module 51 ensures the smoothness and accuracy of the rice milling mechanism 30 during the up and down movement. The ball screw structure has high transmission efficiency and positioning accuracy, and can achieve position adjustment at the micron level.

[0195] In this way, when processing paddy with different particle sizes or hardness, the height of the rice milling mechanism 30 can be finely adjusted to optimize the grinding effect, ensuring both the processing quality and reducing the broken rice rate.

[0196] Further, please refer to Figure 8 , in this embodiment, the feed hopper 21 has two side walls 211 arranged oppositely, and an outlet of the feed hopper 21 is formed between the lower ends of the two side walls 211; the feeding device 2 further includes a feeding rotating shaft 22 arranged at the outlet of the feed hopper 21. The feeding rotating shaft 22 forms a feeding gap with the corresponding side wall 211. The feeding rotating shaft 22 is rotatably arranged around its axis, and the feeding rotating shaft 22 is used to convey the rice grains in the feed hopper 21 towards the feeding gap during its rotation stroke.

[0197] The feed hopper 21 has two oppositely arranged side walls 211, and the two side walls 211 gradually approach each other at the lower end and finally form a narrow outlet.

[0198] The feeding rotating shaft 22 is located between the two side walls 211 and maintains a certain spacing from them to form a feeding gap. The feeding rotating shaft 22 can rotate freely around its own axis and is driven by a motor or other driving device 32. The surface of the feeding rotating shaft 22 can be designed with spiral blades or similar structures as needed to enhance the pushing effect.

[0199] The paddy rice is first poured into the feeding hopper 21. Due to the action of gravity, the paddy rice will naturally slide down along the two side walls 211 and gradually gather at the bottom of the feeding hopper 21, that is, near the outlet. When the feeding rotating shaft 22 starts to rotate, its surface (such as spiral blades) will contact and push the rice grains forward, passing through the feeding gap and entering the rice milling assembly 3.

[0200] By adjusting the rotation speed of the feeding rotating shaft 22, the rate at which the paddy rice enters the rice milling assembly 3 can be precisely controlled, thereby avoiding the blockage problem caused by excessive feeding or the situation where insufficient feeding affects the production efficiency.

[0201] By precisely controlling the rotation speed of the feeding rotating shaft 22 and the size of the feeding gap, the accurate regulation of the inflow rate of the rice grains can be achieved, ensuring the smooth progress of subsequent processing steps. It effectively prevents blockage caused by excessive or too-fast entry of materials, improving the safety and reliability of the equipment operation.

[0202] Further, please continue to refer to Figure 8 , in this embodiment, the two side walls 211 extend downward in a direction approaching each other; the feeding device 2 further includes at least one adjusting plate 23, the adjusting plate 23 is laid on the corresponding side wall 211, and the adjusting plate 23 is slidably arranged along the extending direction of the side wall 211 to adjust the size of the outlet during its sliding stroke.

[0203] The two side walls 211 of the feeding hopper 21 are gradually inclined inward from top to bottom, that is, they extend in a direction approaching each other, so as to guide the paddy rice to concentrate and flow towards the bottom outlet, and the material flow rate can be effectively controlled by changing the outlet width.

[0204] At least one of the side walls 211 is correspondingly provided with one of the adjusting plates 23, the adjusting plate 23 is laid on the surface of the side wall 211 and can slide along the inclined direction of the side wall 211. It can be understood that a slide rail or similar guiding structure is provided between the adjusting plate 23 and the side wall 211 to ensure that the adjusting plate 23 can smoothly move within the set range without deviating from the track.

[0205] By simply moving the adjusting plate 23, the inflow rate of the material can be precisely controlled, avoiding problems such as blockage caused by excessive feeding or insufficient feeding affecting the efficiency, and ensuring the stable operation of the production line.

[0206] Further, please refer to Figure 3 and Figure 10 , in this embodiment, the rice milling machine 100 further includes a detection mechanism 6 located below the rice milling assembly 3. The detection mechanism 6 includes a conveying assembly 61, evacuation teeth 62, a detection device, and a light-shielding cover 64. The conveying assembly 61 includes a conveyor belt extending along the first horizontal direction. The conveying assembly 61 has a first end 61a and a second end 61b in the first horizontal direction. The first end 61a of the conveying assembly 61 is correspondingly arranged at the discharge end 3b of the rice milling part 31. The conveyor belt is used to carry the rice grains conveyed by the discharge end 3b of the rice milling assembly 3 and convey the rice grains between the first end 61a and the second end 61b of the conveying assembly 61. An evacuation station a and a detection station b are sequentially arranged on the conveying assembly 61 in the direction from its first end 61a to the second end 61b; the evacuation teeth 62 are correspondingly arranged at the evacuation station a; the detection device includes an image acquisition device 63 for acquiring an image of the upper surface of the conveyor belt at the detection station b; the light-shielding cover 64 is correspondingly arranged at the detection station b and is arranged above the conveyor belt, and the image acquisition device 63 is arranged in the light-shielding cover 64.

