Coal flow self-adaptive speed regulation system and method
The adaptive speed regulation system, which utilizes lightweight hardware and an intelligent processing layer, solves the problem of high hardware and computing power requirements in existing coal flow speed regulation systems, achieving efficient and stable coal flow speed regulation and reducing energy consumption and equipment wear.
Patent Information
- Application Number
- CN202511603772.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-05
- Publication Date
- 2026-02-17
AI Technical Summary
Existing coal flow speed regulation systems have high hardware requirements, require large computing resources, and are not flexible enough in identifying coal flow status and responding to speed regulation.
By employing a lightweight hardware perception layer and intelligent processing layer, combined with the Retinexformer algorithm and a lightweight YOLOv8 target detection model, adaptive speed regulation is achieved by smoothing coal flow fluctuations through a circular statistical buffer.
It reduces computational load and hardware costs, improves speed regulation stability and response speed, reduces equipment wear and energy consumption, and enhances transportation efficiency.
Smart Images
Figure CN121544533A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the field of coal mine transportation, in particular to a coal flow self-adaptive speed regulation system and method. BACKGROUND
[0002] The current mainstream coal flow speed regulation system generally relies on laser scanning technology, and its typical working process is as follows: first, the cross section of the coal flow is scanned by a laser beam to obtain profile point cloud data; then, the trapezoidal integral method is used to calculate the cross-sectional area of the coal flow in real time, and then the volume flow is calculated in combination with the current speed of the conveying belt; finally, the motor drive frequency is adjusted according to the preset flow-belt speed mapping relationship. This technology has high hardware requirements (laser emitter or ultrasonic radar and its supporting equipment), and requires a large amount of computing resources due to the continuous execution of integral operation. SUMMARY
[0003] In order to solve the technical problems of the current mainstream coal flow speed regulation system, such as high hardware requirements (laser emitter or ultrasonic radar and its supporting equipment), and large amount of computing resources due to the continuous execution of integral operation, one purpose of the present application is to provide a lightweight and efficient coal flow self-adaptive speed regulation system and method.
[0004] To achieve the above purpose, an embodiment of the present application provides a coal flow self-adaptive speed regulation system, comprising: a belt conveyor; a control unit; a hardware perception layer arranged at a key node of the belt conveyor, the hardware perception layer being configured to collect coal flow image data on the belt conveyor in real time; an intelligent processing layer arranged in the control unit, configured to receive the coal flow image data, classify the coal flow state into multiple categories through a target detection model, count the frequency of occurrence of each category within a preset period based on a ring-shaped statistical buffer, and issue a speed regulation instruction according to the dominant category with the highest frequency of occurrence; a control execution layer arranged in the control unit, configured to receive the speed regulation instruction issued by the intelligent processing layer and issue a control instruction to regulate the speed of the belt conveyor.
[0005] In the above technical solution, the control unit comprises an explosion-proof industrial computer or an embedded computing platform.
[0006] In the above technical solution, the hardware perception layer is an industrial camera with high dynamic range and wide dynamic range.
[0007] In the above technical solution, the intelligent processing layer comprises: an image preprocessing module using a Retinexformer algorithm to enhance image quality; A coal flow state recognition module outputs the coverage state of the coal flow in the image with enhanced quality and classifies it into four categories of "large coal flow", "medium coal flow", "small coal flow" or "empty load"; A decision control module is internally provided with a ring-shaped statistical buffer with a preset period, continuously records and counts the frequency of occurrence of various states in the ring-shaped statistical buffer, and issues a corresponding speed regulation instruction according to the category with the highest frequency of occurrence.
[0008] In the technical solution, the control execution layer includes an explosion-proof frequency converter, which adjusts the running speed of the belt conveyor in real time according to the received speed regulation instruction.
[0009] Embodiments of the present application provide a coal flow self-adaptive speed regulation method, comprising: S1. The hardware perception layer collects coal flow image data at key nodes of the belt conveyor in real time; S2. The intelligent processing layer receives the coal flow image data collected by the hardware perception layer, and finally issues a corresponding speed regulation instruction by analyzing the coal flow image data; S3. The control execution layer receives the speed regulation instruction issued by the intelligent processing layer, and issues a control instruction to regulate the speed of the belt conveyor; S4. In the decision control module, the ring-shaped statistical buffer is a first-in first-out queue, and when it is found that the empty load state continuously occurs in a preset period, a speed regulation instruction is issued to enter the ultra-low speed standby mode.
