CN115570228B - Intelligent feedback control method and system for welding pipeline gas supply - Google Patents
Intelligent feedback control method and system for welding pipeline gas supply Download PDFInfo
- Publication number
- CN115570228B CN115570228B CN202211467931.3A CN202211467931A CN115570228B CN 115570228 B CN115570228 B CN 115570228B CN 202211467931 A CN202211467931 A CN 202211467931A CN 115570228 B CN115570228 B CN 115570228B
- Authority
- CN
- China
- Prior art keywords
- clustering
- pipeline
- welding
- data
- feedback control
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B23—MACHINE TOOLS; METAL-WORKING NOT OTHERWISE PROVIDED FOR
- B23K—SOLDERING OR UNSOLDERING; WELDING; CLADDING OR PLATING BY SOLDERING OR WELDING; CUTTING BY APPLYING HEAT LOCALLY, e.g. FLAME CUTTING; WORKING BY LASER BEAM
- B23K3/00—Tools, devices, or special appurtenances for soldering, e.g. brazing, or unsoldering, not specially adapted for particular methods
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B23—MACHINE TOOLS; METAL-WORKING NOT OTHERWISE PROVIDED FOR
- B23K—SOLDERING OR UNSOLDERING; WELDING; CLADDING OR PLATING BY SOLDERING OR WELDING; CUTTING BY APPLYING HEAT LOCALLY, e.g. FLAME CUTTING; WORKING BY LASER BEAM
- B23K3/00—Tools, devices, or special appurtenances for soldering, e.g. brazing, or unsoldering, not specially adapted for particular methods
- B23K3/08—Auxiliary devices therefor
Abstract
The invention discloses a feedback control method and a feedback control system for welding pipeline gas supply, which are characterized in that firstly, data of each sensor in a welding process are collected through a multi-dimensional sensor, current, voltage and wire feeding speed data are selected, clustering is carried out through a clustering algorithm, the number of clusters is iterated, and the optimal cluster number is selected based on real welding conditions; training a clustering model based on the optimal clustering number, and giving a gas supply speed interval of each type of data standard pipeline; finally, matching a standard pipeline gas supply speed interval according to multi-dimensional sensing data acquired by a sensor in real time, and finishing feedback control; the method is based on a clustering algorithm model, combines multi-dimensional sensing data acquired in real time, iterates clustering numbers, designs a root mean square error index capable of reflecting actual welding conditions on the basis of a traditional evaluation index as a clustering evaluation index, and can select the clustering number closest to the actual condition in the iterative clustering number process.
Description
Technical Field
The invention belongs to the technical field of intelligent welding, and particularly relates to an intelligent feedback control method and system for welding pipeline gas supply.
Background
With the continuous development of welding process, more and more large-scale factories provide welding shielding gas with stable flow rate for a welding robot in a pipeline gas supply mode in the gas shielding welding process. Compare in traditional bottled air feed, the pipeline air feed is compared in traditional bottled air feed, has more concentrated, more stable advantage, but has the actual problem that the open air overshoots, the gas of turning off spills over equally, and these problems have directly led to the pipeline air feed and have had a large amount of extravagant circumstances, consequently need an effectual pipeline air feed feedback control method, carry out real-time feedback and effective control to the problem that probably exists in the pipeline air feed process.
With the continuous development of the internet of things technology, multidimensional sensing data are collected through different types of sensors, so that the real-time monitoring of the welding process becomes possible, the pipeline gas flow velocity in the welding process is analyzed and controlled in real time by using the data collected by the sensors, and a direction is provided for the real-time feedback control of the gas supply of the welding pipeline.
Chinese patent CN114951981A (published japanese 20200830) provides an automatic control method for laser welding shielding gas, wherein a core control module consisting of a proportional solenoid valve, a gas flow sensor and a micro control unit realizes automatic regulation control of laser welding shielding gas; according to the invention, the automatic matching and the automatic control of the flow of the protective gas are realized according to the laser output energy in the welding process, the consumption of the welding protective gas can be reduced, the welding production and manufacturing cost consumption is reduced, and the gas saving effect can be realized while the protection effect and the welding quality of a welding area are not influenced.
Chinese patent CN111098000A (published japanese 20200505) provides a flow control method for shielding gas in welding equipment, which actively controls the flow of added gas by checking the pressure in a tube through a pressure switch; the welding current is detected by the current transformer to automatically match with the gas flow, and the gas flow is controlled by a computer or a PLC analog quantity module and the opening of a valve body of the flow control valve is controlled in a voltage mode so as to control the gas flow; and when the gas flow and the welding current deviate from the set values, judging that the welding current is abnormal, and outputting an abnormal signal.
