CN117910620A - Tundish molten steel weight prediction method and system thereof - Google Patents
Tundish molten steel weight prediction method and system thereof Download PDFInfo
- Publication number
- CN117910620A CN117910620A CN202311837118.5A CN202311837118A CN117910620A CN 117910620 A CN117910620 A CN 117910620A CN 202311837118 A CN202311837118 A CN 202311837118A CN 117910620 A CN117910620 A CN 117910620A
- Authority
- CN
- China
- Prior art keywords
- molten steel
- tundish
- data
- model
- steel weight
- 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
- 229910000831 Steel Inorganic materials 0.000 title claims abstract description 377
- 239000010959 steel Substances 0.000 title claims abstract description 377
- 238000000034 method Methods 0.000 title claims abstract description 51
- 230000007246 mechanism Effects 0.000 claims abstract description 86
- 239000007788 liquid Substances 0.000 claims abstract description 80
- 238000012545 processing Methods 0.000 claims abstract description 14
- 241000251468 Actinopterygii Species 0.000 claims description 45
- 238000013179 statistical model Methods 0.000 claims description 32
- 238000012549 training Methods 0.000 claims description 29
- 239000012530 fluid Substances 0.000 claims description 25
- 238000009826 distribution Methods 0.000 claims description 18
- 230000000007 visual effect Effects 0.000 claims description 15
- 230000002159 abnormal effect Effects 0.000 claims description 9
- 230000000739 chaotic effect Effects 0.000 claims description 9
- 230000006870 function Effects 0.000 claims description 9
- 230000009466 transformation Effects 0.000 claims description 9
- 238000010276 construction Methods 0.000 claims description 8
- 238000006243 chemical reaction Methods 0.000 claims description 6
- 238000007789 sealing Methods 0.000 claims description 6
- 238000007781 pre-processing Methods 0.000 claims description 5
- 238000012217 deletion Methods 0.000 claims description 4
- 230000037430 deletion Effects 0.000 claims description 4
- 238000013528 artificial neural network Methods 0.000 claims description 3
- 238000012937 correction Methods 0.000 claims description 3
- 230000003247 decreasing effect Effects 0.000 claims description 3
- 238000000605 extraction Methods 0.000 claims description 3
- 238000013507 mapping Methods 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 238000005303 weighing Methods 0.000 abstract description 14
- 230000008569 process Effects 0.000 abstract description 7
- 238000009851 ferrous metallurgy Methods 0.000 abstract description 2
- 238000009749 continuous casting Methods 0.000 description 4
- 238000004519 manufacturing process Methods 0.000 description 4
- 238000005266 casting Methods 0.000 description 3
- 238000013079 data visualisation Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000004907 flux Effects 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 239000002893 slag Substances 0.000 description 2
- 230000003321 amplification Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 239000003795 chemical substances by application Substances 0.000 description 1
- 238000007599 discharging Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 239000000178 monomer Substances 0.000 description 1
- 238000003199 nucleic acid amplification method Methods 0.000 description 1
- 238000009628 steelmaking Methods 0.000 description 1
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Continuous Casting (AREA)
Abstract
The invention belongs to the technical field of ferrous metallurgy, and discloses a tundish molten steel weight prediction method and a tundish molten steel weight prediction system. The method comprises the following steps: establishing a plurality of tundish molten steel weight mechanism models; according to a plurality of tundish molten steel weight mechanism models, a model database is established; according to the number of the current tundish, matching a corresponding tundish molten steel weight mechanism model in a model database; and inputting the current liquid level data of the molten steel of the tundish into a corresponding tundish molten steel weight mechanism model, and predicting the molten steel weight to obtain a current tundish molten steel weight prediction result. The system comprises a factory server, a data server and a Web server, wherein the data server comprises a data processing unit, a model building unit, a database unit and a molten steel weight prediction unit. The invention solves the problems of poor data accuracy and quality, low intelligent degree, large hardware cost investment and complex weighing process in the prior art.
Description
Technical Field
The invention belongs to the technical field of ferrous metallurgy, and particularly relates to a tundish molten steel weight prediction method and a tundish molten steel weight prediction system.