[0010] In the technical solution, step S2 specifically comprises: S21. The image preprocessing module performs image repair and enhancement through the Retinexformer algorithm; S22. The coal flow state recognition module generates and fuses multiple feature maps by using a target detection model with YOLOv8 as the backbone network and improved by deep lightening, and classifies multiple coal flow images into four categories of "large coal flow", "medium coal flow", "small coal flow" or "empty load" through the feature maps; S23. The decision control module is internally provided with a ring-shaped statistical buffer with a preset period, which continuously records and counts the frequency of occurrence of various categories in the ring-shaped statistical buffer, and issues a corresponding speed regulation instruction according to the category with the highest frequency of occurrence.
[0011] In the technical solution, a multi-dimensional collaborative attention mechanism module is introduced in the coal flow state recognition module to enhance the model's perception ability of coal flow targets in complex scenes.
[0012] Additional aspects and advantages of the present application will become apparent in the light of the following description section, or will be appreciated by practice of the present application. BRIEF DESCRIPTION OF DRAWINGS
[0013] The above and / or additional aspects and advantages of the present application will become apparent and more readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings in which: Fig. 1 is a workflow diagram of the system of the present application; Fig. 2 is a network structure diagram of the real coal flow state recognition module of the present application; Fig. 3 is a system architecture diagram of the present application; DETAILED DESCRIPTION
[0014] In order to enable a clearer understanding of the above-mentioned purposes, features and advantages of the present application, the present application will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.
[0015] In the following description, a large number of specific details are set forth in order to facilitate a thorough understanding of the present application, however, the present application can also be implemented in other ways different from those described herein, and therefore, the scope of protection of the present application is not limited by the specific embodiments disclosed below.
[0016] Reference will now be made to the following Figs. 1 to 3 to describe a coal flow self-adaptive speed regulation system and method according to some embodiments of the present application.
[0017] As shown in Figs. 1 to 3 , the embodiments of the present application provide a coal flow self-adaptive speed regulation system, which comprises: a belt conveyor; a control unit comprising an explosion-proof industrial computer or an embedded computing platform; a hardware perception layer composed of embedded image acquisition devices arranged at key nodes (such as discharge points and transfer points) of the belt conveyor, the embedded image acquisition devices adopt high dynamic range and wide dynamic range industrial cameras, which are packaged in explosion-proof, dustproof and moisture-proof protective covers to adapt to the harsh underground environment, the industrial cameras collect coal flow video on the belt of the belt conveyor in real time, and transmit the coal flow image data through a gigabit Ethernet or a 5G wireless module; an intelligent processing layer arranged in the explosion-proof industrial computer or the embedded computing platform, the intelligent control layer comprising an image preprocessing module, a coal flow state recognition module and a decision control module The intelligent processing layer comprises: An image preprocessing module is used to solve the problem of image quality degradation caused by uneven underground lighting, dust interference and dark environment by using the Retinexformer algorithm for image enhancement. The Retinexformer algorithm first estimates and decomposes the input image into illumination and reflection components, and then uses the illumination-guided transformer module inside the Retinexformer algorithm to repair and enhance the illumination map, effectively suppressing noise, eliminating color distortion, and significantly improving the visibility of dark details, providing high-quality image input for subsequent recognition.
[0018] It should be noted that the Retinexformer algorithm and the illumination-guided transformer module are both existing technologies for enhancing image quality, and will not be described in detail here.