The above invention does not provide a method for feedback control of duct air supply from the perspective of sensor data analysis.
Disclosure of Invention
The purpose of the invention is as follows: aiming at the problems and ideas in the background art, the invention provides an intelligent feedback control method and system for welding pipeline gas supply. And giving out standard gas supply intervals of the gas supply of the welding pipeline under different types of conditions through the trained clustering model, and taking the standard gas supply intervals as feedback control bases. In addition, in order to save data transmission cost, the invention also designs a feedback control model based on reinforcement learning, firstly, the multidimensional sensing data is judged through the feedback control model, and when the judgment needs to execute feedback control, a standard pipeline air supply speed interval is output through the clustering model and is used for adjusting the pipeline air supply speed.
The technical scheme is as follows:
an intelligent feedback control method for welding pipeline gas supply comprises the following steps:
s1, collecting various sensor data in a welding process through a multi-dimensional sensor, and preprocessing the sensor data; the preprocessing comprises outlier culling and missing value filling; the sensor data comprises current, voltage, wire feeding speed and pipeline gas supply speed, wherein each sensor data comprises a uniquely determined index corresponding to the current, voltage, wire feeding speed and pipeline gas supply speed;
s2, selecting the current, voltage and wire feeding speed data acquired in the step S1, clustering by a clustering algorithm, iteratively clustering the number, and selecting the optimal clustering number based on the real welding condition;
s3, training a clustering model based on the optimal clustering number, and endowing each type of data with a standard pipeline gas supply speed interval;
s4, a feedback control model is built based on reinforcement learning, the feedback control model inputs four state data of current, voltage, wire feeding speed and pipeline air supply speed and outputs execution actions, and the execution actions comprise execution control and non-execution control; when the output is execution control, transmitting data of each sensor to the clustering model trained in the step S3, and acquiring and outputting a matched standard pipeline air supply speed interval; when the actual air supply speed of the pipeline exceeds the upper limit, the speed is adjusted through the controller, and when the actual air supply speed of the pipeline is lower than the lower limit, an alarm signal is sent out; when the output is not control execution, the data is not transmitted.
Further, in the step S1, abnormal values in the sensor data are removed by using a box chart method.
Further, the method for supplementing the missing value in step S1 is as follows:
s1.1, counting the welding time of each welding line in the same process based on the same number of the welding lines needing to be welded in the same process, and recording the counting as
Wherein n represents the nth weld;
step S1.2, when sensing occursWhen the device data is missing, the welding seam of the missing part is judged based on the time length set in the step S1.1; when the missing value belongs to the missing of the continuous time period, recording the distance S between the missing starting time point and the starting time of the corresponding welding seam and the duration L of the missing value as positioning information, and recording the positioning information as
(ii) a When the missing value belongs to the single point missing, recording the distance between the missing time point and the initial time point of the welding seam as positioning information, and recording the distance as the positioning information
；
And S1.3, recording time sequence data of the welding process of another identical workpiece, determining the same position of the identical welding seam according to the welding time length set and the positioning information, and finding corresponding data to be filled to the missing value.
Further, the method for selecting the optimal number of clusters in step S2 includes the following steps:
s2.1, dividing the current, the voltage and the wire feeding speed with indexes into a training set dfTrain and a testing set dfTest;
s2.2, based on the number of the dfTracin iterative clustering of the training set, and correspondingly finding out the gas supply speeds of the pipelines with the same number according to a plurality of indexes contained in each class; the air supply speed of the pipeline in each class is counted and the average value is calculated and recorded as
；
S2.3, in the process of each iterative clustering number, predicting the clustering category of the test set dfTest through a trained clustering algorithm; selecting the air supply speed of the pipeline corresponding to all indexes in each class, and recording the air supply speed as
，
，
Representing the number of indexes in the test set; selecting the optimal clustering number by taking the root mean square error Gasrmse as an evaluation index, wherein the optimal clustering number is specifically represented as follows:
and S2.4, iterating the clustering number, calculating the root mean square error Gasrmse in each iteration process, and finding out the corresponding class number when the value of the Gasrmse is minimum, namely the optimal clustering number.