Background
Ladle is generally referred to as ladle. The ladle is used for carrying molten steel in front of an open hearth furnace, an electric furnace or a converter in a steel plant or a foundry for casting operation. The tundish is a refractory container used in short-flow steelmaking, and is used for receiving molten steel poured from a ladle first and then distributing the molten steel into each crystallizer through a tundish nozzle. The steel flux of the ladle flowing into the tundish is matched with the steel flux of the tundish flowing into the crystallizer so as to keep the tonnage of the tundish stable. The weight information of the continuous casting tundish molten steel is critical to the processes of casting, quick change and the like, and the process parameters of automatic control of the liquid level of the tundish molten steel, pull speed control, tail blank discharging and the like of operators are directly influenced in the casting process, so that the production quality of the continuous casting blank is influenced, the weight of the tundish molten steel is required to be predicted and calculated, and the stable control of the tonnage of the tundish is served.
At present, the traditional mode is to adopt a weighing mode to acquire the weight of a tundish, a plurality of weighing sensors are arranged below the tundish, the weighing sensors output corresponding analog signals, the analog signals are sent into a PLC (programmable logic controller) weighing module to be subjected to filtering amplification processing and converted into digital signals, and then the digital signals are subjected to data exchange with a CPU (central processing unit) of the PLC to finish the weighing task of the weight of the tundish molten steel. However, this weighing method has some disadvantages: the covering agent and the steel slag in the tundish are contained in the total weight, so that the problem that the weight data obtained by the weighing module is larger than the actual molten steel weight exists, the weight weighing accuracy is poor, the hidden danger of slag falling from a crystallizer exists, and the final production quality is affected; meanwhile, the weighing sensor is also affected and interfered by the field environment, so that the weighing data is abnormal, and the data quality is low; moreover, the existing weighing mode is low in intelligent degree only by means of a hardware sensor, large in hardware cost investment and complex in weighing process, and cannot meet the requirements of intelligent factory construction.
Disclosure of Invention
The invention aims to solve the problems of poor data accuracy and quality, low intelligent degree, large hardware cost investment and complex weighing process in the prior art, and provides a tundish molten steel weight prediction method and a tundish molten steel weight prediction system.
The technical scheme adopted by the invention is as follows:
A tundish molten steel weight prediction method comprises the following steps:
According to the tundish physical data and the serial numbers of a plurality of historical tundish, a plurality of corresponding tundish molten steel weight mechanism models are established;
According to a plurality of tundish molten steel weight mechanism models, a model database is established;
according to the number of the current tundish, matching a corresponding tundish molten steel weight mechanism model in a model database;
and inputting the current liquid level data of the molten steel of the tundish into a corresponding tundish molten steel weight mechanism model, and predicting the molten steel weight to obtain a current tundish molten steel weight prediction result.
Further, according to tundish physical data of a plurality of historical tundish and the serial numbers thereof, a plurality of corresponding tundish molten steel weight mechanism models are established, and the method comprises the following steps:
Constructing a corresponding tundish structure three-dimensional model according to the historical tundish physical data of the tundish;
Constructing a tundish molten steel fluid domain three-dimensional model according to the tundish structure three-dimensional model;
constructing a tundish molten steel volume statistical model according to the three-dimensional model of the tundish molten steel fluid domain;
and constructing a tundish molten steel weight mechanism model according to the tundish molten steel volume statistical model, and taking the number of the tundish as a matching label of the corresponding tundish molten steel weight mechanism model.
Further, the physical data of the tundish comprise a structural drawing, an actual size and a three-dimensional scanning video of the tundish;
according to the physical data of the tundish, constructing a three-dimensional model of the tundish structure, comprising the following steps:
Carrying out three-dimensional modeling according to the structural drawing of the tundish, and constructing an initial three-dimensional model of the tundish structure;
Performing size adjustment on the initial three-dimensional model of the tundish structure according to the actual size of the tundish to obtain the three-dimensional model of the tundish structure with the actual size;
and carrying out model correction on the actual-size three-dimensional model of the tundish structure according to the three-dimensional scanning video of the tundish to obtain a final three-dimensional model of the tundish structure.