[0019] A coal flow state recognition module uses a target detection model based on YOLOv8 as the backbone network and improved by deep lightening. The model uses the RepGhost module to reconstruct the feature extraction network, generates and fuses multiple feature maps through structural reparameterization technology, significantly reduces the model parameter quantity and computational quantity (FLOPs) while ensuring feature richness; defines BiFPN_Concat2 to replace the original Concat in YOLOv8Neck, integrates a bidirectional cross-scale connection mechanism and a fast normalization fusion strategy, and constructs an efficient multi-scale feature fusion framework; at the same time, the MCA module is introduced to improve the attention to coal flow features in three dimensions of channel, height and width, and enhance the model's perception ability of coal flow targets in complex scenes. The coal flow state recognition module finally outputs the coverage state of the coal flow in the image and classifies it into four categories: "large coal flow", "medium coal flow", "small coal flow" or "empty load".
[0020] It should be noted that the target detection model based on YOLOv8 as the backbone network and improved by deep lightening is an existing technology, and its working principle will not be described in detail here; the target detection model uses the RepGhost module to reconstruct the feature extraction network, which is also an existing technology, and will not be described in detail here; the working principle of the MCA module is an existing technology, and will not be described in detail here.
[0021] A decision control module receives the continuous state sequence output by the recognition module. To eliminate transient fluctuations and avoid frequent switching of the speed control command, a ring-shaped statistical buffer (FIFO queue) with a length of a preset period is set in the module to continuously record and count the frequency of occurrence of each type of state in the time window. The decision engine determines the state with the highest frequency of occurrence in the period as the dominant state based on the "majority rule", and generates a corresponding belt speed control command according to a preset "state-belt speed" mapping table (large coal flow: 100% speed, medium coal flow: 70% speed, small coal flow: 30% speed, and empty load: 10% speed).
[0022] A control execution layer is arranged in the control unit, and the control execution layer includes an explosion-proof frequency converter that communicates with the intelligent processing layer through Profibus-DP or Ethernet / IP protocol. The explosion-proof frequency converter adjusts the input frequency and voltage of the driving motor in real time according to the received belt speed control command, thereby controlling the running speed of the belt conveyor and realizing adaptive matching of the speed and the coal flow. The decision control module of the intelligent processing layer is provided with a ring-shaped statistical buffer, which is a first-in first-out queue with a length of a preset period. The system writes the recognized coal flow state at the end of the buffer every time, and simultaneously removes the oldest state at the head end, thereby realizing sliding update of the buffer. The speed control decision is not responsive to the instantaneous state, but is based on the dominant state with the highest frequency of occurrence in the sliding window. This mechanism ensures the stability and hysteresis of the speed control command, fundamentally avoids frequent operation of the speed control mechanism due to transient fluctuations of the coal flow, and significantly reduces the equipment loss and invalid energy consumption. Compared with the mode in the prior art in which the belt speed is immediately changed as soon as the coal flow state is changed, the speed control frequency is reduced, the calculation amount is greatly reduced, and the cost is reduced.
[0023] For example, the preset period of the ring statistical buffer (FIFO queue) is 60 seconds, and the coverage state of the coal flow in the image is recorded every 1 second. The ring statistical buffer records 60 times, that is, 60 times in each preset period. During the first 60 seconds of the operation of the device, the belt speed corresponding to the large coal flow is operated. After the current 60 seconds of recording is completed, the frequency of the coal flow coverage state in the previous 60 seconds is determined to determine the belt speed of the 61st second. If the frequency of the large coal flow in the previous 60 seconds is the highest, the belt speed of the 61st second remains unchanged. When entering the 61st second, the coal flow coverage state recorded in the first second is removed, and the coal flow coverage state recorded in the 61st second is retained. The frequency of the coal flow coverage state in the second to 61st second is counted again. If the frequency of the medium coal flow in the second to 61st second exceeds that of the large coal flow and becomes the highest frequency, the system automatically sends a belt speed control instruction to change the belt speed of the 62nd second. The belt speed of the 62nd second corresponds to the medium coal flow. By analogy, the belt speed is adjusted by counting the frequency of the coal flow coverage state in 60 seconds. If the frequency in 60 seconds is all empty state and lasts more than 60 seconds, the system automatically sends an instruction to enter the ultra-low speed standby mode to maximize energy saving.
