CN-122007543-A - Inner wall quality control method, system, equipment and medium based on VCR welding

CN122007543ACN 122007543 ACN122007543 ACN 122007543ACN-122007543-A

Abstract

The application relates to the technical field of welding, and provides an inner wall quality control method, a system, equipment and a medium based on VCR welding, wherein the method comprises the steps of collecting process data in a welding process and processing the process data to obtain process information; the method comprises the steps of inputting technological process information into a preset quality prediction model to obtain an inner wall quality prediction result, comparing the inner wall quality prediction result with a preset inner wall quality target value to obtain quality deviation data, if the quality deviation data exceeds a preset deviation threshold value, constructing an optimization control frame, obtaining a current technological state, solving by the optimization control frame according to the current technological state and the quality deviation data to obtain a technological parameter adjustment sequence, generating a technological parameter adjustment instruction according to the technological parameter adjustment sequence, realizing accurate regulation and control of the forming quality of the inner wall of the VCR joint, and ensuring uniformity of the inner wall quality of the circumferential weld and high-standard purity requirements.

Inventors

MAO CHAOBIN
LIU XIN
XU JUN
PENG DINGQIANG
XIE LIHUA

Assignees

季华实验室

Dates

Publication Date: 20260512
Application Date: 20260410

Claims (10)

1. The inner wall quality control method based on VCR welding is characterized by comprising the following steps: Acquiring process data in the welding process and processing the process data to obtain process information; Inputting the technological process information into a preset quality prediction model to obtain an inner wall quality prediction result; comparing the inner wall quality prediction result with a preset inner wall quality target value to obtain quality deviation data; If the quality deviation data exceeds a preset deviation threshold value, an optimization control framework is constructed; acquiring a current process state, and solving by using an optimization control framework according to the current process state and quality deviation data to obtain a process parameter adjustment sequence; And generating a process parameter adjustment instruction according to the process parameter adjustment sequence.
2. The method for controlling the quality of an inner wall based on VCR welding as set forth in claim 1, wherein the steps of collecting process data during welding and processing the process data to obtain process information comprise: Collecting a molten pool area image sequence in the welding process, and extracting the molten pool area image sequence to obtain visual characteristics of a molten pool; acquiring an arc sound signal and a welding gun vibration signal in the welding process, and processing the arc sound signal and the welding gun vibration signal to obtain an acoustic vibration fusion characteristic; collecting welding current, welding voltage, welding speed and real-time angle of a welding gun in the welding process, and converting the real-time angle of the welding gun to obtain position characteristics; and integrating visual characteristics, sound and vibration fusion characteristics, welding current, welding voltage, welding speed and position characteristics of the molten pool to obtain technological process information.
3. The method for controlling the quality of an inner wall based on VCR welding as set forth in claim 1, wherein the step of inputting the process information into a preset quality prediction model to obtain an inner wall quality prediction result comprises the following steps: Intercepting the technological process information according to a preset time length to obtain a time sequence characteristic sequence; Inputting the time sequence characteristic sequence into a preset quality prediction model to obtain an inner wall depression depth predicted value and an inner wall oxidation risk index predicted value; and integrating the predicted value of the depth of the concave inner wall and the predicted value of the oxidation risk index of the inner wall to obtain the predicted result of the quality of the inner wall.
4. The method for controlling the quality of an inner wall based on VCR welding as set forth in claim 1, wherein if the quality deviation data exceeds a preset deviation threshold, constructing an optimized control framework, specifically comprising: Setting adjustment range constraints of a prediction time domain, a control time domain and process parameters, and taking a quality prediction model as an internal prediction model; Setting an optimization function aiming at minimizing the mass deviation of the inner wall and the adjustment amplitude of the process parameters; and constructing an optimization control framework based on the prediction time domain, the control time domain, the adjustment range constraint, the optimization function and the internal prediction model.
5. The method for controlling the quality of the inner wall based on VCR welding as set forth in claim 1, wherein the step of obtaining the current process state and solving the current process state and the quality deviation data by using an optimization control framework to obtain a process parameter adjustment sequence comprises the following steps: Acquiring process information and an inner wall quality prediction result at the current moment to serve as a current process state; Inputting the current process state and quality deviation data into an optimization control framework to solve and obtain a plurality of process parameter adjustment amounts; And arranging and integrating the plurality of process parameter adjustment amounts to obtain a process parameter adjustment sequence.
6. The method for controlling the quality of an inner wall based on VCR welding as set forth in claim 1, wherein the generating the process parameter adjustment command according to the process parameter adjustment sequence comprises: Extracting a first adjustment amount in a process parameter adjustment sequence, wherein the adjustment amount comprises a welding current adjustment amount and a welding speed adjustment amount; converting the welding current adjustment amount and the welding speed adjustment amount to obtain a welding current adjustment instruction and a welding speed adjustment instruction; and generating a process parameter adjustment instruction according to the welding current adjustment instruction and the welding speed adjustment instruction.
7. The method for controlling the quality of an inner wall based on VCR welding as set forth in claim 1, further comprising, after generating the process parameter adjustment command according to the process parameter adjustment sequence: calculating for a plurality of times by utilizing a quality prediction model in a preset time period to obtain a plurality of prediction results; Calculating to obtain the variance of the multiple prediction results based on the multiple prediction results, and taking the variance of the multiple prediction results as uncertainty measurement; if the uncertainty measure exceeds a preset uncertainty threshold, marking the process data in the corresponding time period as a high-value sample; Acquiring actual inner wall quality detection data corresponding to the high-value sample, and forming a new training sample by the high-value sample and the corresponding actual inner wall quality detection data; And performing incremental training on the quality prediction model by using the new training sample so as to optimize the quality prediction model.
8. The inner wall quality control system based on VCR welding is characterized by comprising: the acquisition module is used for acquiring the technical process data in the welding process and processing the technical process data to obtain technical process information; The prediction module is used for inputting the technological process information into a preset quality prediction model so as to obtain an inner wall quality prediction result; the comparison module is used for comparing the inner wall quality prediction result with a preset inner wall quality target value to obtain quality deviation data; the construction module is used for constructing an optimization control frame if the quality deviation data exceeds a preset deviation threshold value; The solving module is used for obtaining the current process state, and solving the current process state and the quality deviation data by utilizing the optimization control framework to obtain a process parameter adjustment sequence; and the generating module is used for generating a process parameter adjusting instruction according to the process parameter adjusting sequence.
9. The inner wall quality control device based on VCR welding is characterized by comprising a memory and at least one processor, wherein instructions are stored in the memory; At least one of the processors invokes the instructions in the memory to cause the inner wall quality control device to perform the steps of the VCR welding-based inner wall quality control method as recited in any of claims 1-7.
10. A computer readable storage medium having instructions stored thereon, which when executed by a processor, perform the steps of the VCR welding-based inner wall quality control method according to any of claims 1-7.