Further, according to the three-dimensional model of the tundish structure, a three-dimensional model of the tundish molten steel fluid domain is constructed, and the method comprises the following steps:
extracting a plurality of side rings of a molten steel area of the three-dimensional model of the tundish structure by using a volume extraction tool;
extracting a plurality of source surfaces and a plurality of sealing surfaces corresponding to the molten steel area according to a plurality of side rings of the molten steel area;
extracting a molten steel fluid domain according to a plurality of source surfaces and a plurality of sealing surfaces of the molten steel region;
and constructing a three-dimensional model of the molten steel fluid domain of the tundish according to the molten steel fluid domain.
Further, according to the three-dimensional model of the liquid steel fluid domain of the tundish, a statistical model of the liquid steel volume of the tundish is constructed, and the method comprises the following steps:
Acquiring a plurality of molten steel liquid level height data and a plurality of molten steel volume data corresponding to the molten steel liquid level heights one by one according to the three-dimensional model of the molten steel liquid domain of the tundish;
Carrying out data preprocessing on a plurality of molten steel liquid level height data and a plurality of molten steel volume data to obtain an effective data pair training set;
establishing an initial tundish molten steel volume statistical model by using IAFSA-AdaBoost algorithm;
Inputting the effective data into an initial tundish molten steel volume statistical model for optimization training to obtain a fitting relation curve of molten steel liquid level height and molten steel volume, wherein the formula of the fitting relation curve is as follows:
V(h)＝Ah2+Bh+C
Wherein V (h) is a molten steel volume function; h is the liquid level height of molten steel; A. b, C are fitting constants;
And adjusting parameters of the initial tundish molten steel volume statistical model according to the fitting relation curve to obtain a final tundish molten steel volume statistical model.
Further, the data preprocessing comprises the following steps:
taking the liquid level height data of the molten steel as an independent variable, taking the liquid volume data of the molten steel as an independent variable, and taking the liquid level height data of the molten steel and the corresponding liquid volume data of the molten steel as an initial data pair;
deleting the repeated data pairs from the plurality of initial data pairs to obtain a first data pair training set without repeated data;
Abnormal data pair deletion is carried out on the first data pair training set by using a 3 sigma principle, and a second data pair training set without abnormal data is obtained;
deleting the missing data pair of the second data pair training set to obtain a third data pair training set without missing data;
and carrying out normalization processing on the third data pair training set to obtain an effective data pair training set.
Further, introducing a Circle chaotic sequence initialization and dynamic reverse learning strategy and improving a traditional artificial fish swarm algorithm based on a Cauchy distribution self-adaptive artificial fish visual field to obtain IAFSA optimizing algorithm, and optimizing network parameters of an AdaBoost neural network by using the IAFSA optimizing algorithm to obtain a IAFSA-AdaBoost algorithm;
The Circle chaotic sequence initialization formula is:
Wherein x i+1,j+1 is the initial position of the circular chaotic mapping artificial fish school; x i,j is the initial position of the artificial fish school generated randomly; mod (-) is a mod function; i is an indication quantity of artificial fish; j is a dimension indicating quantity;
The formula of the dynamic reverse learning strategy is:
x′ij(t)＝k(aj(t)+bj(t))-xij(t)
Wherein x' ij(t)、xij (t) is the reverse position and the forward position of the j-th dimension of the i-th artificial fish respectively; a j(t)、bj (t) is the upper bound and the lower bound of the j-th dimension of the current artificial fish school respectively; k is a decreasing inertia factor, and k=0.9-0.5D/D max;D、Dmax is the current iteration number and the maximum iteration number, respectively; t is a time indication quantity;
the formula for updating the visual field range of the self-adaptive artificial fish based on the Cauchy distribution is as follows:
Wherein v (x i) is a visual field range updating function of the artificial fish; v is the original field of view of the artificial fish; v C(xi) is the visual field range of the artificial fish in the latter half of iteration after the Kexiong distribution transformation; x i (t) is the position of the ith artificial fish; D. d max is the current iteration number and the maximum iteration number respectively;
the formula of the field of view after the Kexiong distribution transformation is:
Wherein v C(xi) is the visual field range of the artificial fish after the Cauchy distribution transformation in the latter half of the iteration; f (x i) is the fitness value of the ith artificial fish; f worst is the worst fitness value of the artificial fish; gamma is a scale parameter, and the smaller the value thereof, the steeper the cauchy distribution probability density curve, and gamma=1/(pi·v origin);vorigin is an initial field value.