Further, in step S3, the model is retrained based on the training set dfTrain, and 1/4 digit and 3/4 digit of the air supply speed of the pipeline in each category are calculated, so that the standard interval of the air supply speed of the pipeline in each category is represented as: [1/4 bit number, 3/4 bit number ].
Further, the method for building the feedback control model in step S4 includes:
s4.1, creating a control environment, wherein the control environment comprises four corresponding state data, namely current, voltage, wire feeding speed and pipeline air supply speed; labeling the output execution action, wherein the execution control is 1, and the non-execution control is 0;
s4.2, formulating a reward function R according to different actions; the method comprises the following specific steps:
wherein action represents output action, v represents air supply speed of the pipeline at the current moment,
representing the average value of the air supply speed of the pipeline at the previous t moment; when the output action is not executing control, action =0, and the reward function R =0; when the output action is execution control, action =1, and the reward function R =
；
S4.3, establishing a Q network model, wherein the Q network model comprises 1 layer of gru and not less than 1 layer of linear layer, and the Relu function is selected as an activation function; the input of the Q network model is a control environment, the output is an execution action, and the execution action with the maximum softmax probability is selected as the output;
and S4.4, inputting training set data, training the reinforced model established in the step S4.3, solving the sum of reward functions R generated by selecting the execution action in each training, and finally selecting the execution action strategy corresponding to the maximum reward function R as a Q network model parameter.
An intelligent feedback control method and system for welding pipeline gas supply comprises a multidimensional sensor data acquisition system, a cloud platform and a pipeline gas supply feedback control system; the multi-dimensional sensor data acquisition system acquires multi-dimensional sensing data in real time, a clustering model and a feedback control model are deployed on a cloud platform, the multi-dimensional sensing data are transmitted to the feedback control model to judge whether feedback control needs to be executed or not, and when the feedback control needs to be executed, the multi-dimensional sensing data are transmitted to the clustering model to acquire and control a standard pipeline air supply speed interval; and when the control is not executed, the multi-dimensional sensing data is not transmitted.
Compared with the prior art, the technical scheme adopted by the invention has the following beneficial effects:
(1) Aiming at the acquired real-time multi-dimensional sensing data, the invention provides a preprocessing method for removing abnormal values and supplementing missing values, and the continuous availability of the sensing data is ensured.
(2) The method is based on a clustering algorithm model, combines multidimensional sensing data acquired in real time, iterates the clustering number, designs a root mean square error index which can reflect the actual welding condition on the basis of the traditional evaluation index as a clustering evaluation index, and can select the clustering number which is closest to the actual condition in the process of iterating the clustering number.
(3) The invention further designs the standard pipeline air supply speed interval of each category through the trained clustering model, and recommends the standard pipeline air supply speed interval to the pipeline air supply feedback control system, thereby realizing the real-time detection, feedback and control of the pipeline air supply speed.
(4) Because the pipeline gas supply speed in the actual welding process is not frequently changed, the change frequency of the standard pipeline gas supply speed interval recommended by the clustering model is not high, if the recommended interval is continuously output in the actual data transmission process, a large amount of redundant data can be generated, and the data transmission cost is greatly improved. Aiming at the problem, the invention designs a matched feedback control model, obtains a Q network model based on reinforcement learning, judges the acquired sensor data through the model, controls a clustering model to output a standard pipeline gas supply speed interval when judging that the feedback control is needed, and performs specific control according to the corresponding interval. The model is established, so that the data transmission cost is greatly reduced, and accurate feedback control can be effectively realized.
Drawings
FIG. 1 is a block diagram of a welding line gas feed feedback control system provided by the present invention;
FIG. 2 is a diagram of a clustering model framework provided by the present invention;
FIG. 3 is a schematic diagram of a feedback control model building process provided by the present invention.
Detailed Description
The invention is further explained below with reference to the drawings.
The invention provides a welding pipeline gas supply feedback control method aiming at the problems of unstable gas supply, overshooting when gas is opened and gas is closed and overflowing in the traditional welding process, and the welding pipeline gas supply feedback control method is controlled by a welding pipeline gas supply feedback control system shown in figure 1. The system mainly comprises a multi-dimensional sensor data acquisition system, a cloud platform and a pipeline air supply feedback control system. The multi-dimensional sensor data acquisition system acquires multi-dimensional sensing data in real time, the cloud platform is used for deploying a clustering model and a feedback control model, the multi-dimensional sensing data are firstly input into the feedback control model for judgment, when the judgment that feedback control needs to be executed, the control data are input into the clustering model, and a standard pipeline air supply speed interval is recommended to the pipeline air supply feedback control system. The pipeline gas supply feedback control system judges whether the current actual pipeline gas supply speed exceeds the limit or not, and performs corresponding control and alarm through the control module. The working principle of each module of the system is described in detail in conjunction with the specific embodiment.