Description

Inner wall quality control method, system, equipment and medium based on VCR welding Technical Field The application relates to the technical field of welding, in particular to an inner wall quality control method, system, equipment and medium based on VCR welding. Background In the manufacturing process of semiconductor chips, the conveying system of special gases such as silane, phosphane, arsine and the like has extremely high requirements on the air tightness and purity of pipeline connection. VCR (face seal) joints are widely used in the semiconductor te pipe system for their excellent sealing performance and anti-vibration capability. The core of the quality is the forming quality of the inner wall of the welding seam, and the inner wall is required to be free of pits and extremely thin in oxide layer. The existing VCR welding quality control method is mainly used for executing welding according to a preset technological parameter curve, and cannot be adjusted according to state changes in the actual welding process. The partial improvement method adopts feedback control based on a single sensor such as molten pool vision, arc voltage and the like, adjusts the welding process by observing the external molten pool form or arc electrical parameters, has limited information content of the single sensor, and cannot effectively regulate and control the quality index of the inner wall which cannot be directly observed. Disclosure of Invention The present application aims to improve at least one technical problem in the background art. The application provides an inner wall quality control method based on VCR welding, which comprises the steps of collecting process data in the welding process and processing the process data to obtain process information; Inputting the technological process information into a preset quality prediction model to obtain an inner wall quality prediction result; comparing the inner wall quality prediction result with a preset inner wall quality target value to obtain quality deviation data; If the quality deviation data exceeds a preset deviation threshold value, an optimization control framework is constructed; Acquiring a current process state, and solving by using an optimization control framework according to the current process state and quality deviation data to obtain a process parameter adjustment sequence; And generating a process parameter adjustment instruction according to the process parameter adjustment sequence. According to some technical schemes of the application, the process data in the welding process are collected and processed to obtain the process information, which comprises the following steps: Collecting a molten pool area image sequence in the welding process, and extracting the molten pool area image sequence to obtain visual characteristics of a molten pool; acquiring an arc sound signal and a welding gun vibration signal in the welding process, and processing the arc sound signal and the welding gun vibration signal to obtain an acoustic vibration fusion characteristic; collecting welding current, welding voltage, welding speed and real-time angle of a welding gun in the welding process, and converting the real-time angle of the welding gun to obtain position characteristics; and integrating visual characteristics, sound and vibration fusion characteristics, welding current, welding voltage, welding speed and position characteristics of the molten pool to obtain technological process information. According to some embodiments of the present application, the inputting the process information into a preset quality prediction model to obtain an inner wall quality prediction result specifically includes: Intercepting the technological process information according to a preset time length to obtain a time sequence characteristic sequence; Inputting the time sequence characteristic sequence into a preset quality prediction model to obtain an inner wall depression depth predicted value and an inner wall oxidation risk index predicted value; and integrating the predicted value of the depth of the concave inner wall and the predicted value of the oxidation risk index of the inner wall to obtain the predicted result of the quality of the inner wall. According to some embodiments of the present application, if the quality deviation data exceeds a preset deviation threshold, an optimization control framework is constructed, which specifically includes: Setting adjustment range constraints of a prediction time domain, a control time domain and process parameters, and taking a quality prediction model as an internal prediction model; Setting an optimization function aiming at minimizing the mass deviation of the inner wall and the adjustment amplitude of the process parameters; and constructing an optimization control framework based on the prediction time domain, the control time domain, the adjustment range constraint, the optimization function and the internal p