CN-121360646-B - Intelligent medium adding method and computer equipment suitable for coal preparation plant

CN121360646BCN 121360646 BCN121360646 BCN 121360646BCN-121360646-B

Abstract

The application discloses an intelligent medium adding method and computer equipment suitable for a coal preparation plant, wherein the method comprises the following steps: cameras are symmetrically arranged in the center of the top of the medium warehouse and on the two sides of the travelling crane beam in the moving direction, synchronously shoot images and transmit the images to an AI controller for pretreatment. When the medium adding industrial agent receives the medium adding operation instruction, the medium adding industrial agent analyzes the characteristics of the crown block track and the medium stack in the image, and associates the characteristics with the position instruction sequence of the common controller to form a VLC dynamic association model. Based on the planned optimal suction point and the collision-free shortest path, the AI controller converts the optimal suction point and the collision-free shortest path into a pulse width instruction to control the crown block to move. After the crown block grabs the medium and sends the medium into the dense medium barrel, the control system determines the clean water amount according to the weight of the medium, and controls the clean water pipe and the air pipe to inject water into the barrel and blow air to form dense medium suspension coal dressing. The medium stack identification, intelligent crown block scheduling and accurate medium adding can be realized, and the problems of low efficiency, poor precision, insufficient reliability and the like of the traditional manual medium adding are solved.

Inventors

BAO ZHEN
GUO GE
LI QIEN
YUE SHUAIJUN

Assignees

国能智深控制技术有限公司

Dates

Publication Date: 20260508
Application Date: 20251021

Claims (8)