S1, collecting various sensor data in the welding process in real time through a multi-dimensional sensor data collecting system, and preprocessing the sensor data. Preprocessing includes outlier culling and missing value padding. In the embodiment, a box graph method is adopted to remove abnormal values in the sensor data, and a filling method is adopted to fill missing values caused by network delay of acquisition equipment and the like in the acquisition process. The specific method comprises the following steps:
s1.1, counting the welding time of each welding line in the same process based on the same number of the welding lines needing to be welded in the same process, and recording the counting as
Wherein n represents the nth weld;
s1.2, when the data of the sensor is lost, judging the welding line of the lost part based on the time length set in the step S1.1; when the missing value belongs to the missing of the continuous time period, recording the distance S between the missing starting time point and the starting time of the corresponding welding seam and the duration L of the missing value as positioning information, and recording the positioning information as
(ii) a When the missing value belongs to the single point missing, recording the distance between the missing time point and the initial time point of the welding seam as positioning information, and recording the distance as the positioning information
；
And S1.3, recording time sequence data of the welding process of another same workpiece, determining missing values according to the welding time length set and the positioning information, finding the same position of the same welding line, and filling the corresponding data to the missing values.
The sensor data collected in real time in this embodiment includes current, voltage, wire feed speed, and pipeline gas supply speed, where each piece of sensor data includes a uniquely determined index corresponding to the current, voltage, wire feed speed, and pipeline gas supply speed. On the basis, the embodiment provides an improved clustering algorithm model for providing a standard pipeline gas supply speed interval. In particular, the amount of the solvent to be used,
s2.1, dividing the current, the voltage and the wire feeding speed with indexes into a training set dfTrain and a testing set dfTest;
s2.2, based on the number of the dfTracin iterative clustering of the training set, and correspondingly finding out the gas supply speeds of the pipelines with the same number according to a plurality of indexes contained in each class; the air supply speed of the pipeline in each class is counted and the average value is calculated and recorded as
；
S2.3, in the process of each iterative clustering number, predicting the clustering category of the test set dfTest through a trained clustering algorithm; selecting the air supply speed of the pipeline corresponding to all indexes in each class, and recording the air supply speed as
，
，
Representing the number of indexes in the test set; selecting the optimal clustering number by taking root mean square error Gasrmse as an evaluation index, wherein the optimal clustering number is specifically expressed as follows:
and S2.4, iterating the clustering number, calculating the root mean square error Gasrmse in each iteration process, and finding out the corresponding class number when the value of the Gasrmse is minimum, namely the optimal clustering number.
In this embodiment, the finally selected optimal clustering number is the clustering condition closest to the actual welding condition, and in this case, the input of the corresponding current, voltage and wire feeding speed can more accurately recommend the standard pipeline gas supply speed interval for the subsequent feedback control.
S3, after the optimal clustering number is obtained, retraining the model based on the training set dfTrain, and calculating 1/4 digit number and 3/4 digit number of the air supply speed of the pipeline in each category, wherein the standard pipeline air supply speed interval in each category is expressed as: [1/4 bit number, 3/4 bit number ]. The 1/4 digit number and the 3/4 digit number are respectively selected as the upper limit and the lower limit of the standard pipeline gas supply speed interval, the optimal range of the recommended pipeline gas supply speed can be more accurately represented, the given interval is closer to the actual welding working condition, and the standard welding process is better met.
