CN114118355A - Stainless steel pickling process parameter control method based on neural network - Google Patents
Stainless steel pickling process parameter control method based on neural network Download PDFInfo
- Publication number
- CN114118355A CN114118355A CN202111172831.3A CN202111172831A CN114118355A CN 114118355 A CN114118355 A CN 114118355A CN 202111172831 A CN202111172831 A CN 202111172831A CN 114118355 A CN114118355 A CN 114118355A
- Authority
- CN
- China
- Prior art keywords
- data
- pickling
- neural network
- steel
- convolutional neural
- 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.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- C—CHEMISTRY; METALLURGY
- C23—COATING METALLIC MATERIAL; COATING MATERIAL WITH METALLIC MATERIAL; CHEMICAL SURFACE TREATMENT; DIFFUSION TREATMENT OF METALLIC MATERIAL; COATING BY VACUUM EVAPORATION, BY SPUTTERING, BY ION IMPLANTATION OR BY CHEMICAL VAPOUR DEPOSITION, IN GENERAL; INHIBITING CORROSION OF METALLIC MATERIAL OR INCRUSTATION IN GENERAL
- C23G—CLEANING OR DE-GREASING OF METALLIC MATERIAL BY CHEMICAL METHODS OTHER THAN ELECTROLYSIS
- C23G1/00—Cleaning or pickling metallic material with solutions or molten salts
- C23G1/02—Cleaning or pickling metallic material with solutions or molten salts with acid solutions
- C23G1/08—Iron or steel
- C23G1/081—Iron or steel solutions containing H2SO4
-
- C—CHEMISTRY; METALLURGY
- C23—COATING METALLIC MATERIAL; COATING MATERIAL WITH METALLIC MATERIAL; CHEMICAL SURFACE TREATMENT; DIFFUSION TREATMENT OF METALLIC MATERIAL; COATING BY VACUUM EVAPORATION, BY SPUTTERING, BY ION IMPLANTATION OR BY CHEMICAL VAPOUR DEPOSITION, IN GENERAL; INHIBITING CORROSION OF METALLIC MATERIAL OR INCRUSTATION IN GENERAL
- C23G—CLEANING OR DE-GREASING OF METALLIC MATERIAL BY CHEMICAL METHODS OTHER THAN ELECTROLYSIS
- C23G1/00—Cleaning or pickling metallic material with solutions or molten salts
- C23G1/02—Cleaning or pickling metallic material with solutions or molten salts with acid solutions
- C23G1/08—Iron or steel
- C23G1/086—Iron or steel solutions containing HF
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Chemical & Material Sciences (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Software Systems (AREA)
- Mechanical Engineering (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Organic Chemistry (AREA)
- Metallurgy (AREA)
- Chemical Kinetics & Catalysis (AREA)
- General Chemical & Material Sciences (AREA)
- Materials Engineering (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Quality & Reliability (AREA)
- Cleaning And De-Greasing Of Metallic Materials By Chemical Methods (AREA)
Abstract
The invention discloses a stainless steel pickling process parameter control method based on a neural network, which comprises the following steps: establishing an acid concentration prediction model, and training and testing; after the head of the strip steel enters the pickling area, the multilayer perceptron receives actually measured pickling data sent by the production line PLC, historical data of a steel coil in a preorder procedure and steel coil information data sent by a material tracking program, and processes the data to form a characteristic vector; when the head of the strip steel enters the meter, the convolutional neural network receives image information and processes the image information to form a characteristic vector, and when the head of the strip steel enters the pickling area but does not reach the meter, the system uses the historical accumulated characteristic vector of the strip steel with the corresponding specification; and the full connection layer performs characteristic combination and regression on the characteristic vector formed by the multilayer perceptron and the convolutional neural network, predicts the pickling process parameters and issues the prediction result to the acid control PLC. The method can improve the accuracy of acid concentration prediction and improve the pickling quality of the surface of the strip steel.
Description
Technical Field
The invention relates to the technical field of hot-rolled stainless steel pickling, in particular to a stainless steel pickling process parameter control method based on a neural network.