[0207] The conveying assembly 61 includes the conveyor belt extending along the first horizontal direction (i.e., the direction parallel to the ground). The conveyor belt has a first end 61a and a second end 61b. The first end 61a of the conveying assembly 61 is directly correspondingly arranged at the discharge end 3b of the rice milling part 31, for receiving the rice grains coming out of the rice milling assembly 3 and conveying them along the conveyor belt from the first end 61a to the second end 61b.

[0208] The evacuation station a and the detection station b are sequentially arranged on the conveyor belt from the first end 61a to the second end 61b, ensuring that the rice grains can be properly dispersed before being detected.

[0209] The evacuation teeth 62 are arranged at the evacuation station a, for dispersing the rice grains that may be in a group or aggregated discharged from the rice milling part 31, ensuring that each rice grain can be individually captured by the detection device for an image.

[0210] The evacuation teeth 62 can work by mechanical vibration or other means to ensure that the rice grains are evenly distributed on the entire conveyor belt.

[0211] The detection device mainly includes the image acquisition device 63, for acquiring an image of the rice grains on the upper surface of the conveyor belt at the detection station b. The image acquisition device 63 can be a high-resolution camera or other devices suitable for shooting the details of small objects, and is installed inside the light-shielding cover 64 to ensure stable light in the shooting environment and reduce external interference.

[0212] The light-shielding cover 64 ensures that the image acquisition device 63 can obtain clear rice grain images under optimal lighting conditions.

[0213] After the rice grains are processed by hulling, they first fall onto the conveyor belt on the first end 61a of the conveyor assembly 61 through the discharge end 3b of the hulling assembly 3. The conveyor belt runs continuously, transporting the rice grains smoothly from the first end 61a to the second end 61b, passing through the evacuation station a and the detection station b during this process. When the rice grains reach the evacuation station a, the evacuation teeth 62 start to work, dispersing any aggregated rice grains to ensure that each rice grain can be individually subjected to subsequent detection. The dispersed rice grains continue to move forward to the detection station b, where the image acquisition device 63 takes pictures of the rice grains on the conveyor belt. The captured rice grain images are transmitted to a computer system for further analysis, and the system automatically identifies and classifies the quality grades of the rice grains according to preset standards, such as whether there are broken grains, impurities, etc.

[0214] By monitoring the product after hulling in real time, quality problems can be detected and solved in a timely manner, significantly improving the quality of the final product. Moreover, the detection process is highly automated, reducing the need for manual intervention and improving work efficiency. In addition, by using image acquisition technology and data analysis algorithms, the state of the rice grains can be accurately judged, avoiding subjective errors in traditional methods.

[0215] Further, please refer to Figure 3 and Figure 4 , in an embodiment, the rice huller 100 further includes a material transfer mechanism 7 disposed between the hulling assembly 3 and the detection mechanism 6. The material transfer mechanism 7 includes a guide plate 71, and the guide plate 71 is inclined downward from top to bottom in a direction approaching the first end 61a of the conveyor assembly 61.

[0216] The guide plate 71 is inclined downward from top to bottom in a direction approaching the first end 61a of the conveyor assembly 61. In this way, the gravity can be utilized to guide the rice grains to slide naturally onto the conveyor belt, avoiding the problems of material accumulation or blockage.

[0217] The material of the guide plate 71 is usually selected as a smooth and wear-resistant material, such as stainless steel or engineering plastics, to reduce the frictional loss and breakage risk when the rice grains move on its surface.

[0218] By reasonably designing the guide plate 71, the material transfer efficiency can be significantly improved, avoiding accumulation. The smooth surface of the guide plate 71 and the appropriate inclination angle can reduce the collision and friction of the rice grains during transfer, thereby reducing the breakage rate.

[0219] Please refer to Figure 11 and Figure 12In another embodiment, the material transfer mechanism 7 includes a material transfer hopper 72 and a material guide plate 73. The upper opening of the material transfer hopper 72 is arranged corresponding to the discharge end 3b of the rice milling assembly 3, and the lower opening of the material transfer hopper 72 is arranged corresponding to the first end 61a of the conveying assembly 61. The material guide plate 73 is rotatably installed at the lower opening of the material transfer hopper 72 around a rotation axis in the second horizontal direction. The material guide plate 73 sequentially has a first material guiding position, a closed position, and a second material guiding position during its rotation stroke. In the first material guiding position, the material guide plate 73 is inclined downward from top to bottom toward the direction close to the first end 61a of the conveying assembly 61. In the closed position, the material guide plate 73 closes the lower opening of the material transfer hopper 72. In the second material guiding position, the material guide plate 73 is inclined downward from top to bottom toward the direction away from the first end 61a of the conveying assembly 61.