And S4, in the actual welding process, the gas supply speed of the pipeline is not frequently changed, the change frequency of the standard pipeline gas supply speed interval recommended by the clustering model is not high, if the recommended interval is continuously output in the actual data transmission process, a large amount of redundant data can be generated, and the data transmission cost is greatly increased. This embodiment has further designed feedback control model in order to solve accurate feedback control's problem. As shown in fig. 3, a feedback control model is built based on reinforcement learning, the feedback control model inputs four state data of current, voltage, wire feeding speed and pipeline air supply speed, and outputs execution actions, wherein the execution actions include execution control and non-execution control; when the output is execution control, transmitting the data of each sensor to the clustering model trained in the step S3, and acquiring and outputting a matched standard pipeline gas supply speed interval; when the actual air supply speed of the pipeline exceeds the upper limit, the speed is adjusted through the controller, and when the actual air supply speed of the pipeline is lower than the lower limit, an alarm signal is sent out; when the output is not executing control, the data is not transmitted. In particular, the amount of the solvent to be used,
s4.1, creating a control environment which comprises four corresponding state data, namely current, voltage, wire feeding speed and pipeline air supply speed; labeling the output execution action, wherein the execution control is 1, and the non-execution control is 0;
s4.2, formulating a reward function R according to different actions; the method comprises the following specific steps:
wherein action represents output action, v represents air supply speed of the pipeline at the current moment,
representing the average value of the air supply speed of the pipeline at the previous t moment; when the output action is not executing control, action =0, and the reward function R =0; when the output action is execution control, action =1, and the reward function R =
；
S4.3, establishing a Q network model, wherein the Q network model comprises 1 layer of gru and not less than 1 layer of linear layer, and the activation function is a Relu function; and the input of the Q network model is a control environment, the output is an execution action, and the execution action with the maximum softmax probability is selected as the output.
And S4.4, inputting training set data, training the reinforced model established in the step S4.3, solving the sum of reward functions R generated by selecting the execution action in each training, and finally selecting the execution action strategy corresponding to the maximum reward function R as a Q network model parameter.
The foregoing is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and amendments can be made without departing from the principle of the present invention, and these modifications and amendments should also be considered as the protection scope of the present invention.
Claims (3)
1. An intelligent feedback control method for welding pipeline gas supply is characterized by comprising the following steps:
s1, collecting data of various sensors in a welding process through a multi-dimensional sensor, and preprocessing the data of the sensors; the preprocessing comprises outlier culling and missing value filling; the sensor data comprises current, voltage, wire feeding speed and pipeline gas supply speed, wherein each sensor data comprises a uniquely determined index corresponding to the current, voltage, wire feeding speed and pipeline gas supply speed; the method of supplementing the missing values is as follows:
s1.1, counting the welding time of each welding line in the same process based on the same number of the welding lines needing to be welded in the same process, and recording the counting as [ T ] 1 ，T 2 ，...，T n ]Wherein n represents the nth weld;
s1.2, when the data of the sensor is lost, judging the welding line of the lost part based on the time length set in the step S1.1; when the missing value belongs to the missing of the continuous time period, recording the distance S between the missing starting time point and the starting time of the corresponding welding seam and the duration L of the missing value as positioning information, and recording as S, L](ii) a When the missing value belongs to the single point missing, recording the distance between the missing time point and the initial time point of the welding seam as positioning information, and recording as S p ；
S1.3, recording time sequence data of a welding process of another same workpiece, determining a missing value according to a welding time length set and positioning information, finding the same position of the same welding line, finding corresponding data and filling the corresponding data to the missing value;
