Search

CN-121994335-A - Double-scale collaborative self-learning weighing system and method for alloy charging

CN121994335ACN 121994335 ACN121994335 ACN 121994335ACN-121994335-A

Abstract

The invention relates to a double-weighing collaborative self-learning weighing system and a method for alloy feeding, which belong to the technical field of metallurgy. After each charging, the micro-weighing unit is used for accurately weighing the weight of the retained residual materials, and meanwhile, the main scale is used for weighing the final blanking weight, and the actual value of the aerial material column is reversely calculated and recorded by combining the reading when vibration is stopped. The system continuously collects the data, correlates the data with parameters such as material type, feeding flow and the like, and establishes a dynamic air material column quantity prediction model for different materials by utilizing a self-learning algorithm. When feeding is performed next time, the system directly calls the predicted value which is learned and optimized to control the vibration stopping point, so that high-precision and self-adaptive feeding control is realized, and the metering error problem caused by the change of the aerial material column is fundamentally solved.

Inventors

  • SHI CAIXIA
  • ZHAO YUNFENG
  • GAO ZHONGJIANG

Assignees

  • 中冶赛迪工程技术股份有限公司

Dates

Publication Date
20260508
Application Date
20251231

Claims (10)

  1. 1. A double-scale collaborative self-learning weighing system for alloy feeding is characterized by comprising an alloy bin (1), a vibrating feeder (2), a weighing hopper (4) and a PLC control system (5), wherein the alloy bin (1) is arranged above the vibrating feeder (2), the weighing hopper (4) is arranged below the vibrating feeder (2), a micro-weighing unit (3) is arranged between the vibrating feeder (2) and the weighing hopper (4), an upper inlet of the micro-weighing unit (3) is in sealing connection with a discharge hole of the vibrating feeder (2) through a first flexible connecting piece (6), a lower outlet of the micro-weighing unit (3) is in sealing connection with a feed hole of the weighing hopper (4) through a second flexible connecting piece (9), the micro-weighing unit (3) is provided with a weighing module (7) and a stop valve (8) for opening and closing the lower outlet of the micro-weighing unit, the PLC control system (5) is used for receiving weight signals of the micro-weighing unit (3) and the weighing hopper (4), and performing collaborative metering and prediction algorithm and prediction error from a vibratory feeding column (2) is controlled by using a prediction algorithm.
  2. 2. The double-scale collaborative self-learning weighing system for alloy charging according to claim 1, wherein the PLC control system (5) is connected with an environmental humidity sensor and is used for acquiring environmental humidity information, and taking the environmental humidity information as one of associated parameters, storing the environmental humidity information together with the actual value G_actual of the air material column theory and participating in the establishment and updating of the prediction model.
  3. 3. A dual-scale collaborative self-learning weighing system for alloy charging according to claim 1, characterized in that the weighing accuracy of the weighing module (7) is higher than the weighing accuracy of the weighing hopper (4).
  4. 4. The double-scale collaborative self-learning weighing system for alloy charging according to claim 1, wherein the weighing range of the weighing module (7) is 1-5% of the maximum weighing value of the weighing hopper (4).
  5. 5. A dual balance collaborative self-learning weighing system for alloy additions according to claim 1 wherein said first flexible connection (6) and said second flexible connection (9) are made of canvas, rubber or flexible sealing material.
  6. 6. The dual-scale collaborative self-learning weighing system for alloy charging according to claim 1, wherein the PLC control system (5) is further configured such that during self-learning, the input variables of the predictive model further include an average feed flow of the vibratory feeder (2) prior to stopping vibration to build a dynamic relationship model of air column volume-feed flow.
  7. 7. The dual scale collaborative self-learning weighing system for alloy charging of claim 1, wherein the PLC control system (5) is further configured to train historical data using a machine learning regression algorithm to update the predictive model, wherein recent data is weighted higher than early data.
  8. 8. A double-scale cooperative self-learning weighing method for alloy charging is characterized in that the double-scale cooperative self-learning weighing system for alloy charging is based on any one of claims 1-7, and comprises the following steps: s1, before starting feeding, a PLC control system (5) sends out a command to open a stop valve (8); s2, a PLC control system (5) sends out a command to start a vibrating feeder (2) to convey materials to a weighing hopper (4), and the weighing hopper monitors the weight in real time; s3, when the reading of the weighing hopper (4) reaches (target weight-vibration stopping predicted value), the PLC control system (5) sends out an instruction to stop the vibration feeder (2), and synchronously sends out an instruction to close the stop valve (8) under the micro-weighing unit (3); s4, after vibration is stopped, weighing by a micro-weighing unit (3) and obtaining the weight W1 of the retained remainder; S5, obtaining the final weight W2 after the materials in the weighing hopper (4) are stable; S6, error analysis and data recording, namely calculating a theoretical actual value G_actual=W1+ (W2-Wt) of the current air material column, wherein Wt is the reading of the weighing hopper (4) at the moment of stopping vibration in the step S3, calculating a system weighing error E=W-W2, wherein W is an initial weighing value set by the PLC control system (5), and storing the G_actual and E in a historical database after being related to the material type; S7, model self-learning and updating, namely updating an aerial material column prediction model of the corresponding material through a self-learning algorithm based on a historical database; And S8, optimizing control, namely calling the updated prediction model to output a predicted value for vibration stopping judgment in the step S3 when the material is fed next time.
  9. 9. The method of claim 8, wherein in step S3, when the reading of the weighing module of the weighing hopper (4) is close to (target weight-pre-vibration stopping set value), the PLC control system (5) controls the vibration feeder (2) to switch from high-speed feeding to low-speed or inching feeding, and then accurately stops vibrating.
  10. 10. The method of claim 8, wherein the step S4 is performed in synchronization with the closing of the first control command and the second control command, and the difference between the times of the first control command and the second control command is less than 0.5 seconds.