Further, according to the tundish molten steel volume statistical model, a tundish molten steel weight mechanism model is constructed, and the number of the tundish is used as a matching label of the corresponding tundish molten steel weight mechanism model, and the method comprises the following steps:
according to the mechanism among the weight, the volume and the density of molten steel of the tundish, a tundish molten steel weight mechanism model is constructed, and the mechanism formula of the tundish molten steel weight mechanism model is as follows:
M＝ρ·V(h)
wherein M is a molten steel weight prediction result; ρ is the density of molten steel; v (h) is the volume of molten steel; h is the liquid level height of molten steel;
establishing a link relation between a tundish molten steel weight mechanism model and a corresponding tundish molten steel volume statistical model;
and taking the number of the tundish as a matching label of a corresponding tundish molten steel weight mechanism model.
Further, the current tundish molten steel level height data is input into a matched tundish molten steel weight mechanism model to predict the molten steel weight, and the method comprises the following steps:
acquiring the current liquid level data of the molten steel of the tundish, and inputting the liquid level data of the molten steel into a tundish molten steel weight mechanism model;
According to the liquid level data of the molten steel, a corresponding tundish molten steel volume statistical model is called, and molten steel volume prediction is carried out to obtain a corresponding molten steel volume;
And according to the molten steel volume returned by the tundish molten steel volume statistical model, predicting the molten steel weight by using a tundish molten steel weight mechanism model to obtain a molten steel weight prediction result.
The tundish molten steel weight prediction system is applied to a tundish molten steel weight prediction method and comprises a factory server, a data server and a Web server, wherein the factory server, the data server and the Web server are sequentially connected, and the Web server is connected with a plurality of external query terminals;
The factory server is used for collecting tundish physical data and numbers of a plurality of historical tundish, and the current tundish molten steel level height data and numbers, and sending the collected data to the data server;
The data server is used for establishing a tundish molten steel weight mechanism model according to the historical tundish physical data and the serial numbers thereof of the tundish sent by the factory server, matching the corresponding tundish molten steel weight mechanism model according to the current tundish molten steel level data and the serial numbers thereof sent by the factory server, carrying out molten steel weight prediction, and sending the current tundish molten steel level data, the serial numbers thereof and the corresponding molten steel weight prediction results to the Web server;
the Web server is used for providing Web pages, and visually displaying the molten steel liquid level height data, the number and the corresponding molten steel weight prediction result to the query terminal through the Web pages;
The data server comprises a data processing unit, a model building unit, a database unit and a molten steel weight prediction unit, wherein the data processing unit is respectively connected with the factory server, the model building unit and the molten steel weight prediction unit;
The data processing unit is used for converting the data format of the tundish physical data and the serial numbers of the historical tundish sent by the factory server and the current tundish molten steel liquid level data and the serial numbers thereof, sending the tundish physical data and the serial numbers after the data format conversion to the model construction unit and sending the current tundish molten steel liquid level data after the data format conversion to the molten steel weight prediction unit;
The model construction unit is used for establishing a plurality of corresponding tundish molten steel weight mechanism models according to the historical tundish physical data and the serial numbers of the tundish physical data, and sending metadata of the tundish molten steel weight mechanism models to the database unit;
the database unit is used for establishing a model database according to the received metadata of the weight mechanism models of the plurality of tundish molten steels;
The molten steel weight prediction unit is used for matching the corresponding tundish molten steel weight mechanism model in the model database according to the number of the current tundish, inputting the current tundish molten steel level data into the corresponding tundish molten steel weight mechanism model, predicting the molten steel weight to obtain the current tundish molten steel weight prediction result, and sending the current tundish molten steel level data, the number thereof and the corresponding molten steel weight prediction result to the Web server.