s2, selecting the current, voltage and wire feeding speed data acquired in the step S1, clustering by a clustering algorithm, iteratively clustering the number, and selecting the optimal clustering number based on the real welding condition; the method for selecting the optimal cluster number comprises the following steps:
s2.1, dividing the current, the voltage and the wire feeding speed with indexes into a training set dfTrain and a testing set dfTest;
s2.2, based on the number of the dfTracin iterative clustering of the training set, and correspondingly finding out the gas supply speeds of the pipelines with the same number according to a plurality of indexes contained in each class; counting the air supply speed of the pipeline in each class, and calculating an average value which is recorded as Gasmean;
s2.3, in the process of each iterative clustering number, predicting the clustering category of the test set dfTest through a trained clustering algorithm; selecting the air supply speed of the pipeline corresponding to all indexes in each class and recording the air supply speed as dfTestGas i ，i∈[1，len(dfTest)]Len (dfTest) represents the number of indexes in the test set; selecting the optimal clustering number by taking the root mean square error Gasrmse as an evaluation index, wherein the optimal clustering number is specifically represented as follows:
dfTestGas i -Gasmean type represents each of the dfTestGas i Making a difference with Gasmean of the cluster type;
s2.4, iterating the number of clusters, calculating root mean square error Gasrmse in each iteration process, and finding out the corresponding class number when the value of Gasrmse is minimum, namely the optimal number of clusters;
s3, training a clustering model based on the optimal clustering number, and endowing each type of data with a standard pipeline gas supply speed interval; specifically, the model is retrained based on the training set dfTrain, and 1/4 digit number and 3/4 digit number of the air supply speed of the pipeline in each category are calculated, so that the standard interval of the air supply speed of the pipeline in each category is represented as: [1/4 bit number, 3/4 bit number ];
s4, building a feedback control model based on reinforcement learning, wherein the feedback control model inputs four state data of current, voltage, wire feeding speed and pipeline air supply speed and outputs execution actions, and the execution actions comprise execution control and non-execution control; when the output is execution control, transmitting data of each sensor to the clustering model trained in the step S3, and acquiring and outputting a matched standard pipeline air supply speed interval; when the actual air supply speed of the pipeline exceeds the upper limit, the speed is adjusted through the controller, and when the actual air supply speed of the pipeline is lower than the lower limit, an alarm signal is sent out; when the output is not executing control, the data is not transmitted; specifically, the feedback control model building method includes:
s4.1, creating a control environment which comprises four corresponding state data, namely current, voltage, wire feeding speed and pipeline air supply speed; labeling the output execution action, wherein the execution control is 1, and the non-execution control is 0;
s4.2, formulating a reward function R according to different actions; the method comprises the following specific steps:
wherein action represents output action, v represents air supply speed of the pipeline at the current moment,
representing the average value of the air supply speed of the pipeline at the previous t moment; when the output action is not executing control, action =0, and the reward function R =0; when the output action is execution control, action =1, and the reward function
S4.3, establishing a Q network model, wherein the Q network model comprises 1 layer of gru and not less than 1 layer of linear layer, and the activation function is a Relu function; the input of the Q network model is a control environment, the output is an execution action, and the execution action with the maximum softmax probability is selected as the output;
and S4.4, inputting training set data, training the reinforced model established in the step S4.3, solving the sum of reward functions R generated by selecting the execution action in each training, and finally selecting the execution action strategy corresponding to the maximum reward function R as a Q network model parameter.
2. The intelligent feedback control method for welding pipeline gas supply according to claim 1, characterized in that abnormal values in sensor data are removed by a box chart method in the step S1.
