CN-121696590-B - Welding process state monitoring system and method based on multi-source sensing

CN121696590BCN 121696590 BCN121696590 BCN 121696590BCN-121696590-B

Abstract

The application provides a state monitoring system and method for a welding process based on multi-source sensing, which are used for synchronously collecting multi-source welding data in the welding process, constructing a welding state feature set according to the multi-source welding data, carrying out transition characterization on the stability change trend of the welding process according to the welding state feature set to obtain a state transition feature space, carrying out evolution modeling on the state transition feature space to obtain a welding state evolution track of the welding process, identifying a state type of the welding process based on the welding state evolution track, evaluating the deviation degree of the welding process under the state type, predicting the deviation degree to obtain abnormal risk probability, carrying out real-time monitoring and early warning on the welding process according to the abnormal risk probability, and outputting a welding parameter optimization instruction. According to the technical scheme provided by the application, the stability change trend of the welding process can be described based on the multi-source sensing information, so that the welding state type identification and abnormal risk prediction are realized, and the reliability of the welding process is improved.

Inventors

Wu Rigenbayila
LI HONG
AN PUGUANG
WANG YURONG
ZENG XIULI
HAN WENHUA

Assignees

包头职业技术学院

Dates

Publication Date: 20260505
Application Date: 20260209

Claims (9)

1. A welding process state monitoring method based on multi-source sensing is characterized by comprising the following steps: Synchronously acquiring multi-source welding data in a welding process, and constructing a welding state characteristic set in the welding process according to the multi-source welding data; Performing transition characterization on the stability change trend of the welding process according to the welding state feature set to obtain a state transition feature space, performing weighted fusion on state transition feature sequences corresponding to all the welding state features in the state transition feature space according to the credibility weight of the corresponding sensing channel to obtain a fusion feature sequence, and mapping the fusion feature sequence into a welding state evolution track of the welding process through a long-short-time memory network, wherein the welding state evolution track is a state path sequence for describing the continuous change rule of the welding state along with time; Recognizing a welding state type based on the welding state evolution track, evaluating the deviation degree of a welding process under the state type, constructing a mapping model between the historical deviation degree and abnormal risk probability based on historical welding sample data and marked abnormal event information, and inputting the deviation degree into the mapping model for prediction to obtain the abnormal risk probability of the welding process; And carrying out real-time monitoring and early warning on the welding process according to the abnormal risk probability, and outputting a welding parameter optimization instruction.
2. The method of claim 1, wherein the multi-source welding data includes welding current data, puddle infrared temperature field data, puddle high speed visual images, gun and workpiece vibration signals.
3. The method of claim 1, wherein the set of welding state characteristics includes a current average, an arc fluctuation index, a short circuit duty cycle, a weld pool area, a spatter density, a weld pool center temperature, a cooling rate, and a vibration RMS value during welding.
4. The method for monitoring the state of a welding process based on multi-source sensing according to claim 1, wherein the step of performing transition characterization on the stability variation trend of the welding process according to the welding state feature set to obtain a state transition feature space specifically comprises: For each type of welding state feature in the welding state feature set, acquiring sequence data corresponding to the welding state feature; Performing adjacent transition measurement on the sequence data to obtain a state transition characteristic sequence corresponding to the welding state characteristics, and further obtaining a state transition characteristic sequence corresponding to each type of welding state characteristics; And constructing a state transition feature space through state transition feature sequences corresponding to all the welding state features.
5. The method of monitoring a state of a welding process based on multi-source sensing of claim 1, further comprising: acquiring the signal-to-noise ratio and the historical stability contribution degree of each sensing channel in the welding process; and determining the credibility weight of each sensing channel based on the corresponding signal-to-noise ratio and the historical stability contribution degree.
6. The method of claim 1, wherein identifying a state class of a weld based on the weld state evolution trajectory comprises: extracting a plurality of classification features from the welding state evolution track; all classification features are input into a Softmax classifier for classification, and then the state category of welding is identified.
7. A multi-source sensing based welding process state monitoring system for performing a multi-source sensing based welding process state monitoring method as defined in any one of claims 1 to 6, comprising: the characteristic construction module is used for synchronously collecting multi-source welding data in the welding process and constructing a welding state characteristic set in the welding process according to the multi-source welding data; the evolution modeling module is used for carrying out transition characterization on the stability change trend of the welding process according to the welding state feature set to obtain a state transition feature space, carrying out evolution modeling by using the state transition feature space, and further obtaining a welding state evolution track of the welding process; The probability prediction module is used for identifying the state type of welding based on the welding state evolution track, evaluating the deviation degree of the welding process under the state type, and predicting the abnormal risk probability of the welding process through the deviation degree; and the early warning feedback module is used for carrying out real-time monitoring and early warning on the welding process according to the abnormal risk probability and outputting a welding parameter optimization instruction.
8. A computer device comprising a memory storing code and a processor configured to obtain the code and to perform the multisource sensing based welding process state monitoring method of any of claims 1 to 6.
9. A computer readable storage medium storing a computer program, characterized in that the computer program, when executed by a processor, implements the multisource sensing based welding process state monitoring method according to any one of claims 1 to 6.