The beneficial effects of the invention are as follows:
1) According to the tundish molten steel weight prediction method, the tundish molten steel weight mechanism model is established through the tundish physical data of the tundish and the mechanisms among the tundish molten steel weight, the tundish molten steel volume and the tundish molten steel density, so that the intelligent automatic prediction of the tundish molten steel weight is realized, the molten steel weight acquisition mode is simplified, the corresponding molten steel weight can be obtained according to the input molten steel level height data, the intelligent degree and the accuracy and the reliability of the data are improved, meanwhile, a model database is established according to a plurality of tundish molten steel weight mechanism models, the method can be suitable for the tundish of different models, and the practicability of the tundish molten steel weight mechanism model and the applicability to actual scenes are improved.
2) According to the tundish molten steel weight prediction system provided by the invention, historical data and on-site real-time data are acquired through the factory server, the addition of external hardware equipment is avoided, the hardware cost investment is reduced, the influence of external factors such as on-site environment and the like on errors of sensor hardware is avoided, the continuous casting production is better served finally, the real-time on-line automatic prediction of the tundish molten steel weight is provided by the data server, the tundish molten steel weight mechanism model is used for the real-time on-line automatic prediction of the tundish molten steel weight, the intelligent degree of the system is improved, the molten steel weight acquisition process is simplified, the Web server provides Web pages for data visualization, the real-time inquiry and the monitoring of the molten steel weight prediction result by a query terminal are facilitated, and the practicability of the system is improved.
Other advantageous effects of the present invention will be further described in the detailed description.
Drawings
FIG. 1 is a block flow diagram of a tundish molten steel weight prediction method in the present invention.
FIG. 2 is a block diagram showing the construction of a tundish molten steel weight prediction system according to the present invention.
Detailed Description
The invention is further illustrated by the following description of specific embodiments in conjunction with the accompanying drawings.
Example 1:
As shown in fig. 1, the embodiment provides a tundish molten steel weight prediction method, which includes the following steps:
According to the tundish physical data and the numbers of a plurality of historical tundish, a plurality of corresponding tundish molten steel weight mechanism models are established, and the method comprises the following steps:
constructing a corresponding tundish structure three-dimensional model according to the historical tundish physical data of the tundish; the tundish physical data comprises a structural drawing, an actual size and a three-dimensional scanning video of the tundish;
according to the physical data of the tundish, constructing a three-dimensional model of the tundish structure, comprising the following steps:
Carrying out three-dimensional modeling according to the structural drawing of the tundish, and constructing an initial three-dimensional model of the tundish structure;
Performing size adjustment on the initial three-dimensional model of the tundish structure according to the actual size of the tundish to obtain the three-dimensional model of the tundish structure with the actual size;
carrying out model correction on the actual-size three-dimensional model of the tundish structure according to the three-dimensional scanning video of the tundish to obtain a final three-dimensional model of the tundish structure;
according to the three-dimensional model of the tundish structure, constructing a three-dimensional model of the tundish molten steel fluid domain, comprising the following steps:
extracting a plurality of side rings of a molten steel area of the three-dimensional model of the tundish structure by using a volume extraction tool;
extracting a plurality of source surfaces and a plurality of sealing surfaces corresponding to the molten steel area according to a plurality of side rings of the molten steel area;
extracting a molten steel fluid domain according to a plurality of source surfaces and a plurality of sealing surfaces of the molten steel region;
Constructing a three-dimensional model of the liquid steel fluid domain of the tundish according to the liquid steel fluid domain;
By digital three-dimensional realization of the physical structure of the tundish, the accurate and true restoration of the three-dimensional model of the liquid steel fluid domain of the tundish to the liquid steel fluid domain is improved, a three-dimensional model support is provided for the subsequent liquid steel volume calculation, and the authenticity and accuracy of the liquid steel liquid level height data and the liquid steel volume data are ensured;