Background
At present, the yield of stainless steel crude steel in China breaks through 3000 million tons, and the capacity accounts for more than 50% of the global capacity. The annealing and pickling treatment of the hot-rolled stainless steel coil is a necessary process for the subsequent processing of the hot-rolled stainless steel strip. The hot-rolled coil annealing pickling line (HAPL) recovers the physical property and the mechanical property of a steel plate by carrying out heat treatment on a hot-rolled stainless steel coil, achieves the aims of eliminating surface iron scales and controlling surface roughness by carrying out surface treatment of dephosphorization, shot blasting and pickling, and provides qualified raw materials for subsequent processes.
As a treatment line for continuous production, in order to ensure the excellent performance and quality of treated products, the treatment line needs to be realized by controlling and optimizing various parameters of a process section, wherein the process section comprises an annealing furnace, a scale breaker, a shot blasting machine and a pickling area, and a TV value (T represents the thickness of strip steel, V represents the running speed of the strip steel, and the TV is the product of the thickness of the strip steel and the running speed of the strip steel) is a key index for measuring the annealing and pickling process of stainless steel. The speed V of the process section in the TV value is determined by the annealing furnace due to the self characteristics of the annealing furnace, and how to accurately control the process parameters of the pickling section is realized according to the actual process setting of the preorder process section on the premise that the speed V of the process section is determined for the pickling section of the last area in the process section, thereby having great significance for ensuring the product quality.
The control target of the pickling section is to remove iron scales and pollutants on the surface of the strip steel, passivate the surface of the stainless steel, endow the stainless steel with corrosion resistance, and simultaneously avoid the over-pickling or under-pickling phenomena. The pickling of hot-rolled stainless steel generally adopts a mode of "sulfuric acid stage + mixed acid stage (hydrofluoric acid + nitric acid)" (hereinafter, the chemical formula is used). The key process control parameter of the pickling section is H2SO4Concentration, H2SO4Temperature, HF concentration, HF temperature, HNO3Concentration, HNO3And (3) temperature. At present, an on-line acid analyzer can detect the concentration of various acid liquids on line, but the on-line acid concentration analyzer is not equipped in most production lines at present due to high price and difficult maintenance.
Disclosure of Invention
The invention aims to provide a stainless steel pickling process parameter control method based on a neural network aiming at the defects in the existing pickling section key process parameter control technology.
The technical scheme adopted for realizing the purpose of the invention is as follows:
a stainless steel pickling process parameter control method based on a neural network is characterized in that an acid concentration prediction model is established, the acid concentration prediction model comprises a multilayer perceptron, two branches of a convolutional neural network and a full connection layer for connecting the two branches, and the acid concentration prediction model comprises the following steps:
the multilayer perceptron receives actually measured pickling data, historical data of a steel strip in a preorder procedure and steel coil information data, and processes the actually measured pickling data, the historical data and the steel coil information data to form a first characteristic vector;
the convolutional neural network receives the pickled strip steel image information, processes the pickled strip steel image information and forms a second feature vector;
and the full-connection layer performs characteristic combination and regression on the first characteristic vector and the second characteristic vector, and predicts pickling process parameters.
In the technical scheme, the actually measured pickling data is sent by a production line PLC;
after the head of the strip steel enters the pickling area, the multilayer sensor receives data;
when the head of the strip steel enters the meter inspection instrument, the convolutional neural network receives image information, and when the head of the strip steel enters the pickling area but does not reach the meter inspection instrument, the system uses the historical accumulated characteristic vector of the strip steel with the corresponding specification as a second characteristic vector;
and after the system predicts the pickling process parameters, the pickling process parameters are sent to an acid control PLC.
In the above technical solution, the acid concentration prediction model is created by the following method:
defining input layer nodes of the multilayer perceptron, wherein the input layer nodes comprise the actually measured pickling data, historical data of a steel strip in a preorder process and steel coil information data; defining 2 fully-connected hidden layers, wherein the first hidden layer is provided with 10 nodes, the second hidden layer is provided with 6 nodes, relu is used as an activation function, and the number of nodes of an output layer is 6;
defining a convolutional neural network, wherein an input layer of the convolutional neural network is a three-channel color image, firstly defining a first-stage fully-connected convolutional layer, carrying out convolutional calculation on pixel point matrix data of an image picture, and using relu as an activation function; then defining a pooling layer and a full-connection layer to jointly form extraction of the level features in the surface quality image, and defining 6 nodes of the output layer of the convolutional neural network;
the multilayer perceptron and the convolutional neural network operate independently before being connected, after the definition of the multilayer perceptron and the convolutional neural network is completed, a full connection layer is established to perform feature combination on the multilayer perceptron and the convolutional neural network, regression is performed on the full connection layer, and predicted pickling process parameters are output.