3. An intelligent feedback control system for welding pipeline gas supply is characterized by comprising a multidimensional sensor data acquisition system, a cloud platform and a pipeline gas supply feedback control system; the multi-dimensional sensor data acquisition system acquires multi-dimensional sensing data in real time, a clustering model and a feedback control model provided by the method of any one of claims 1-2 are deployed on a cloud platform, firstly, the multi-dimensional sensing data are transmitted to the feedback control model to judge whether feedback control needs to be executed, and when the feedback control needs to be executed, the multi-dimensional sensing data are transmitted to the clustering model to acquire and control a standard pipeline air supply speed interval; and when the control is not executed, the multi-dimensional sensing data is not transmitted.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211467931.3A CN115570228B (en) | 2022-11-22 | 2022-11-22 | Intelligent feedback control method and system for welding pipeline gas supply |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211467931.3A CN115570228B (en) | 2022-11-22 | 2022-11-22 | Intelligent feedback control method and system for welding pipeline gas supply |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115570228A CN115570228A (en) | 2023-01-06 |
| CN115570228B true CN115570228B (en) | 2023-03-17 |
Family
ID=84588480
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211467931.3A Active CN115570228B (en) | 2022-11-22 | 2022-11-22 | Intelligent feedback control method and system for welding pipeline gas supply |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115570228B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116275643B (en) * | 2023-02-03 | 2023-11-07 | 苏芯物联技术(南京)有限公司 | Intelligent recognition method for execution condition of welding process |
| CN116038076B (en) * | 2023-03-28 | 2023-10-13 | 苏芯物联技术(南京)有限公司 | Welding shielding gas control method based on scene matching |
| CN117226224B (en) * | 2023-11-16 | 2024-01-23 | 苏芯物联技术(南京)有限公司 | Welding protection air flow real-time intelligent control method |
| CN117428302B (en) * | 2023-12-20 | 2024-02-20 | 苏芯物联技术(南京)有限公司 | Intelligent dynamic control method and system for air supply flow rate of welded pipeline |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN207043515U (en) * | 2017-06-05 | 2018-02-27 | 上海和宗焊接设备制造有限公司 | Inexpensive IGBT single tubes single-end ortho-exciting inverter argon arc welding machine and its circuit |
| CN109530862A (en) * | 2018-11-23 | 2019-03-29 | 中车工业研究院有限公司 | A kind of Intelligent welding control method and device |
| CN112183555A (en) * | 2020-09-23 | 2021-01-05 | 唐山松下产业机器有限公司 | Method and system for detecting welding quality, electronic device and storage medium |
| EP3799988A1 (en) * | 2019-09-12 | 2021-04-07 | Illinois Tool Works Inc. | System for and method of labeling weld monitoring time periods using machine learning techniques |
| CN113084311A (en) * | 2021-04-09 | 2021-07-09 | 吴忠市黄河电焊机有限公司 | Group control transmission monitoring system for multifunctional digital welding machine |
| CN114309901A (en) * | 2020-09-28 | 2022-04-12 | 林肯环球股份有限公司 | Method and system for controlling welding fume extraction |
Family Cites Families (24)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8592722B2 (en) * | 2008-11-03 | 2013-11-26 | Illinois Tool Works Inc. | Weld parameter interface |
| US20120095941A1 (en) * | 2010-02-11 | 2012-04-19 | Illinois Tool Works Inc. | Method and Apparatus For Welding Cost Calculator |
| CN108027911A (en) * | 2015-07-29 | 2018-05-11 | 伊利诺斯工具制品有限公司 | Promote the system and method that welding is service software |
| JP7048382B2 (en) * | 2018-03-28 | 2022-04-05 | 株式会社神戸製鋼所 | Control method and control device for gas shielded arc welding |
| CN108877947B (en) * | 2018-06-01 | 2021-10-15 | 重庆大学 | Depth sample learning method based on iterative mean clustering |
| CN111142474A (en) * | 2018-11-05 | 2020-05-12 | 苏芯物联技术(南京)有限公司 | Health monitoring system of data control machine tool |
| US11373262B2 (en) * | 2019-03-25 | 2022-06-28 | Illinois Tool Works Inc. | Systems and methods for providing part quality confidence |
| CN110543816B (en) * | 2019-07-23 | 2021-08-03 | 浙江工业大学 | Self-adaptive face image clustering method based on spectral clustering and reinforcement learning |
| US20210060680A1 (en) * | 2019-08-28 | 2021-03-04 | Lincoln Global, Inc. | Systems and methods providing coordinated dual power outputs supporting a same welding or auxiliary power process |
| CN110653457A (en) * | 2019-09-30 | 2020-01-07 | 桥兴能源科技(广东)有限公司 | Energy-saving comprehensive control system for welding gas |
| CN111098000B (en) * | 2020-01-13 | 2023-10-31 | 广州欣大津汽车智能科技有限公司 | Method for controlling flow rate of shielding gas in welding equipment and gas flow rate control device |
| CN110991568B (en) * | 2020-03-02 | 2020-07-31 | 佳都新太科技股份有限公司 | Target identification method, device, equipment and storage medium |
| CN111790967A (en) * | 2020-03-19 | 2020-10-20 | 何苗 | Gas control system for welding, control method thereof and welding system |
| US20210299777A1 (en) * | 2020-03-30 | 2021-09-30 | Illinois Tool Works Inc. | Systems and methods for control of welding parameter command value limits |
| CN112186799B (en) * | 2020-09-22 | 2022-07-26 | 中国电力科学研究院有限公司 | Distributed energy system autonomous control method and system based on deep reinforcement learning |
| US20220014963A1 (en) * | 2021-03-22 | 2022-01-13 | Shu-Ping Yeh | Reinforcement learning for multi-access traffic management |