[0220] The material transfer hopper 72 has an upper opening and a lower opening. The upper opening is directly arranged corresponding to the discharge end 3b of the rice milling assembly 3 for receiving the rice grains after whitening treatment.

[0221] The lower opening is arranged corresponding to the first end 61a of the conveying assembly 61 to ensure that the rice grains can be smoothly transferred onto the conveyor belt for subsequent detection.

[0222] The material guide plate 73 is installed at the lower opening of the material transfer hopper 72 and can rotate around a rotation axis in the second horizontal direction.

[0223] The material guide plate 73 has three different working positions: the first material guiding position, the closed position, and the second material guiding position to meet different operation requirements.

[0224] Please refer to Figure 11 , when the material guide plate 73 is in the first material guiding position, the material guide plate 73 is inclined downward from top to bottom toward the direction close to the first end 61a of the conveying assembly 61. At this time, after the rice grains fall from the upper opening of the material transfer hopper 72, they slide along the surface of the material guide plate 73 and finally fall on the first end 61a of the conveying assembly 61, starting to move toward the evacuation station a to facilitate guiding them to the detection system when the surface of the rice grains needs to be inspected.

[0225] When the material guide plate 73 is in the closed position, it completely covers and closes the lower opening of the material transfer hopper 72. This position can be used to temporarily stop the flow of materials, for example, during equipment maintenance or adjustment, to prevent the rice grains from continuing to enter the conveyor belt.

[0226] In addition, when the material guide plate 73 is in the closed position, it can also be used as a safety measure in the emergency shutdown state to avoid material overflow or equipment damage.

[0227] Please refer to Figure 12 When the material guide plate 73 is in the second material guiding position, the material guide plate 73 is inclined downward from top to bottom in a direction away from the first end 61a of the conveying assembly 61. In this way, the material can be guided to other collection points to collect white rice or for special treatment.

[0228] It can be understood that the rotation of the material guide plate 73 can be achieved by manual operation (such as a handle or a knob) or an automatic control system (such as an electric motor drive). In the automatic control mode, the position of the material guide plate 73 can be automatically switched according to a preset program or sensor feedback to adapt to different production scenarios.

[0229] In this way, the control of multiple material flow directions can be achieved through a simple mechanical structure to meet different production requirements. The above description is only an exemplary embodiment of the present application, and does not limit the patent scope of the present application. Any equivalent structural transformation made under the technical concept of the present application by using the content of the specification and drawings of the present application, or any direct / indirect application in other related technical fields is included in the patent protection scope of the present application.

Claims

1. An intelligent control method for a rice milling machine, characterized in that, The method includes: Constructing a digital twin system of a rice milling machine, the digital twin system including a physical layer, a virtual layer, and an intelligent service layer, and the virtual layer including a simulation model based on the contact dynamics between the abrasive belt and the rice grains; Collecting the operation data of the rice milling machine through the physical layer, and performing spatio-temporal alignment on the operation data and the simulation model to obtain target data; Predicting the change trend of the broken rice rate during the rice milling process based on the target data and the simulation model, and generating an initial parameter combination of the abrasive belt pressure, rotational speed, and gap of the whitening chamber; Iteratively updating the initial parameter combination through the reinforcement learning algorithm of the intelligent service layer to obtain a target parameter combination; Adjusting the actuator of the rice milling machine according to the target parameter combination to complete the intelligent control of the rice milling machine.

2. The method according to claim 1, wherein The step of predicting the change trend of the broken rice rate during the rice milling process based on the target data and the simulation model, and generating an initial parameter combination of the abrasive belt pressure, rotational speed, and gap of the whitening chamber includes: Inputting the target data into the simulation model to obtain the stress distribution of the rice grain group in the whitening chamber; Counting the number of rice grains with stress greater than the preset rice grain compressive strength threshold in each time window according to the stress distribution to obtain broken rice probability data; Constructing a broken rice rate prediction surface based on the broken rice probability data, and obtaining an initial safety parameter range of the abrasive belt pressure, rotational speed, and gap of the whitening chamber according to the broken rice rate prediction surface; Combining with a fault countermeasure knowledge graph, screening parameter combinations that meet the requirements of whitening uniformity within the initial safety parameter range as the initial parameter combination, and the fault countermeasure knowledge graph is generated according to historical tension data, historical motion trajectory data, and historical fault case data.