1. The intelligent mediating method suitable for the coal preparation plant is characterized by being applied to a control system, wherein the control system comprises a common controller, an AI controller and mediating industrial intelligent agents, and the method comprises the following steps: arranging a top panoramic camera at the top center of a medium warehouse, symmetrically arranging side wall dynamic tracking cameras on two side walls of the medium warehouse along the movement direction of a crown block beam, so that the top panoramic camera and the side wall dynamic tracking cameras synchronously shoot images, and transmitting the shot images to an AI controller for preprocessing, wherein the shooting range of the top panoramic camera covers the middle axis area of the crown block beam and the medium pile area in the medium warehouse, the top panoramic camera and the side wall dynamic tracking cameras adopt multi-view CV cameras, and the AI controller extracts color features and texture features based on the images acquired by the multi-view CV cameras and identifies magnetic characteristics through the color features and the texture features; When the medium adding industrial agent receives the medium adding operation instruction, the medium adding industrial agent analyzes the motion track of the crown block and the visual characteristics of the medium stack form in the image preprocessed by the AI controller by using a computer visual technology, and correlates the analyzed visual characteristics with a position instruction sequence sent by a common controller for controlling the crown block to form a VLC dynamic correlation model containing the mapping relation among the vision, the position and the instruction, wherein the position instruction sequence is acquired by the medium adding industrial agent on the common controller for controlling the crown block; The mediating industry agent determines all candidate suction points to which the sucker possibly moves based on the VLC dynamic association model; Screening the optimal suction points minimizing the optimization cost from candidate suction points by the mediating industrial agent through a multi-objective optimization cost function, wherein the multi-objective optimization cost function is as follows: P is the real-time coordinates of candidate suction points determined based on the VLC dynamic association model in a global coordinate system, the function f (P) is a multi-objective optimization cost function, For the euclidean distance of the current position of the suction cup to the candidate suction point, The magnetic content of the medium that is the candidate suction point, For the medium height of the candidate suction points, , Respectively a distance weight coefficient, a medium quality weight coefficient and a medium height weight coefficient, ; The method comprises the steps that a non-collision shortest path is planned by an intervening industrial intelligent agent based on a VLC dynamic association model and fed back to an AI controller, the AI controller converts the non-collision shortest path into a pulse width instruction, and a common controller of the crown block is controlled based on the pulse width instruction, so that the crown block is driven by the common controller to move to an optimal suction point along the non-collision shortest path, and a sucker is corresponding to the crown block; The control system controls the crown block to suck media in the media stack through the sucker at an optimal suction point, the control system determines the required clean water amount according to the weight of the media added into the media stack, controls the clean water pipe to flow into the media stack through the common controller of the clean water pipe, and controls the air pipe to blow air into the media stack through the common controller of the air pipe to form heavy media suspension for coal dressing operation, wherein the clean water pipe is arranged above the media stack, the air pipe is arranged in the media stack, and the crown block, the clean water pipe and the air pipe are respectively corresponding to the common controller.
2. The method of claim 1, wherein the mediating industrial agent comprises a lightweight U-Net semantic segmentation model, wherein the mediating industrial agent uses computer vision techniques to resolve visual features of crown block motion trajectories and media stack morphology in images preprocessed by the AI controller, comprising: The medium adding industrial agent uses a lightweight U-Net semantic segmentation model to carry out pixel-level classification on the image preprocessed by the AI controller to obtain a sucker area, a medium stack area, a ground area and a crown block area; The medium adding industrial intelligent agent obtains a three-dimensional contour of the medium stack in the medium stack area through an Ojin threshold contour extraction algorithm, and determines visual features of the medium stack form based on the three-dimensional contour; In the crown block area, the mediating industry intelligent agent converts the pixel level position change of the crown block into a continuous motion track of a physical space through the space-time correlation of multi-frame images, and extracts the visual characteristics of the crown block motion track based on the converted continuous motion track.
3. The method of claim 2, wherein the mediating industrial agent correlates the resolved visual features with a sequence of position commands issued by a generic controller controlling the crown block, comprising: The method comprises the steps that an industrial intelligent agent calculates the coordinates of characteristic points under a global coordinate system by combining calibrated top panoramic camera external parameters and side wall dynamic tracking camera external parameters on the basis of a multi-view triangulation principle, wherein the pixel coordinates of the same characteristic points are captured by images synchronously shot by a top panoramic camera and a side wall dynamic tracking camera, and the coordinates of the characteristic points under the global coordinate system are calculated until visual features of the crown block movement track and the medium pile form analyzed in the images are unified to the global coordinate system; the industrial intelligent agent associates the visual characteristics after unifying the coordinates with a position instruction sequence sent by a common controller for controlling the crown block.
4. The method of claim 1, wherein the crown block performs orthogonal motion only for X, Y and Z axes, and wherein the mediated industry agent plans a collision-free shortest path, comprising: The medium-adding industrial intelligent body plans a collision-free shortest path according to the principle of firstly far axis and then near axis based on an improved path optimizing algorithm.
5. The method of claim 4, wherein the crown block corresponds to a servo motor, the servo motor corresponds to X, Y and a Z-axis, each axis corresponds to a preset pulse distance conversion factor, the crown block corresponds to a preset maximum acceleration and a preset maximum speed, the AI controller converting the collision-free shortest path to a pulse width command, comprising: the AI controller decomposes the collision-free shortest path into target distances which need to be moved in X, Y and Z axis directions respectively; The AI controller calculates the target pulse number required by each axis of the servo motor to reach the target distance according to the preset pulse distance conversion coefficient; the AI controller determines speed curves of the crown block at different stages according to the collision-free shortest path, the preset maximum acceleration and the preset maximum speed of the crown block; the AI controller calculates the pulse width modulation duty ratio corresponding to each stage according to the speed curves of the crown block in different stages; and the AI controller packages the calculated target pulse number and the corresponding pulse width modulation duty ratio into a pulse width instruction.
6. The method of any one of claims 1 to 5, wherein before the mediation industry agent receives the mediation operation instruction, the method further comprises: an operator inputs the medium adding amount and a heavy medium system to be subjected to medium adding operation through language input or text on an interactive interface of the medium adding industrial intelligent body, and the medium adding industrial intelligent body generates a medium adding operation instruction based on the medium adding amount and the heavy medium system to be subjected to medium adding operation; and controlling the AI controller based on the mediation operation instruction so as to enable an operation result of the AI controller to be fed back to an operator through the mediation industry intelligent agent, wherein the mediation industry intelligent agent is correspondingly provided with an interactive interface.
7. The method of claim 6, wherein the AI controller includes an edge calculation unit that transmits the captured image to the AI controller for preprocessing, comprising: And transmitting the shot image to an edge computing unit, and preprocessing the shot image by adopting a bilateral filtering algorithm.
8. A computer device comprising a storage medium, a processor and a computer program stored on the storage medium and executable on the processor, characterized in that the processor implements the method of intelligent mediating applicable to coal preparation plants according to any one of claims 1 to 7 when executing the computer program.