[0024] The whole system integrates network communication function, and builds underground ring network through industrial switch to upload the identification results, belt speed information, device state and other data of each node to the ground safety production monitoring center in real time. The monitoring center software platform provides man-machine interactive interface for real-time monitoring of system running state, query of historical data, setting of speed regulation parameters, fault diagnosis and alarm, and forms a complete intelligent management and control closed loop.
[0025] Corresponding to the above-mentioned coal flow self-adaptive speed regulation system, the application also discloses a coal flow self-adaptive speed regulation method, comprising: S1. The hardware perception layer collects the coal flow video at the key nodes of the belt conveyor belt in real time, and transmits the coal flow image data through a gigabit Ethernet or a 5G wireless module; S2. The intelligent processing layer receives the coal flow image data collected by the hardware perception layer, and finally sends a corresponding speed regulation instruction by analyzing the coal flow image data; S21. The image preprocessing module in the intelligent processing layer receives the coal flow image data collected by the hardware perception layer, and adopts the Retinexformer algorithm for image enhancement in view of the image quality degradation problems caused by uneven underground illumination, dust interference and dark environment. The Retinexformer algorithm first estimates and decomposes the illumination component and the reflection component of the input image, and then repairs and enhances the illumination map through the illumination-guidedtransformer module in the Retinexformer algorithm, effectively suppresses noise, eliminates color distortion, and significantly improves the visibility of dark details, providing high-quality image input for subsequent identification.
[0026] S22. The coal flow state recognition module of the intelligent processing layer uses a target detection model based on YOLOv8 as the backbone network and improved in depth and lightness. The model uses the RepGhost module to reconstruct the feature extraction network, generates and fuses multiple feature maps through structural reparameterization technology, greatly reduces the model parameter quantity and calculation quantity (FLOPs) on the premise of ensuring the richness of the features; at the same time, the MCA module is introduced to improve the attention to the coal flow features in the channel, height and width dimensions, and enhance the perception ability of the model to the coal flow targets in complex scenes. The coal flow state recognition module finally outputs the coverage state of the coal flow in the image, and classifies it into four categories: "large coal flow", "medium coal flow", "small coal flow" or "empty load".
[0027] S23. The decision control module of the intelligent processing layer records and counts the frequency of each type of state appearing in the time window through the built-in ring statistical buffer (FIFO queue) with a length of the preset period. The decision engine determines the state with the highest frequency in the period as the dominant state based on the "majority rule", and generates the corresponding belt speed control instruction according to the preset "state-belt speed" mapping table (large coal flow: 100% speed, medium coal flow: 70% speed, small coal flow: 30% speed, empty load: 10% speed).
[0028] The control execution layer communicates with the intelligent processing layer through Profibus-DP or Ethernet / IP protocol, and its core is the explosion-proof frequency converter. The frequency converter directly responds to the real-time belt speed instruction output by the intelligent processing layer, which is the mapping of the dominant state of the coal flow in the current ring buffer sliding window. By synchronously adjusting the frequency and voltage of the driving motor, the adaptive matching of the belt conveyor running speed and the coal flow is realized.
[0029] The present application has the following advantages: 1. The dominant state statistics and stepless speed regulation strategy based on the ring statistical buffer are invented, which smooths the coal flow state fluctuations through the preset period statistical window, dynamically adjusts the belt speed according to the dominant type in the period, and introduces the empty load automatic standby mechanism, realizes the accurate matching of "coal flow-belt speed", the operation amount decreases exponentially, the equipment is simple, and the system energy consumption, transportation equipment wear and tear, operation efficiency and stability are effectively reduced; 2. The pre-trained model is used to classify the coal flow state in real time, replacing the traditional laser scanning and integral calculation, which greatly reduces the algorithm complexity and hardware cost, while ensuring the recognition accuracy, significantly improving the system response speed and economy; 3. Retinexformer low light enhancement and lightweight convolutional network are fused to effectively overcome the poor imaging conditions such as underground dark light and dust, improve the image quality and recognition robustness; 4. By introducing the RepGhost lightweight module, the model calculation amount and storage overhead are greatly reduced under the premise of maintaining high precision, supporting real-time deployment and stable operation on embedded devices.