3. The method according to claim 2, wherein The step of constructing a broken rice rate prediction surface based on the broken rice probability data, and obtaining an initial safety parameter range of the abrasive belt pressure, rotational speed, and gap of the whitening chamber according to the broken rice rate prediction surface includes: According to the broken rice probability data, fitting the non-linear relationship between the abrasive belt pressure, rotational speed, and gap of the whitening chamber and the broken rice rate through a Gaussian process regression algorithm to generate a broken rice rate prediction surface; Defining a safety threshold boundary on the broken rice rate prediction surface according to a preset broken rice rate threshold and a preset standard deviation of whitening uniformity to obtain an initial safety parameter range.

4. The method according to claim 2, wherein The step of iteratively updating the initial parameter combination through the reinforcement learning algorithm of the intelligent service layer to obtain a target parameter combination includes: Based on the broken rice rate data, whitening uniformity detection value, and energy consumption sensor feedback data in the broken rice rate prediction surface, defining a multi-objective reward function: Performing parameter exploration in the initial parameter combination through a deep deterministic policy gradient algorithm to generate a Pareto optimal solution set of the abrasive belt pressure and rotational speed; Selecting candidate parameter combinations with a broken rice rate less than a preset broken rice rate threshold and unit mass energy consumption less than a preset energy consumption threshold from the Pareto optimal solution set; Update the candidate parameter combination based on the historical rectification data in the fault countermeasure knowledge graph, and return the step of generating the Pareto optimal solution set of the abrasive belt pressure and rotational speed by performing parameter exploration in the initial parameter combination through the deep deterministic policy gradient algorithm until the multi-objective reward function converges to obtain the target parameter combination.

5. The method according to claim 4, characterized in that, The step of updating the candidate parameter combination based on the historical rectification data in the fault countermeasure knowledge graph includes: When the wear amount detected by the abrasive belt edge wear sensor exceeds the preset threshold, obtain the abrasive belt rectification roller compensation amount according to the historical rectification data in the fault countermeasure knowledge graph; Convert the abrasive belt rectification roller compensation amount into a whitening chamber gap adjustment amount according to the linear conversion relationship, and the linear conversion relationship is established based on the geometric relationship between the abrasive belt rectification roller and the whitening chamber; Update the candidate parameter combination based on the whitening chamber gap adjustment amount.

6. The method according to claim 1, characterized in that, The actuator includes a whitening chamber gap adjustment mechanism, an abrasive belt drive motor, and a ball screw slide table. The step of adjusting the actuator of the rice milling machine according to the target parameter combination to complete the intelligent control of the rice milling machine includes: Convert the target abrasive belt pressure value in the target parameter combination into a pulse width modulation signal; Control the whitening chamber gap adjustment mechanism according to the pulse width modulation signal; Based on the target rotational speed value in the target parameter combination, adjust the input voltage of the abrasive belt drive motor through a fuzzy proportional-integral-derivative controller; Adjust the ball screw slide table according to the target gap value in the target parameter combination so that the whitening chamber gap reaches the target gap value to complete the intelligent control of the rice milling machine.

7. The method according to any one of claims 1 to 6, characterized in that, The physical layer is deployed with a three-axis force sensor and an industrial camera. The step of aligning the operation data and the simulation model in space and time to obtain the target data includes: Perform Butterworth low-pass filtering on the tension data collected by the three-axis force sensor to obtain the filtered tension data; Align the time stamps of the rice grain images collected by the industrial camera and the filtered tension data to obtain a synchronized data packet; Map the rice grain position coordinates in the synchronized data packet to the three-dimensional coordinate system of the simulation model through a feature point matching algorithm to obtain the target data.

8. An intelligent control device for a rice milling machine, characterized in that, The device includes: A digital twin construction module for constructing a digital twin system of the rice milling machine. The digital twin system includes a physical layer, a virtual layer, and an intelligent service layer. The virtual layer includes a simulation model based on the contact dynamics between the abrasive belt and the rice grains; A data fusion module for collecting the operation data of the rice milling machine through the physical layer and aligning the operation data and the simulation model in space and time to obtain the target data; A prediction module for predicting the change trend of the broken rice rate during the rice milling process based on the target data and the simulation model, and generating an initial parameter combination of the abrasive belt pressure, rotational speed, and whitening chamber gap; A parameter optimization module for iteratively updating the initial parameter combination through the reinforcement learning algorithm of the intelligent service layer to obtain the target parameter combination; A control module, configured to adjust an actuator of the rice milling machine according to the target parameter combination, so as to complete the intelligent control of the rice milling machine.

9. An intelligent control device for a rice milling machine, characterized in that, The device includes: a memory, a processor, and a computer program stored on the memory and executable on the processor, where the computer program is configured to implement the steps of the rice milling machine intelligent control method according to any one of claims 1 to 7.

10. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the rice milling machine intelligent control method according to any one of claims 1 to 7.

Images