Description

Intelligent medium adding method and computer equipment suitable for coal preparation plant Technical Field The application relates to the technical field of coal, in particular to an intelligent medium adding method and computer equipment suitable for a coal preparation plant. Background Dense medium coal dressing is a method of sorting with liquid with a density between that of clean coal and gangue (or middlings) as medium. Has the advantages of high separation efficiency, wide feeding granularity range, easy operation and the like. The principle of dense medium coal dressing is that clean coal with density lower than that of the medium floats, while gangue or middling coal with density higher than that of the medium sinks, and then different products are collected respectively. An important task in dense medium coal dressing is thus the formulation of dense medium suspensions. The heavy medium suspension is typically formulated from a particulate solid heavy mass and water. Because coal itself carries heavy matters, the loss of the heavy matters is unavoidable, and new media are required to be continuously added in the production process. At present, coal preparation plants are commonly provided with special people for adding media, preparing heavy media suspension and conveying heavy media suspension, and the work is basically operated manually, so that the labor and material consumption is high. And manual operation stability is poor, and the efficiency of adding medium is low, can not in time satisfy the production demand, is unfavorable for high-efficient separation. Disclosure of Invention In view of the above, the application provides an intelligent medium adding method and computer equipment suitable for a coal preparation plant, which can realize medium stack identification, intelligent overhead travelling crane scheduling and accurate medium adding and solve the problems of low efficiency, poor precision, insufficient reliability and the like of the traditional manual medium adding. According to one aspect of the present application, there is provided an intelligent mediating method applicable to a coal preparation plant, applied to a control system, the control system including a general controller, an AI controller and a mediating industrial agent, the method comprising: Arranging a top panoramic camera at the center of the top of a medium warehouse, and symmetrically arranging side wall dynamic tracking cameras along the movement direction of a crown block beam on two side walls of the medium warehouse, so that the top panoramic camera and the side wall dynamic tracking cameras synchronously shoot images, and transmitting the shot images to an AI controller for preprocessing, wherein the shooting range of the top panoramic camera covers the center axis area of the crown block beam and the medium stacking area in the medium warehouse; When the medium adding industrial agent receives the medium adding operation instruction, the medium adding industrial agent analyzes the motion track of the crown block and the visual characteristics of the medium stack form in the image preprocessed by the AI controller by using a computer visual technology, and correlates the analyzed visual characteristics with a position instruction sequence sent by a common controller for controlling the crown block to form a VLC dynamic correlation model containing the mapping relation among the vision, the position and the instruction, wherein the position instruction sequence is acquired by the medium adding industrial agent on the common controller for controlling the crown block; The method comprises the steps that an mediating industry intelligent agent plans an optimal suction point for sucking media and feeds a collision-free shortest path back to an AI controller based on a VLC dynamic association model, the AI controller converts the collision-free shortest path into a pulse width instruction, and controls a common controller of an overhead travelling crane based on the pulse width instruction so that the overhead travelling crane is driven by the common controller to move to the optimal suction point along the collision-free shortest path, wherein the overhead travelling crane is correspondingly provided with a sucker; The control system controls the crown block to suck media in the media stack through the sucker at an optimal suction point, the control system determines the required clean water amount according to the weight of the media added into the media stack, controls the clean water pipe to flow into the media stack through the common controller of the clean water pipe, and controls the air pipe to blow air into the media stack through the common controller of the air pipe to form heavy media suspension for coal dressing operation, wherein the clean water pipe is arranged above the media stack, the air pipe is arranged in the media stack, and the crown block, the clean water pipe and the air pipe are res