Description

Welding process state monitoring system and method based on multi-source sensing Technical Field The application relates to the technical field of welding state monitoring, in particular to a welding process state monitoring system and method based on multi-source sensing. Background The stability and welding quality of a welding process, which is a typical high energy density, strong non-linear and strong disturbance industrial process, are highly dependent on the synergistic effect of various factors such as arc state, molten pool thermal behavior, weld seam forming process and mechanical motion stability. The existing welding process monitoring technology mainly adopts a single sensor or a small amount of sensing information to monitor, for example, the state judgment is carried out only based on current and voltage signals, temperature signals or visual images, the problems of single information dimension, weak anti-interference capability and poor adaptability to complex working conditions exist, meanwhile, the existing technology mainly adopts static threshold judgment or transient state analysis, lacks the system modeling capability of dynamic evolution characteristics and stability change trend of the welding process, is difficult to accurately describe the process mechanism of the welding from stable to abnormal evolution, and is easy to generate misjudgment and missed judgment. Therefore, how to characterize the stability variation trend of the welding process based on the multi-source sensing information, so as to realize the welding state type identification and abnormal risk prediction, so as to improve the reliability of the welding process is a difficult problem faced by the prior art. Disclosure of Invention The application provides a welding process state monitoring system and method based on multi-source sensing, which can be used for describing the stability change trend of a welding process based on multi-source sensing information, so that welding state type identification and abnormal risk prediction are realized, and the reliability of the welding process is improved. In a first aspect, the present application provides a method for monitoring a state of a welding process based on multi-source sensing, comprising the steps of: Synchronously acquiring multi-source welding data in a welding process, and constructing a welding state characteristic set in the welding process according to the multi-source welding data; Performing transition characterization on the stability change trend of the welding process according to the welding state feature set to obtain a state transition feature space, and performing evolution modeling by using the state transition feature space to further obtain a welding state evolution track of the welding process; recognizing a welding state type based on the welding state evolution track, evaluating the deviation degree of a welding process under the state type, and predicting the abnormal risk probability of the welding process through the deviation degree; And carrying out real-time monitoring and early warning on the welding process according to the abnormal risk probability, and outputting a welding parameter optimization instruction. In some embodiments, the multi-source welding data includes welding current data, puddle infrared temperature field data, puddle high speed visual images, gun and workpiece vibration signals. In some embodiments, the set of welding state characteristics includes a current average, an arc fluctuation index, a short circuit duty cycle, a puddle area, a spatter density, a puddle center temperature, a cooling rate, and a vibration RMS value during welding. In some embodiments, performing transition characterization on the stability variation trend of the welding process according to the welding state feature set, and obtaining a state transition feature space specifically includes: For each type of welding state feature in the welding state feature set, acquiring sequence data corresponding to the welding state feature; Performing adjacent transition measurement on the sequence data to obtain a state transition characteristic sequence corresponding to the welding state characteristics, and further obtaining a state transition characteristic sequence corresponding to each type of welding state characteristics; And constructing a state transition feature space through state transition feature sequences corresponding to all the welding state features. In some embodiments, further comprising: acquiring the signal-to-noise ratio and the historical stability contribution degree of each sensing channel in the welding process; and determining the credibility weight of each sensing channel based on the corresponding signal-to-noise ratio and the historical stability contribution degree. In some embodiments, performing evolution modeling using the state transition feature space, and further obtaining a welding state evolution track of the