[0030] In the present application, the terms "mounting", "connecting", "connecting", "fixing" and the like should be understood broadly, for example, "connecting" can be fixed connection, or detachable connection, or integral connection; "connecting" can be directly connected, or indirectly connected through an intermediate medium. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.
[0031] In the description of the present application, it should be understood that the positions or position relationships indicated by the terms "inner", "outer" and the like are based on the positions or position relationships shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the devices or units referred to must have a particular direction, be constructed and operated in a particular position, therefore, it cannot be understood as a limitation to the present application.
[0032] In the description of the present application, the description of the terms "one embodiment", "some embodiments", "specific embodiments" and the like means that the specific features, structures, materials or characteristics described in conjunction with the embodiments or examples are included in at least one embodiment or example of the present application. In the present application, the illustrative description of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.
[0033] The above is only the preferred embodiment of the present application, and is not used to limit the present application. For those skilled in the art, the present application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.
Claims
1. A coal flow adaptive speed regulation system, characterized in that, include: Belt conveyor; Control unit; The hardware sensing layer is located at key nodes of the belt conveyor and is used to collect coal flow image data on the belt conveyor in real time. The intelligent processing layer, located within the control unit, is used to receive the coal flow image data, classify the coal flow state into multiple categories through a target detection model, count the frequency of each category within a preset period based on a circular statistical buffer, and issue a speed adjustment command based on the dominant category with the highest frequency of occurrence. A control execution layer is located within the control unit. The control execution layer receives speed adjustment commands from the intelligent processing layer and issues control commands to adjust the speed of the belt conveyor.
2. The coal flow adaptive speed regulation system according to claim 1, characterized in that, The control unit includes an explosion-proof industrial computer or an embedded computing platform.
3. The coal flow adaptive speed regulation system according to claim 2, characterized in that, The hardware sensing layer is an industrial camera with high dynamic range and wide dynamic range.
4. The coal flow adaptive speed regulation system according to claim 3, characterized in that, The intelligent processing layer includes: The image preprocessing module uses the Retinexformer algorithm to enhance image quality; The coal flow status recognition module outputs the coverage status of the coal flow in the enhanced image and classifies it into four categories: "large coal flow", "medium coal flow", "small coal flow" or "no load". The decision control module has a preset periodic circular statistical buffer that continuously records and counts the frequency of various states within the circular statistical buffer, and issues corresponding speed adjustment commands based on the category with the highest frequency.
5. The coal flow adaptive speed regulation system according to claim 4, characterized in that, The control execution layer includes an explosion-proof frequency converter, which adjusts the running speed of the belt conveyor in real time according to the received speed adjustment command.
6. A method for adaptive speed regulation of coal flow, characterized in that, include: S1. The hardware sensing layer collects coal flow image data at key nodes of the belt conveyor in real time; S2. The intelligent processing layer receives coal flow image data collected by the hardware perception layer, and finally issues corresponding speed regulation commands by analyzing the coal flow image data; S3. The control execution layer receives the speed adjustment command from the intelligent processing layer and issues control commands to adjust the speed of the belt conveyor. S4. In the decision control module, the circular statistical buffer is a first-in-first-out queue. When the idle state is detected to occur continuously within a preset period, a speed adjustment command to enter the ultra-low speed standby mode is issued.
7. The coal flow adaptive speed regulation system according to claim 6, characterized in that, Step S2 specifically includes: S21. The image preprocessing module performs image restoration and enhancement using the Retinex Former algorithm; S22. The coal flow status identification module uses a target detection model with YOLOv8 as the backbone network and has been deeply lightweighted and improved to generate and fuse multiple feature maps. The feature maps are used to classify multiple coal flow images into four categories: "large coal flow", "medium coal flow", "small coal flow" or "empty". S23. The decision control module has a preset periodic circular statistical buffer, which continuously records and counts the frequency of each category in the circular statistical buffer, and issues the corresponding speed adjustment command based on the category with the highest frequency.
8. The coal flow adaptive speed regulation system according to claim 6, characterized in that, The coal flow state recognition module incorporates a multi-dimensional collaborative attention mechanism module to enhance the model's ability to perceive coal flow targets in complex scenarios.