In the technical scheme, the steel coil information data is sent by a material tracking program and comprises the type of strip steel, the width of the strip steel, the thickness of the strip steel, the coiling temperature and/or the process section speed of the steel coil in hot continuous rolling, and the historical data of the preorder process comprises the outlet plate temperature of an annealing furnace, the tension of a scale breaker, the reduction 1 of the scale breaker, the reduction 2 of the scale breaker, the reduction 3 of the scale breaker, the tension of a shot blasting machine and/or the rotating speed of the shot blasting machine; the actually measured pickling data comprises the liquid level of the sulfuric acid tank, the liquid level of the mixed acid tank, the temperature of the sulfuric acid heat exchanger, the temperature of the mixed acid tank heat exchanger and/or acid liquid sampling and testing data.
In the above technical solution, the size of the three-channel color image is 299 × 3, and the unit is a pixel, which respectively represents the width, height, and depth of the input image.
In the technical scheme, the predicted pickling process parameters comprise the acid liquor concentration, the temperature and the metal ion concentration of the sulfuric acid tank and the acid liquor concentration, the temperature and the metal ion concentration of the mixed acid tank.
In the above technical solution, the method for training and testing the acid concentration prediction model includes the following steps:
by collecting 50 rolls of production process parameters of an annealing and pickling project, surface quality image data after pickling and acid liquid sampling and testing data as a data set, wherein the acid liquid sampling and testing data are target data, according to the following steps of 4: 1, dividing a data set into a training set and a testing set, and training and testing the acid concentration prediction model.
In the technical scheme, when the multilayer perceptron processes data, the parameter information is processed into a standardized mean value and a standardized standard deviation, and in the convolutional neural network, the image information is graded according to quality testing personnel and is marked, and the grade is totally classified into 6 grades.
In the technical scheme, the multilayer perceptron scales the category data and the numerical data to a range of [0,1], and then the category data and the numerical data form a first feature vector;
the convolutional neural network scales the image information to the range [0,1], which is then composed into a second feature vector.
In the technical scheme, in the acid concentration prediction model training process, the model is evaluated by using mean absolute percentage error loss (MAPE), and an Adam optimizer is used for optimizing the model on the attenuation of the learning rate.
Compared with the prior art, the invention has the beneficial effects that:
1. aiming at the problem that the acid concentration is difficult to detect on line in the production process of the stainless steel hot rolled strip steel annealing and pickling treatment line, the invention establishes the acid concentration prediction model, takes the mixed data as input, processes the image data in the branch of the acid concentration prediction model, processes the category data and the parameter data in the multilayer perceptron, and combines and regresses the feature vectors respectively by the full connecting layer.
2. The method can more accurately predict the acid liquor concentration and the metal ion concentration of the sulfuric acid tank, the acid liquor concentration of the mixed acid tank and the metal ion concentration in the pickling process in real time, further improve the online control precision of the pickling process section, solve the quality problems of insufficient pickling, over pickling and the like easily occurring in production, and have remarkable application value to the production process.
Drawings
FIG. 1 is a diagram showing a structure of an acid concentration prediction model.
FIG. 2 is a flow chart of the pickling process control.
Detailed Description
The present invention will be described in further detail with reference to specific examples. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Example 1
Establishing an acid concentration prediction model, which comprises the following specific steps:
defining input layer nodes of the multilayer sensor, wherein the input layer nodes comprise steel coil information data (comprising strip steel type, strip steel width, strip steel thickness, coiling temperature and process section speed of a steel coil in hot continuous rolling) sent by a material tracking program, historical data of a preorder process (comprising outlet plate temperature of an annealing furnace, tension of a scale breaker, reduction 1 of the scale breaker, reduction 2 of the scale breaker, reduction 3 of the scale breaker, tension of a shot blasting machine and rotating speed of the shot blasting machine, wherein the three reduction of the scale breaker respectively correspond to the reduction of three rollers which can be independently adjusted below the scale breaker) and acid washing measured data (sulfuric acid tank liquid level, mixed acid tank liquid level, sulfuric acid heat exchanger temperature, mixed acid tank heat exchanger temperature and/or acid liquor sampling test data); defining 2 fully-connected hidden layers, wherein the first hidden layer is provided with 10 nodes, the second hidden layer is provided with 6 nodes, relu is used as an activation function, and the number of nodes of an output layer is 6;
defining a convolutional neural network, wherein an input layer of the convolutional neural network is a three-channel color image, the size of the convolutional neural network is 299 x 3, the unit of the convolutional neural network is a pixel, the width, the height and the depth of the input image are respectively represented, firstly, defining a first-stage fully-connected convolutional layer, carrying out convolutional calculation on image picture pixel point matrix data, and using relu as an activation function; then defining a pooling layer and a full-connection layer to jointly form extraction of the level features in the surface quality image, and defining 6 nodes of the output layer of the convolutional neural network;
the multilayer perceptron and the convolutional neural network are independently operated before being connected, after the definition of the multilayer perceptron and the convolutional neural network is completed, a full connection layer is established to carry out feature combination on the multilayer perceptron and the convolutional neural network, regression is carried out on the full connection layer, and predicted acid washing technological parameters are output, wherein the acid washing technological parameters comprise acid liquor concentration, temperature and metal ion concentration of a sulfuric acid tank and acid liquor concentration, temperature and metal ion concentration of a mixed acid tank.
Example 2
Training and testing an acid concentration prediction model:
in order to train the model, 50 production process parameters of an annealing and pickling project, pickled surface quality image data and acid liquid sampling and testing data are collected, wherein the acid liquid sampling and testing data are target data. According to the following steps of 4: a 1 ratio divides the data set into a training set and a test set.
In the multilayer perceptron, the actually measured parameter information of the strip steel production is processed into a standardized mean value and a standardized standard deviation. In the convolutional neural network, for a surface quality image, judging a grade according to quality inspection personnel and marking, wherein the grade is totally divided into 6 grades, and specifically, the grade can be as follows: the grades can be classified according to different specifications of a steel mill, such as 1 grade, 1A grade, 2A grade, 3 grade and 3A grade.
When the data is preprocessed, the category data (the category of steel grade) and the numerical data are scaled to the range of [0,1] so as to facilitate better training and convergence. And then the feature vectors are formed into a feature vector which is used as the first input of the mixed input neural network. The image data is also scaled to the range [0,1] as a second input to the hybrid input neural network.
During model training, the model was evaluated using mean absolute percent error loss (MAPE), while the model was optimized using the attenuation of the learning rate by the Adam optimizer.
And after the model training reaches the standard, taking the model as an independent model process of the HAPL process control system to participate in the online control of HAPL.
Example 3
And (3) applying the acid concentration prediction model to carry out process control:
after the head of the strip steel enters the pickling area, an acid concentration prediction model program processes the actually measured pickling data sent by a production line PLC, the historical data of a steel coil in a preorder process and the steel coil information data sent by a material tracking program to form a characteristic vector after receiving the actually measured pickling data, the historical data of the steel coil in the preorder process and the steel coil information data; after the head of the strip steel enters the meter inspection instrument, the convolutional neural network receives image information, processes the image information and forms a characteristic vector; and the full connection layer performs characteristic combination on the characteristic vector formed by the multilayer perceptron and the convolutional neural network, regresses the characteristic vector, predicts the pickling process parameters and issues the prediction result to the acid control PLC. The method can improve the accuracy of acid concentration prediction, and further improve the surface pickling quality of the strip steel.
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 decorations can be made without departing from the principle of the present invention, and these modifications and decorations should also be regarded as the protection scope of the present invention.
Claims (10)
1. A stainless steel pickling process parameter control method based on a neural network is characterized by comprising the following steps:
establishing an acid concentration prediction model, wherein the acid concentration prediction model comprises a multilayer perceptron, two branches of a convolutional neural network and a full connection layer for connecting the two branches, and the acid concentration prediction model comprises the following steps:
the multilayer perceptron receives actually measured pickling data, historical data of a steel strip in a preorder procedure and steel coil information data, and processes the actually measured pickling data, the historical data and the steel coil information data to form a first characteristic vector;
the convolutional neural network receives the pickled strip steel image information, processes the pickled strip steel image information and forms a second feature vector;
and the full-connection layer performs characteristic combination and regression on the first characteristic vector and the second characteristic vector, and predicts pickling process parameters.
2. The method for controlling parameters in a stainless steel pickling process according to claim 1, wherein the measured pickling data is sent by a production line PLC;
after the head of the strip steel enters the pickling area, the multilayer sensor receives data;
when the head of the strip steel enters the meter inspection instrument, the convolutional neural network receives image information, and when the head of the strip steel enters the pickling area but does not reach the meter inspection instrument, the system uses the historical accumulated characteristic vector of the strip steel with the corresponding specification as a second characteristic vector;
and after the system predicts the pickling process parameters, the pickling process parameters are sent to an acid control PLC.
3. The method for controlling parameters in a stainless steel pickling process according to claim 1, wherein the acid concentration prediction model is created by:
defining an input layer node of the multilayer sensor, wherein the input layer node comprises the actually measured pickling data, historical data of a steel strip in a preorder process and steel coil information data; defining 2 fully-connected hidden layers, wherein the first hidden layer is provided with 10 nodes, the second hidden layer is provided with 6 nodes, relu is used as an activation function, and the number of nodes of an output layer is 6;
defining a convolutional neural network, wherein an input layer of the convolutional neural network is a three-channel color image, firstly defining a first-stage fully-connected convolutional layer, carrying out convolutional calculation on pixel point matrix data of an image picture, and using relu as an activation function; then defining a pooling layer and a full-connection layer to jointly form extraction of the level features in the surface quality image, and defining 6 nodes of the output layer of the convolutional neural network;
the multilayer perceptron and the convolutional neural network operate independently before being connected, after the definition of the multilayer perceptron and the convolutional neural network is completed, a full connection layer is established to perform feature combination on the multilayer perceptron and the convolutional neural network, regression is performed on the full connection layer, and predicted pickling process parameters are output.
4. The stainless steel pickling process parameter control method according to claim 1, wherein the steel coil information data is sent by a material tracking program and comprises a steel strip type, a steel strip width, a steel strip thickness, a coiling temperature and/or a process section speed of a steel coil in hot continuous rolling, and the historical data of the previous procedure comprises an annealing furnace outlet plate temperature, a scale breaker tension, a scale breaker rolling reduction 1, a scale breaker rolling reduction 2, a scale breaker rolling reduction 3, a shot blasting machine tension and/or a shot blasting machine rotating speed; the actually measured pickling data comprises the liquid level of the sulfuric acid tank, the liquid level of the mixed acid tank, the temperature of the sulfuric acid heat exchanger, the temperature of the mixed acid tank heat exchanger and/or acid liquid sampling and testing data.
5. The method for controlling parameters in a stainless steel pickling process according to claim 3, wherein the three-channel color image has dimensions of 299 x 3 in pixels, and represents the width, height and depth of an input image, respectively.
6. The method of claim 1, wherein the predicted pickling process parameters comprise sulfuric acid tank acid concentration, temperature, metal ion concentration, and mixed acid tank acid concentration, temperature, metal ion concentration.
7. The method for controlling parameters in a stainless steel pickling process according to claim 1, wherein the method for training and testing the acid concentration prediction model comprises the following steps:
by collecting 50 rolls of production process parameters of an annealing and pickling project, surface quality image data after pickling and acid liquid sampling and testing data as a data set, wherein the acid liquid sampling and testing data are target data, according to the following steps of 4: 1, dividing a data set into a training set and a testing set, and training and testing the acid concentration prediction model.
8. The method for controlling parameters in stainless steel pickling process according to claim 1, wherein the multi-layer sensor processes the parameter information into a standardized mean value and standard deviation when processing data, and in the convolutional neural network, the image information is graded and labeled according to quality inspector judgment, and is classified into 6 grades in total.
9. The method for controlling parameters in stainless steel pickling process according to claim 1, wherein said multilayer sensor scales category data and numerical data to a [0,1] range and then forms them into a first feature vector;
the convolutional neural network scales the image information to the range [0,1], which is then composed into a second feature vector.
10. The method of claim 7, wherein during the acid concentration prediction model training process, the model is evaluated using mean absolute percentage error loss while an Adam optimizer optimizes the model for the decay of the learning rate.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111172831.3A CN114118355A (en) | 2021-10-08 | 2021-10-08 | Stainless steel pickling process parameter control method based on neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111172831.3A CN114118355A (en) | 2021-10-08 | 2021-10-08 | Stainless steel pickling process parameter control method based on neural network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114118355A true CN114118355A (en) | 2022-03-01 |
Family
ID=80441351
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111172831.3A Pending CN114118355A (en) | 2021-10-08 | 2021-10-08 | Stainless steel pickling process parameter control method based on neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114118355A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115985822A (en) * | 2023-03-21 | 2023-04-18 | 江苏凯威特斯半导体科技有限公司 | High-precision surface quality control system for integrated circuit chip |
| CN116288381A (en) * | 2023-03-16 | 2023-06-23 | 山东钢铁集团日照有限公司 | Closed-loop control method for realizing stable pickling quality by automatically adjusting hydrochloric acid process parameters |
| WO2023226227A1 (en) * | 2022-05-27 | 2023-11-30 | 福建龙氟新材料有限公司 | Automatic batching system for preparing electronic-grade hydrofluoric acid and batching method therefor |
| CN117238420A (en) * | 2023-11-14 | 2023-12-15 | 太原理工大学 | Method and device for predicting mechanical properties of ultrathin strip |
-
2021
- 2021-10-08 CN CN202111172831.3A patent/CN114118355A/en active Pending
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2023226227A1 (en) * | 2022-05-27 | 2023-11-30 | 福建龙氟新材料有限公司 | Automatic batching system for preparing electronic-grade hydrofluoric acid and batching method therefor |
| CN116288381A (en) * | 2023-03-16 | 2023-06-23 | 山东钢铁集团日照有限公司 | Closed-loop control method for realizing stable pickling quality by automatically adjusting hydrochloric acid process parameters |
| CN115985822A (en) * | 2023-03-21 | 2023-04-18 | 江苏凯威特斯半导体科技有限公司 | High-precision surface quality control system for integrated circuit chip |
| CN115985822B (en) * | 2023-03-21 | 2023-06-09 | 江苏凯威特斯半导体科技有限公司 | High-precision surface quality control system for integrated circuit chip |
| CN117238420A (en) * | 2023-11-14 | 2023-12-15 | 太原理工大学 | Method and device for predicting mechanical properties of ultrathin strip |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114118355A (en) | Stainless steel pickling process parameter control method based on neural network | |
| KR100527788B1 (en) | Continuous Pickling Method and Continuous Pickling Device | |
| WO2023130666A1 (en) | Strip steel plate convexity prediction method based on data-driving and mechanism model fusion | |
| CN100577315C (en) | Device for forecasting and controlling material quality of roll line | |
| CN107179749A (en) | Hot dip zinc product whole process method of quality control | |
| CN106825069B (en) | A kind of cold-strip steel high precision plates shape surface roughness on-line intelligence control method | |
| CN114329940A (en) | Continuous casting billet quality prediction method based on extreme learning machine | |
| CN110434172B (en) | Load distribution calculation method for continuous rolling of furnace coil and finishing mill group | |
| CN109935280A (en) | A kind of blast-melted quality prediction system and method based on integrated study | |
| CN101934295A (en) | Pre-calculation method for controlled cooling of thick plate after rolling | |
| CN115121626B (en) | Hot-rolled strip steel transient hot roll shape forecasting method based on error compensation | |
| CN110222825B (en) | Cement product specific surface area prediction method and system | |
| CN114888092A (en) | Cold rolling deformation resistance prediction method based on cross-process data platform | |
| Mazur et al. | Quality control system for a hot-rolled metal surface. | |
| US20240002964A1 (en) | Method and system for determining converter tapping quantity | |
| CN111553226A (en) | Method for extracting river monitoring section water surface width based on remote sensing interpretation technology | |
| Jin et al. | Identification of impacting factors of surface defects in hot rolling processes using multi-level regression analysis | |
| CN214556274U (en) | Online detection device for yield elongation of cold-rolled strip steel | |
| CN117171936A (en) | Slab quality prediction method for extracting real-time characteristic value of crystallizer based on defect mechanism | |
| CN114139292A (en) | Process control method for thickness of paint film of steel plate pretreatment based on big data analysis platform | |
| CN109032097A (en) | A kind of cold-strip steel galvanized wire course control method for use | |
| CN110174409B (en) | Medium plate periodic defect control method based on real-time detection result | |
| Vozmilov et al. | Using computer vision to recognize defects on the surface of hot-rolled steel | |
| CN105160127A (en) | Cast steel plate (CSP) flow hot continuous rolling production line based performance forecast method for Q235B steel | |
| CN115703131A (en) | Use method of cold-rolled sheet yield elongation on-line detection device |
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 |