Description

Double-scale collaborative self-learning weighing system and method for alloy charging Technical Field The invention belongs to the technical field of ferrous metallurgy, and relates to a double-scale collaborative self-learning weighing system and method for alloy feeding. Background The external refining device of the steel-making workshop is a key device for producing high-quality steel, and various alloy additives (such as ferrosilicon, ferromanganese, aluminum and the like) need to be accurately added into molten steel in the process. The accuracy of the alloy addition directly determines the chemical composition and mechanical properties of the final product, thereby affecting the grade and cost of the steel. Currently, alloy charging systems for external refining units typically employ a "vibratory feeder + scale hopper" mode. The working process comprises the steps of starting the vibrating feeder, conveying and throwing materials in the alloy bin into a weighing hopper below the vibrating feeder, weighing the weighing hopper in real time, and sending out a command by the control system to stop the vibrating feeder when the weight is close to a target value. However, this mode has an inherent technical pain point of the "aerial stock column" problem. That is, after the vibration feeder stops vibrating, the material (air material column) which is thrown out from the feeder but not falls into the weighing hopper, and the material which falls down due to aftershock at the moment of stopping the feeder can continuously fall into the weighing hopper, so that the actual weighing value is larger than the target value set by the control system, and the weighing is out of tolerance. The accumulated effect of the errors is obvious under the working conditions of frequent feeding and small-batch weighing, so that the alloy components are controlled inaccurately, and the product quality of high-end steel is seriously affected. In the prior art, a method of stopping the vibration feeder in advance is mostly adopted to offset the drop, but the method depends on experience parameters, the drop is a dynamic variable quantity, and for materials with different granularity, different water content and different mobility, the time of stopping the vibration feeder in advance needs to be repeatedly debugged, the adaptability is poor, the precision is unstable, and the problem can not be fundamentally solved. For high value alloys or steel grades with severe requirements for composition, such errors can lead to fluctuations in alloy yield, inaccurate composition control, and even product rejection. In summary, in order to solve the deficiencies of the prior art, it is currently needed to provide a dual-weighing collaborative self-learning weighing system and method for alloy charging, which solve the problem of out-of-tolerance precision caused by the 'overhead material column' and residual shock materials. Disclosure of Invention Therefore, the invention aims to provide a double-scale collaborative self-learning weighing system and a double-scale collaborative self-learning weighing method for alloy charging, which are used for carrying out double-scale collaborative metering by introducing a micro-weighing unit and establishing a dynamic prediction model by utilizing a self-learning algorithm, so that the metering error problem caused by an aerial material column is fundamentally solved. In order to achieve the above purpose, the present invention provides the following technical solutions: The invention provides a double-scale collaborative self-learning weighing system for alloy feeding, which comprises an alloy bin, a vibrating feeder, a micro-weighing unit, a weighing hopper and a PLC control system, wherein the alloy bin is arranged above the vibrating feeder, the weighing hopper is arranged below the vibrating feeder, the micro-weighing unit is arranged between the vibrating feeder and the weighing hopper, an upper inlet of the micro-weighing unit is in sealing connection with a discharge hole of the vibrating feeder through a first flexible connecting piece, a lower outlet of the micro-weighing unit is in sealing connection with a feed hole of the weighing hopper through a second flexible connecting piece, the micro-weighing unit is provided with a weighing module and a stop valve for opening and closing the lower outlet of the micro-weighing unit, and the PLC control system is used for receiving weight signals of the micro-weighing unit and the weighing hopper, establishing a dynamic prediction model by utilizing a self-learning algorithm after collaborative metering, predicting a metering error from an air column and controlling the stop valve and the vibrating feeder. Further, the PLC control system is connected with an environmental humidity sensor and is used for acquiring environmental humidity information, taking the environmental humidity information as one of associated parameters, storing the environmen