| CN113610232B (en) * | 2021-09-28 | 2022-02-22 | 苏州浪潮智能科技有限公司 | Network model quantization method and device, computer equipment and storage medium |
| CN113870260B (en) * | 2021-12-02 | 2022-04-12 | 苏芯物联技术(南京)有限公司 | Welding defect real-time detection method and system based on high-frequency time sequence data |
| CN114528949A (en) * | 2022-03-24 | 2022-05-24 | 哈尔滨理工大学 | Parameter optimization-based electric energy metering abnormal data identification and compensation method |
| CN114951981A (en) * | 2022-05-20 | 2022-08-30 | 重庆理工大学 | Automatic control method for laser welding protective gas |
| CN115147448A (en) * | 2022-05-20 | 2022-10-04 | 南京理工大学 | Image enhancement and feature extraction method for automatic welding |
| CN114755962A (en) * | 2022-06-16 | 2022-07-15 | 苏芯物联技术(南京)有限公司 | Intelligent welding data acquisition system and method based on multiple sensors |
| CN114969553B (en) * | 2022-07-19 | 2022-10-28 | 苏芯物联技术(南京)有限公司 | Welding cost and process parameter comprehensive intelligent recommendation method based on knowledge graph |
| CN115293383A (en) * | 2022-10-09 | 2022-11-04 | 西南石油大学 | Game theory fused transformer risk cause analysis method |
-
2022
- 2022-11-22 CN CN202211467931.3A patent/CN115570228B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN207043515U (en) * | 2017-06-05 | 2018-02-27 | 上海和宗焊接设备制造有限公司 | Inexpensive IGBT single tubes single-end ortho-exciting inverter argon arc welding machine and its circuit |
| CN109530862A (en) * | 2018-11-23 | 2019-03-29 | 中车工业研究院有限公司 | A kind of Intelligent welding control method and device |
| EP3799988A1 (en) * | 2019-09-12 | 2021-04-07 | Illinois Tool Works Inc. | System for and method of labeling weld monitoring time periods using machine learning techniques |
| CN112183555A (en) * | 2020-09-23 | 2021-01-05 | 唐山松下产业机器有限公司 | Method and system for detecting welding quality, electronic device and storage medium |
| CN114309901A (en) * | 2020-09-28 | 2022-04-12 | 林肯环球股份有限公司 | Method and system for controlling welding fume extraction |
| CN113084311A (en) * | 2021-04-09 | 2021-07-09 | 吴忠市黄河电焊机有限公司 | Group control transmission monitoring system for multifunctional digital welding machine |
Non-Patent Citations (2)
| Title |
|---|
| 徐德进 ; .基于云智能焊接管控的大数据分析系统设计.2018,(第10期),全文. * |
| 杨雨禾 ; 雷正龙 ; 黎炳蔚 ; 陈彦宾 ; 吴军 ; 张黎旭 ; 秦占领 ; .D406A超高强度钢激光-TIG复合填丝焊接气孔特性分析.2017,(第12期),全文. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115570228A (en) | 2023-01-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115570228B (en) | Intelligent feedback control method and system for welding pipeline gas supply | |
| Pham et al. | Artificial intelligence in engineering | |
| CN110023850B (en) | Method and control device for controlling a technical system | |
| CN107220734A (en) | CNC Lathe Turning process Energy Consumption Prediction System based on decision tree | |
| CN114048600A (en) | Digital twin-driven multi-model fusion industrial system anomaly detection method | |
| CN112446091A (en) | Artificial neural network-based pulsating pressure prediction method | |
| CN111898867B (en) | Airplane final assembly production line productivity prediction method based on deep neural network | |
| CN112180832A (en) | Artificial intelligent wine picking device and wine picking method thereof | |
| CN108427400A (en) | A kind of aircraft airspeed pipe method for diagnosing faults based on neural network Analysis design | |
| CN111476278A (en) | Intelligent gas concentration detection system | |
| CN115481726A (en) | Industrial robot complete machine health assessment method and system | |
| CN115146798A (en) | Assembly robot full-process monitoring and assisting method and system based on body data | |
| CN115129003A (en) | Automatic production line intelligent monitoring system based on self-learning time-varying digital twin | |
| Guan et al. | Application of a novel PNN evaluation algorithm to a greenhouse monitoring system | |
| CN113418229A (en) | Heating management and control valve intelligent regulation and control system based on convolution-self-coding neural network | |
| CN116503029B (en) | Module data cooperative processing method and system for automatic driving | |
| CN113126488A (en) | Intelligent substance mixing system | |
| CN116859839A (en) | Industrial control method and device based on model training | |
| Xie | Industrial Robot Assembly Line Design Using Machine Vision | |
| CN107480647B (en) | Method for detecting abnormal behaviors in real time based on inductive consistency abnormality detection | |
| CN116203836A (en) | Method for intelligently controlling addition of ammonia water as industrial flue gas absorbent | |
| CN114880930A (en) | Entity equipment digital twin modeling simulation prediction method based on online learning | |
| Fan et al. | Survey of the selection and evaluation for dispatching rules in dynamic job shop scheduling problem | |
| Trinh et al. | Dynamics modeling of industrial robots using transformer networks | |
| CN112816211A (en) | Belt conveyor fault diagnosis method based on PSO-BP algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |