CN102921915A - Slag carry-over detection method and device based on image recognition of vortex on surface of molten steel - Google Patents
Slag carry-over detection method and device based on image recognition of vortex on surface of molten steel Download PDFInfo
- Publication number
- CN102921915A CN102921915A CN2012104090493A CN201210409049A CN102921915A CN 102921915 A CN102921915 A CN 102921915A CN 2012104090493 A CN2012104090493 A CN 2012104090493A CN 201210409049 A CN201210409049 A CN 201210409049A CN 102921915 A CN102921915 A CN 102921915A
- Authority
- CN
- China
- Prior art keywords
- vortex
- slag
- image
- molten steel
- camera
- 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
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B22—CASTING; POWDER METALLURGY
- B22D—CASTING OF METALS; CASTING OF OTHER SUBSTANCES BY THE SAME PROCESSES OR DEVICES
- B22D11/00—Continuous casting of metals, i.e. casting in indefinite lengths
- B22D11/16—Controlling or regulating processes or operations
- B22D11/18—Controlling or regulating processes or operations for pouring
- B22D11/181—Controlling or regulating processes or operations for pouring responsive to molten metal level or slag level
- B22D11/185—Controlling or regulating processes or operations for pouring responsive to molten metal level or slag level by using optical means
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B22—CASTING; POWDER METALLURGY
- B22D—CASTING OF METALS; CASTING OF OTHER SUBSTANCES BY THE SAME PROCESSES OR DEVICES
- B22D41/00—Casting melt-holding vessels, e.g. ladles, tundishes, cups or the like
- B22D41/14—Closures
- B22D41/22—Closures sliding-gate type, i.e. having a fixed plate and a movable plate in sliding contact with each other for selective registry of their openings
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B22—CASTING; POWDER METALLURGY
- B22D—CASTING OF METALS; CASTING OF OTHER SUBSTANCES BY THE SAME PROCESSES OR DEVICES
- B22D41/00—Casting melt-holding vessels, e.g. ladles, tundishes, cups or the like
- B22D41/14—Closures
- B22D41/22—Closures sliding-gate type, i.e. having a fixed plate and a movable plate in sliding contact with each other for selective registry of their openings
- B22D41/38—Means for operating the sliding gate
Landscapes
- Engineering & Computer Science (AREA)
- Mechanical Engineering (AREA)
- Continuous Casting (AREA)
Abstract
The invention relates to the field of automatic control in metallurgical industry, and aims to provide a slag carry-over detection method and a slag carry-over detection device based on image recognition of a vortex on the surface of molten steel. The slag carry-over detection method comprises the following steps of: arranging a camera above a ladle, connecting the camera to an image signal detection unit and an industrial personal computer through a cable in sequence, and processing and extracting an image to obtain three tapping slag inclusion states. Slag carry-over prediction accuracy and stability are improved according to the essence of ladle slag carry-over; the central position and state characteristic quantity information of the vortex on the surface of the molten steel in the ladle are obtained by a free surface vortex recognition method, so that characteristic information is extracted under the complex working condition on a continuous casting site; and a function of detection before slag carry-over is realized, the quality of the molten steel is effectively controlled, and yield is improved.
Description
Technical field
The present invention relates to the Metallurgical Industry Automation control field, particularly a kind of lower slag inspection method and device based on the image recognition of molten steel surface vortex of producing towards continuous casting.
Background technology
Along with the progress of metallurgical technology, improving constantly of iron and steel kind and quality, more and more higher to the requirement of Molten Steel Cleanliness in the continuous casting of iron and steel production.In continuous casting production process, the oxidant in the large bag, impurity mix formation liquid slag, and its proportion only is about 1/3rd of molten steel, therefore can float on molten steel top.In the pouring molten steel later stage, be subjected to the impact of vortex, the slag of melting can enter tundish from large packet flow gradually, affects steel quality, reduces the service life of middle bag, when serious even continuous casting production can't be carried out.In order to improve steel quality, reduce the erosion for tapping hole and slide gate nozzle, the service life of wrapping in the raising and molten steel recovery rate must be identified for the several key states in the large bag casting process.Therefore, since the eighties in last century, in succession develop multiple large bag casting condition recognition technology both at home and abroad, mainly comprised electromagnetic detection method, ultrasonic Detection Method, Infrared Detection Method and vibration detection method.
The electromagnetism method of identification carries out state recognition by the sensor that two concentric annular coils form, and when logical high-frequency alternating electric current in coil, molten steel will have induced field to produce, and its direction can be judged by right-hand screw rule.Because the magnetic conductivity of slag is much smaller than the magnetic conductivity (during 1600 ℃ of left and right sides, their ratio is 1:10000) of molten steel, the magnetic field that induction produces in clean steel water is far longer than the magnetic field that produces in the molten steel that is containing slag.The defective of electromagnetism method of identification is: apparatus structure is complicated, and installation process is more loaded down with trivial details, needs 3 ~ 6 working days, and need to carry out to a certain degree transformation to the continuous casting production equipment when installing, and will incur loss through delay production like this; Service life is short, and maintenance cost is high, and is easily impaired owing to solenoid is worked under abominable hot environment, thereby detection system was lost efficacy, and therefore needs periodic maintenance, more emat coil.
Ultrasonic Detection Method also is a kind of industrial detection method commonly used.Its principle is to utilize slag is arranged in the steel stream and to be that difference between ultrasonic wave emission, the reflected signal realizes the detection to slag without slag.Although this method is on not impact of casting process, because the operating ambient temperature of ultrasonic probe is up to 1500 degrees centigrade, working environment is more abominable, and manufacturing and cost of use are high, also have long period of time from commercial Application.
Infrared Detection Method utilizes the different principle of thermal emissivity rate of molten steel and slag that slag is identified, in the identification of the application that present this type systematic is commonplace and converter, electric furnace steel tapping state, if directly apply in the state recognition of large bag casting process, then must remove long nozzle, steel stream is directly exposed in the air, can cause the secondary oxidation of molten steel, production is very disadvantageous for continuous casting for this.
Method for detecting vibration is to utilize the monitoring molten steel entering the process of tundish from large packet flow, and the vibration that protective casing and motion arm are produced realizes wrapping greatly the identification of casting condition.The slide gate nozzle aperture is larger, and molten steel flow is larger, and corresponding vibration is strong with regard to Shaoxing opera.Mentioned, slag proportion approximately is 1/3rd of clean steel water proportion before, just must be discrepant by clean steel water flow and the mobile vibration that causes of slag therefore.As long as detect this species diversity, just can judge effectively that steel flows down the generation of slag.The bottleneck of vibration detection method is that mainly vibration signal is faint, because the impact capacity of steel stream is limited, therefore the vibration signal of long nozzle collection is easily covered by the interference of continuous casting production on-site environment, thereby causes the wrong judgement of system's generation and then affect Systems balanth.
Summary of the invention
Main purpose of the present invention is to overcome deficiency of the prior art, and a kind of lower slag inspection method and device that casting process is not affected, can effectively control steel quality and raising recovery rate is provided.For solving the problems of the technologies described above, solution of the present invention is:
A kind of lower slag inspection method based on the image recognition of molten steel surface vortex is provided, is that camera is installed in the large side of wrapping, and is connected to successively image signal control unit and industrial computer by cable; Described lower slag inspection method comprises following concrete steps:
The vortex that (A) will wrap greatly interior formation is divided into without slag vortex, mixed slag vortex and full slag vortex three state, and these three kinds of vorticities are normal cast, mixed slag cast and the lower slag three state of corresponding large bag casting process respectively;
(B) by the two dimensional image feature of camera collection molten steel surface vortex, image is carried out preliminary treatment; By based on the Rotational Symmetry Region Segmentation of how much active contour models, detect and the extraction of characteristics of image is carried out in the pivot location based on streamline edge, the direction of Canny operator, carry out the identification of Free Surface characteristics of image by vortex identification and vorticity feature extraction;
(C) after the identification of vortex characteristics of image, to do the vortex characteristic vector that represents after the normalized as four nodes of BP network input layer with area, girth and the vorticity characteristic value in the closely-related vortex of lower slag zone, consist of output node with three tapping slag inclusion state parameters setting, realize training by three layers of feedforward network grader of error Back-Propagation training and to it by design, thereby obtain three kinds of tapping slag inclusion states.
As a kind of improvement, the identification of described Free Surface whirlpool characteristics of image is by condition 1 winding angle condition, i.e. winding angle a=± 2 π, and note is clockwise for just; And fluid satisfies ∠ (I the adjacent area planted agent in condition 2 flow fields
i, I
I-1) ∈ [(0, pi/2), ∠ (I
i, I
I-1) ∈ (pi/2,0)] judge, by satisfy condition simultaneously 1 with 2 of conditions be judged to be whirlpool; Described winding angle
, N wherein
2Be the piece number that square template is divided into, I
iIt is i piece zone streamline tangential direction average.
As a kind of improvement, described image preprocessing process comprises the method for relaxation image denoising and based on the image sharpening of Sobel operator.
As a kind of improvement, the topological structure of described three layers of feedforward network grader is divided into input layer, middle hidden layer and output layer, by the error Back-Propagation training step is:
(a) design input layer and output layer: the nodes of input layer depends on the dimension of vortex characteristic vector, will be after the vortex image recognition and the area in the closely-related vortex of lower slag zone, and girth and vorticity characteristic value ξ
1, ξ
2Be expressed as vortex characteristic vector u=[u after doing normalized
1, u
2, u
3, u
4], thereby as four nodes of BP network input layer; The area in vortex zone and the normalized value of girth be the instantaneous area of vortex and perimeter value than discharge outlet area and girth, vorticity characteristic value ξ
1, ξ
2Normalized value be ξ
1, ξ
2Than 8 grades of gray scale rank values 255; Three the tapping slag inclusion state parameters of output node for having established are namely without slag, mixed slag, full slag;
(b) set the nodes of hidden layer: formula rule of thumb
(n wherein
iBe input layer number, n
0Be output layer node number, a is the Changshu between 1 ~ 10), calculate the span of hidden layer node number;
(c) excitation function adopts the S type function: f (x)=1/ (1+e
-kx), after constantly training reached necessary requirement, training process finished.
As a kind of improvement, the k value of described S type function gets 1.5.
The present invention further provides the lower slag checkout gear based on the image recognition of molten steel surface vortex that is used for realizing preceding method, comprise image capturing system and image processing system, described image capturing system is mounted in the lower end and has the camera directly over the large bag of slide gate nozzle, and image processing system comprises image signal control unit and industrial computer; Camera is connected with image signal control unit by cable, and image signal control unit is connected with industrial computer by optical cable.
As a kind of improvement, described image signal control unit comprises data acquisition module, power management module, mouth of a river control module and onsite alarming module.
As a kind of improvement, described mouth of a river control module one end is connected with slide gate nozzle, and the other end is connected with industrial computer.
As a kind of improvement, described camera is equipped with dust cover and sealing outward.
Compared with prior art, the invention has the beneficial effects as follows:
1, caused and time scoriform attitude essence inseparable with vorticity is started with by the Free Surface vortex slag that inner liquid steel level produces from bale slag-blanking, improved time slag accuracy of the forecast and stability;
2, the recognition methods by the Free Surface vortex obtains center, the status flag amount information of large Baogang water surface vortex, has realized under the on-the-spot complex working condition environment of continuous casting the extraction for characteristic information;
3, design is based on the grader of artificial neural network, by excavation vortex characteristic parameter and the lower slag Relations Among to the training implicit expression of grader, correct is categorized into corresponding molten steel discharged slag state with characteristic parameter, judge the adjusting of control ladle slide gate in billet by the state of lower slag at last, priori measuring ability, more effective control steel quality and the raising recovery rate of " do not descend slag, detect first " have been realized.
Description of drawings
Fig. 1 is the lower slag checkout gear system composition diagram based on the image recognition of molten steel surface vortex;
Fig. 2 is large bag casting condition recognition system implementation schematic diagram;
Fig. 3 is image signal control unit and industrial computer correspondence figure among the present invention;
Fig. 4 is the BP neural network model by three layers of feedforward network grader of error Back-Propagation training;
Reference numeral among the figure is: 1 tundish; 2 molten steel; 3 protective casings; 4 slide gate nozzles; 5 wrap greatly; 6 slags; 7 cameras; 8 image signal control unit; 9 mouth of a river control modules; 10 industrial computers.
The specific embodiment
Below in conjunction with accompanying drawing and the specific embodiment the present invention is described in further detail:
The hardware based on the lower slag checkout gear of molten steel 2 surperficial vortex image recognitions among Fig. 1 forms and mainly comprises image capturing system and image processing system.
Image capturing system forms by being used for gathering the large camera 7 that wraps 5 liquid level image informations.Because system applies is on continuous casting line, the device erecting bed has the large characteristics of dust, therefore need to add dust cover and sealing in camera 7 outsides.
Image processing system comprises image signal control unit 8 and industrial computer 10, image signal control unit 8 integrated data acquisition module, power management module, mouth of a river control module 9 and onsite alarming module.The transfer of data that data acquisition module is responsible for camera 7 is gathered is to industrial computer 10, and simultaneously, data acquisition module also needs to gather tundish 1, wraps 5 weight information, mouth of a river opening information greatly.Power management module provides power supply to other modules in the image signal control unit 8.Mouth of a river control module 9 is used for receiving the signal that opens and closes the mouth of a river and the switching of controlling the mouth of a river from industrial computer 10.The onsite alarming module receives the lower slag alarm signal of industrial computer 10 transmissions and the power supply alarming signal of power management module transmission is also pointed out the site operation personnel with the mode of sound and light alarm.
Picture signal is delivered to image signal control unit 8 by fire resistant shielding cable first, arrives industrial computer 10 by optical cable transmission again, and industrial computer 10 is identified large bag 5 casting conditions by picture signal.
The implementation schematic diagram of the lower slag inspection method that is based on molten steel 2 surperficial vortex image recognitions shown in Figure 2, the concrete correspondence of image signal control unit 8 and industrial computer 10 as shown in Figure 3, the communication of system comprises three kinds of power supply input, signal input and control outputs.Its realization flow is, other modules that power management module is responsible in the image signal control unit 8 provide power supply input signal, and when Power supply was unusual, power management module can send the power supply alarming signal to the onsite alarming module simultaneously.Gather large bag 5 liquid level images by camera 7, this for pattern-recognition signal and tundish 1, wrap 5 weight signal greatly and be uploaded to information acquisition module with mouth of a river aperture signal.Information acquisition module transfers signals to industrial computer 10 after receiving signal.The characteristic information that industrial computer 10 utilizations collect is set up towards large bag 5 casting conditions identification neutral net and is trained, and training utilizes input signal to identify for large bag 5 casting conditions after finishing.
Gathered the two dimensional image feature of molten steel 2 surperficial vortexs in this example by camera 7, image is improved and carries out preliminary treatment, feature extraction and feature identification, the identification that solves molten steel 2 surperficial vortexs is also obtained its position and status information.Will be in the area in the closely-related vortex of lower slag zone after the vortex image recognition, girth and vorticity characteristic value are done the vortex characteristic vector that represents after the normalized as four nodes of BP network input layer, three tapping slag inclusion state parameters setting consist of output node, realize training by three layers of feedforward network grader of error Back-Propagation training and to it by design, thereby obtain three kinds of tapping slag inclusion states.
The concrete steps of this example are as follows:
Step 1: the vortex that will wrap greatly 5 interior formation is divided into without slag vortex, mixed slag vortex and full slag vortex three state, and these three kinds of vorticities are normal cast, mixed slag cast and the lower slag three state of corresponding large bag 5 casting process respectively.
Step 2: gathered the two dimensional image feature of molten steel 2 surperficial vortexs by camera 7, by the method for relaxation image denoising, carry out the image preliminary treatment based on the image sharpening of Sobel operator, by the Rotational Symmetry Region Segmentation based on how much active contour models.Cutting apart of image finished in the evolution that utilizes the effective combining image information exchange of energy functional to cross closed curve, continuity, the closed of object edge have been guaranteed simultaneously, and the topological relation that the usage level set representations has overcome in the curve evolvement process changes a difficult problem, change of shape, baroque object region are cut apart quite effective, be able to be solved cutting apart of Rotational Symmetry zone in the Free Surface vortex image.Based on the streamline edge of Canny operator and direction detects and the extraction of characteristics of image is carried out in the pivot location, have preferably noiseproof feature and higher edge precision, therefore and the linear connection in edge is complete, can farthest obtain in the image streamline marginal information and guarantee its reliability.Carry out the identification of Free Surface characteristics of image by vortex identification and vorticity feature extraction.The identification of described Free Surface whirlpool characteristics of image is by condition 1 winding angle condition, i.e. winding angle a=± 2 π, and note is clockwise for just; And fluid satisfies ∠ (I the adjacent area planted agent in condition 2 flow fields
i, I
I-1) ∈ [(0, pi/2), ∠ (I
i, I
I-1) ∈ (pi/2,0)] judge, by satisfy condition simultaneously 1 with 2 of conditions be judged to be whirlpool.Described winding angle
, N wherein
2Be the piece number that square template is divided into, I
iIt is i piece zone grain direction (being the streamline tangential direction) average.
Step 3: after the identification of vortex characteristics of image, will with the area in the closely-related vortex of lower slag zone, girth and vorticity characteristic value are done the vortex characteristic vector that represents after the normalized as four nodes of BP network input layer, three tapping slag inclusion state parameters setting consist of output node, realize training by three layers of feedforward network grader of error Back-Propagation training and to it by design, thereby obtain three kinds of tapping slag inclusion states.
The topological structure of described three layers of feedforward network grader is divided into input layer, middle hidden layer and output layer, by the error Back-Propagation training step is:
(a) design input layer and output layer: the nodes of input layer depends on the dimension of vortex characteristic vector, will be after the vortex image recognition and the area in the closely-related vortex of lower slag zone, and girth and vorticity characteristic value ξ
1, ξ
2Be expressed as vortex characteristic vector u=[u after doing normalized
1, u
2, u
3, u
4], thereby as four nodes of BP network input layer; The area in vortex zone and the normalized value of girth be the instantaneous area of vortex and perimeter value than discharge outlet area and girth, vorticity characteristic value ξ
1, ξ
2Normalized value be ξ
1, ξ
2Than 8 grades of gray scale rank values 255; Three the tapping slag inclusion state parameters of output node for having established are namely without slag, mixed slag, full slag;
(b) nodes of setting hidden layer: because the hidden layer node number can cause the networking poor fault tolerance to be difficult to correct classification based training sample characteristic vector in addition very little, the hidden layer node number then can increase the training time too much.Formula rule of thumb
(n wherein
iBe input layer number, n
0Be output layer node number, a is the Changshu between 1 ~ 10), calculate the span of hidden layer node number;
(c) the overall error function has just been determined by excitation function fully after topology of networks and training sample data are determined, employing S type function: f (x)=1/ (1+e
-kx), when k=1.5, the error of training is little, converges faster.After constantly training reached necessary requirement, training process finished.
Technical basis of the present invention is:
Be to reach the purpose that control enters tundish 1 slag 6 amounts by for lower slag obtaining in advance constantly for the remarkable meaning of the identification of large bag 5 casting conditions, therefore the most simply dividing is that casting condition is divided into two kinds of normal cast and lower slags.Yet along with the reduction of large bag 5 interior molten steel 2 liquid levels, under the effect of gravity, coriolis force and initial fluid disturbance, steel stream surface can form the Free Surface vortex gradually.Form with vortex and to develop into the slag 6 that can will float on molten steel 2 surfaces after the certain scale and be involved in the vortex core and bring to out head piece and enter tundish 1, also can cause tundish 1 nozzle blocking and reduce continuous casting billet quality and molten steel 2 recovery rates thereby not only affect billet quality.The air that the while vortex is carried under one's arms has increased by the 2 secondary oxidation times of molten steel and has caused billet quality to descend.At this moment the composition of steel stream is molten steel 2, slag 6 and air, and this state is between normal casting condition and the lower scoriform attitude.Comprehensive above the analysis will be wrapped greatly 5 casting process and be divided into normal cast, mixed slag cast, lower slag three state.
In large bag 5 casting process, along with the reduction of large bag 5 interior molten steel 2 liquid levels, because the effect of gravity, coriolis force and initial fluid disturbance, steel stream surface can form the Free Surface vortex gradually.Form with vortex and to develop into the slag 6 that can will float on molten steel 2 surfaces after the certain scale and be involved in the vortex core and bring to out head piece and enter tundish 1.Therefore can grasp Free Surface vortex occurrence positions and state of development by accurate, thereby take the braking measure of vortex that vortex is carried out effective Control and elimination, and then reduce the harm that brings.
In the present invention, introduced the method for image processing, identification and neutral net, in large bag 5 casting process, gather large bag 5 liquid level image informations, image is improved and carries out preliminary treatment, feature extraction and feature identification, the identification that solves molten steel 2 surperficial vortexs is also obtained its position and status information, and with its input as neutral net, by the training of master sample, determine the BP neutral net each function, thereby wrapped greatly 3 states of 5 casting process.
Certainly, what more than enumerate only is a specific embodiment of the present invention, obviously, the invention is not restricted to above embodiment, and many distortion can also be arranged.All distortion that those of ordinary skill in the art can directly derive or associate from content disclosed by the invention all should be thought protection scope of the present invention.
Claims (9)
1. the lower slag inspection method based on the image recognition of molten steel surface vortex is characterized in that, is camera is installed in the large side of wrapping, and is connected to successively image signal control unit and industrial computer by cable; Described lower slag inspection method comprises following concrete steps:
The vortex that (A) will wrap greatly interior formation is divided into without slag vortex, mixed slag vortex and full slag vortex three state, and these three kinds of vorticities are normal cast, mixed slag cast and the lower slag three state of corresponding large bag casting process respectively;
(B) by the two dimensional image feature of camera collection molten steel surface vortex, image is carried out preliminary treatment; By based on the Rotational Symmetry Region Segmentation of how much active contour models, detect and the extraction of characteristics of image is carried out in the pivot location based on streamline edge, the direction of Canny operator, carry out the identification of Free Surface characteristics of image by vortex identification and vorticity feature extraction;
(C) after the identification of vortex characteristics of image, to do the vortex characteristic vector that represents after the normalized as four nodes of BP network input layer with area, girth and the vorticity characteristic value in the closely-related vortex of lower slag zone, consist of output node with three tapping slag inclusion state parameters setting, realize training by three layers of feedforward network grader of error Back-Propagation training and to it by design, thereby obtain three kinds of tapping slag inclusion states.
2. lower slag inspection method according to claim 1 is characterized in that, the identification of described Free Surface whirlpool characteristics of image is by condition 1 winding angle condition, i.e. winding angle a=± 2 π, and note is clockwise for just; And fluid satisfies ∠ (I the adjacent area planted agent in condition 2 flow fields
i, I
I-1) ∈ [(0, pi/2), ∠ (I
i, I
I-1) ∈ (pi/2,0)] judge, by satisfy condition simultaneously 1 with 2 of conditions be judged to be whirlpool; Described winding angle
, N wherein
2Be the piece number that square template is divided into, I
iIt is i piece zone streamline tangential direction average.
3. lower slag inspection method according to claim 1 is characterized in that, described image preprocessing process comprises the method for relaxation image denoising and based on the image sharpening of Sobel operator.
4. lower slag inspection method according to claim 1 is characterized in that, the topological structure of described three layers of feedforward network grader is divided into input layer, middle hidden layer and output layer, by the error Back-Propagation training step is:
(a) design input layer and output layer: the nodes of input layer depends on the dimension of vortex characteristic vector, will be after the vortex image recognition and the area in the closely-related vortex of lower slag zone, and girth and vorticity characteristic value ξ
1, ξ
2Be expressed as vortex characteristic vector u=[u after doing normalized
1, u
2, u
3, u
4], thereby as four nodes of BP network input layer; The area in vortex zone and the normalized value of girth be the instantaneous area of vortex and perimeter value than discharge outlet area and girth, vorticity characteristic value ξ
1, ξ
2Normalized value be ξ
1, ξ
2Than 8 grades of gray scale rank values 255; Three the tapping slag inclusion state parameters of output node for having established are namely without slag, mixed slag, full slag;
(b) set the nodes of hidden layer: formula rule of thumb
(n wherein
iBe input layer number, n
0Be output layer node number, a is the Changshu between 1 ~ 10), calculate the span of hidden layer node number;
(c) excitation function adopts the S type function: f (x)=1/ (1+e
-kx), after constantly training reached necessary requirement, training process finished.
5. lower slag inspection method according to claim 4 is characterized in that, the k value of described S type function gets 1.5.
6. lower slag checkout gear based on the image recognition of molten steel surface vortex of be used for realizing the described method of claim 1, comprise image capturing system and image processing system, it is characterized in that, described image capturing system is mounted in the lower end and has the camera directly over the large bag of slide gate nozzle, and image processing system comprises image signal control unit and industrial computer; Camera is connected with image signal control unit by cable, and image signal control unit is connected with industrial computer by optical cable.
7. lower slag checkout gear according to claim 6 is characterized in that, described image signal control unit comprises data acquisition module, power management module, mouth of a river control module and onsite alarming module.
8. lower slag checkout gear according to claim 7 is characterized in that, described mouth of a river control module one end is connected with slide gate nozzle, and the other end is connected with industrial computer.
9. lower slag checkout gear according to claim 6 is characterized in that, described camera is equipped with dust cover and sealing outward.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012104090493A CN102921915A (en) | 2012-10-23 | 2012-10-23 | Slag carry-over detection method and device based on image recognition of vortex on surface of molten steel |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012104090493A CN102921915A (en) | 2012-10-23 | 2012-10-23 | Slag carry-over detection method and device based on image recognition of vortex on surface of molten steel |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102921915A true CN102921915A (en) | 2013-02-13 |
Family
ID=47636920
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2012104090493A Pending CN102921915A (en) | 2012-10-23 | 2012-10-23 | Slag carry-over detection method and device based on image recognition of vortex on surface of molten steel |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102921915A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107052320A (en) * | 2017-04-12 | 2017-08-18 | 沈阳理工大学 | Auto-pouring System and sprue cup liquid level recognition methods based on image detection |
| WO2018077044A1 (en) * | 2016-10-26 | 2018-05-03 | 宝山钢铁股份有限公司 | Control method and apparatus for inhibiting slag entrapment in ladle in last stage of pouring during continuous casting |
| CN109990972A (en) * | 2019-02-28 | 2019-07-09 | 北京首钢股份有限公司 | A kind of continuous casting water test model and method |
| CN110184401A (en) * | 2019-05-10 | 2019-08-30 | 柳州钢铁股份有限公司 | Tap cinder iron state online test method |
| CN110210427A (en) * | 2019-06-06 | 2019-09-06 | 中国民航科学技术研究院 | A kind of shelter bridge working condition detection system and method based on image processing techniques |
| CN110796046A (en) * | 2019-10-17 | 2020-02-14 | 武汉科技大学 | Intelligent steel slag detection method and system based on convolutional neural network |
| CN111414720A (en) * | 2020-02-17 | 2020-07-14 | 中国空气动力研究与发展中心计算空气动力研究所 | Flow field vortex detection method based on neural network |
| CN114523082A (en) * | 2022-03-10 | 2022-05-24 | 云南曲靖钢铁集团凤凰钢铁有限公司 | Excellent special steel continuous casting process manufacturing system capable of optimizing striking speed |
| CN115404307A (en) * | 2022-09-14 | 2022-11-29 | 福建三宝钢铁有限公司 | Preparation method of electric furnace smelting riveting screw steel ML20 |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1701877A (en) * | 2005-05-11 | 2005-11-30 | 浙江大学 | Continuous casting roughing slag inspection method and device based on vibration monitoring |
| CN101254528A (en) * | 2007-02-08 | 2008-09-03 | 浙江大学 | Bale discharging slag testing automatic control sytem |
| CN101716663A (en) * | 2009-09-17 | 2010-06-02 | 杭州谱诚泰迪实业有限公司 | Remote control system and method for roughing slag detection of continuous casting ladle |
| CN102392098A (en) * | 2011-11-04 | 2012-03-28 | 衡阳镭目科技有限责任公司 | Roughing slag detection system for steel furnace |
| CN102534097A (en) * | 2012-03-04 | 2012-07-04 | 杭州谱诚泰迪实业有限公司 | Converter tapping monitoring control system based on thermal image processing |
-
2012
- 2012-10-23 CN CN2012104090493A patent/CN102921915A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1701877A (en) * | 2005-05-11 | 2005-11-30 | 浙江大学 | Continuous casting roughing slag inspection method and device based on vibration monitoring |
| CN101254528A (en) * | 2007-02-08 | 2008-09-03 | 浙江大学 | Bale discharging slag testing automatic control sytem |
| CN101716663A (en) * | 2009-09-17 | 2010-06-02 | 杭州谱诚泰迪实业有限公司 | Remote control system and method for roughing slag detection of continuous casting ladle |
| CN102392098A (en) * | 2011-11-04 | 2012-03-28 | 衡阳镭目科技有限责任公司 | Roughing slag detection system for steel furnace |
| CN102534097A (en) * | 2012-03-04 | 2012-07-04 | 杭州谱诚泰迪实业有限公司 | Converter tapping monitoring control system based on thermal image processing |
Non-Patent Citations (2)
| Title |
|---|
| 江俊，等: "自由表面旋涡的识别方法", 《浙江大学学报（工学版）》 * |
| 谭大鹏，等: "基于振动信号测量的连铸下渣检测系统", 《浙江大学学报（工学版）》 * |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP3533535A4 (en) * | 2016-10-26 | 2020-04-22 | Baoshan Iron & Steel Co., Ltd. | Control method and apparatus for inhibiting slag entrapment in ladle in last stage of pouring during continuous casting |
| WO2018077044A1 (en) * | 2016-10-26 | 2018-05-03 | 宝山钢铁股份有限公司 | Control method and apparatus for inhibiting slag entrapment in ladle in last stage of pouring during continuous casting |
| CN107983928A (en) * | 2016-10-26 | 2018-05-04 | 宝山钢铁股份有限公司 | Con casting ladle cast latter stage suppresses slag control method and device |
| KR20190062603A (en) * | 2016-10-26 | 2019-06-05 | 바오샨 아이론 앤 스틸 유한공사 | Control method and apparatus for suppressing slag inclusion in the final stage of pouring of continuous casting ladle |
| US11154926B2 (en) | 2016-10-26 | 2021-10-26 | Baoshan Iron & Steel Co., Ltd. | Control method and apparatus for inhibiting slag entrapment in ladle in last stage of pouring during continuous casting |
| KR102251636B1 (en) * | 2016-10-26 | 2021-05-14 | 바오샨 아이론 앤 스틸 유한공사 | Control method and apparatus for suppressing the mixing of slag in the final stage of pouring of continuous casting ladle |
| CN107983928B (en) * | 2016-10-26 | 2019-11-22 | 宝山钢铁股份有限公司 | Con casting ladle is poured latter stage inhibition slag control method and device |
| CN107052320A (en) * | 2017-04-12 | 2017-08-18 | 沈阳理工大学 | Auto-pouring System and sprue cup liquid level recognition methods based on image detection |
| CN109990972A (en) * | 2019-02-28 | 2019-07-09 | 北京首钢股份有限公司 | A kind of continuous casting water test model and method |
| CN110184401B (en) * | 2019-05-10 | 2020-12-25 | 柳州钢铁股份有限公司 | On-line detection method for slag iron state of blast furnace |
| CN110184401A (en) * | 2019-05-10 | 2019-08-30 | 柳州钢铁股份有限公司 | Tap cinder iron state online test method |
| CN110210427A (en) * | 2019-06-06 | 2019-09-06 | 中国民航科学技术研究院 | A kind of shelter bridge working condition detection system and method based on image processing techniques |
| CN110796046A (en) * | 2019-10-17 | 2020-02-14 | 武汉科技大学 | Intelligent steel slag detection method and system based on convolutional neural network |
| CN110796046B (en) * | 2019-10-17 | 2023-10-10 | 武汉科技大学 | Intelligent steel slag detection method and system based on convolutional neural network |
| CN111414720A (en) * | 2020-02-17 | 2020-07-14 | 中国空气动力研究与发展中心计算空气动力研究所 | Flow field vortex detection method based on neural network |
| CN111414720B (en) * | 2020-02-17 | 2023-01-20 | 中国空气动力研究与发展中心计算空气动力研究所 | Flow field vortex detection method based on neural network |
| CN114523082A (en) * | 2022-03-10 | 2022-05-24 | 云南曲靖钢铁集团凤凰钢铁有限公司 | Excellent special steel continuous casting process manufacturing system capable of optimizing striking speed |
| CN114523082B (en) * | 2022-03-10 | 2023-08-18 | 云南曲靖钢铁集团凤凰钢铁有限公司 | Manufacturing system for continuous casting process of high-quality special steel |
| CN115404307A (en) * | 2022-09-14 | 2022-11-29 | 福建三宝钢铁有限公司 | Preparation method of electric furnace smelting riveting screw steel ML20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102921915A (en) | Slag carry-over detection method and device based on image recognition of vortex on surface of molten steel | |
| CN102319883B (en) | Method for controlling on-line prediction of continuous casting blank quality | |
| CN102896289B (en) | System and method for realizing real-time tracking of casting blank | |
| TW201224786A (en) | Method of efficacy anticipation and failure examination for an apparatus | |
| CN101251749A (en) | Ladle roughing slag detection, control method and system | |
| CN105522133B (en) | A kind of continuous casting billet bonds the decision method of behavior with copper plate of crystallizer | |
| CN103639385B (en) | Based on breakout prediction method and the system of least square | |
| CN202845727U (en) | Slagging detection device based on molten steel surface vortex image identification | |
| CN102343427A (en) | Breakout prediction method for slab continuous casting mold based on withdrawal resistance | |
| CN101972841B (en) | Continuous casting ladle slag automatic control system and control method | |
| CN105978145B (en) | Secondary system of intelligent substation information fusion system | |
| Tchórzewska-Cieślak | Water supply system reliability management | |
| CN202158839U (en) | Online real-time plate width detection device for float glass | |
| CN106950946B (en) | A kind of hydrometallurgy exception control method based on optimization principles | |
| CN103065433B (en) | Monitoring and alarming device for departure from nucleate boiling ratio (DNBR) of reactor core of pressurized water reactor | |
| CN101787969B (en) | Numerical control system of minimum flow valve of governing feedwater pump and minimum flow control method thereof | |
| CN103021180B (en) | Correction speed based traffic jam monitoring and forecasting method for macroscopic traffic flow discrete model | |
| CN117291551B (en) | Environmental monitoring early warning system based on digital visualization | |
| CN117234170B (en) | Method and device for monitoring and managing key technological parameters of steel | |
| CN203725730U (en) | Steel ladle slagging detection device | |
| CN102398006B (en) | Continuous casting blank speed detection device and detection method thereof | |
| CN108607968A (en) | A kind of forecasting procedure of the continuous casting tundish slag based on Lower proterozoic group | |
| CN116993166A (en) | Park safety risk monitoring method | |
| CN201871705U (en) | Continuous casting ladle roughing slag automatic control device | |
| CN105067028B (en) | Circulating cooling system and its monitoring method with on-line condition monitoring function |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20130213 |