According to the three-dimensional model of the liquid steel fluid domain of the tundish, a statistical model of the liquid steel volume of the tundish is built, and the method comprises the following steps:
acquiring a plurality of molten steel liquid level height data and a plurality of molten steel volume data corresponding to the molten steel liquid level heights one by one according to the three-dimensional model of the molten steel liquid domain of the tundish; in the three-dimensional model of the liquid steel fluid domain of the tundish, only one piece of liquid steel volume data of each liquid steel fluid level is matched with the liquid steel fluid domain of the tundish, so that a certain corresponding relation exists between the liquid steel fluid level data and the liquid steel volume data, and the corresponding relation can be learned through a statistical model to realize the prediction of the liquid steel volume data;
Carrying out data preprocessing on a plurality of molten steel liquid level height data and a plurality of molten steel volume data to obtain an effective data pair training set, wherein the method comprises the following steps:
taking the liquid level height data of the molten steel as an independent variable, taking the liquid volume data of the molten steel as an independent variable, and taking the liquid level height data of the molten steel and the corresponding liquid volume data of the molten steel as an initial data pair;
deleting the repeated data pairs from the plurality of initial data pairs to obtain a first data pair training set without repeated data;
Abnormal data pair deletion is carried out on the first data pair training set by using a 3 sigma principle, and a second data pair training set without abnormal data is obtained;
the formula for abnormal data pair deletion using the 3σ principle is:
wherein sigma is the standard deviation of a sample; x i′ is a sample value, wherein each monomer data in the abnormal data pair is a sample; Is the sample mean value; i' is a sample indication; i is the total number of samples;
deleting the missing data pair of the second data pair training set to obtain a third data pair training set without missing data;
Normalizing the training set of the third data to obtain an effective data pair training set;
Wherein q i′_new is normalized sample data; q i′ is the pre-normalization sample data; i' is a sample indication; q max、qmin is the maximum and minimum values of the sample data, respectively;
Establishing an initial tundish molten steel volume statistical model by using IAFSA-AdaBoost algorithm; introducing a Circle chaotic sequence initialization and dynamic reverse learning strategy and improving a traditional artificial fish swarm algorithm based on a Cauchy distribution self-adaptive artificial fish visual field to obtain IAFSA optimizing algorithm, and optimizing network parameters of an AdaBoost neural network by using the IAFSA optimizing algorithm to obtain a IAFSA-AdaBoost algorithm;
The Circle chaotic sequence initialization formula is:
Wherein x i+1,j+1 is the initial position of the circular chaotic mapping artificial fish school; x i,j is the initial position of the artificial fish school generated randomly; mod (-) is a mod function; i is an indication quantity of artificial fish; j is a dimension indicating quantity;
The formula of the dynamic reverse learning strategy is:
x′ij(t)＝k(aj(t)+bj(t))-xij(t)
Wherein x' ij(t)、xij (t) is the reverse position and the forward position of the j-th dimension of the i-th artificial fish respectively; a j(t)、bj (t) is the upper bound and the lower bound of the j-th dimension of the current artificial fish school respectively; k is a decreasing inertia factor, and k=0.9-0.5D/D max;D、Dmax is the current iteration number and the maximum iteration number, respectively; t is a time indication quantity;
the formula for updating the visual field range of the self-adaptive artificial fish based on the Cauchy distribution is as follows:
Wherein v (x i) is a visual field range updating function of the artificial fish; v is the original field of view of the artificial fish; v C(xi) is the visual field range of the artificial fish in the latter half of iteration after the Kexiong distribution transformation; x i (t) is the position of the ith artificial fish; D. d max is the current iteration number and the maximum iteration number respectively;
the formula of the field of view after the Kexiong distribution transformation is:
Wherein v C(xi) is the visual field range of the artificial fish after the Cauchy distribution transformation in the latter half of the iteration; f (x i) is the fitness value of the ith artificial fish; f worst is the worst fitness value of the artificial fish; gamma is a scale parameter, the smaller the value of the gamma is, the steeper the Koxie distribution probability density curve is, and gamma=1/(pi·v origin);vorigin is an initial view value;
Inputting the effective data into an initial tundish molten steel volume statistical model for optimization training to obtain a fitting relation curve of molten steel liquid level height and molten steel volume, wherein the formula of the fitting relation curve is as follows:
V(h)＝Ah2+Bh+C
Wherein V (h) is a molten steel volume function; h is the liquid level height of molten steel; A. b, C are fitting constants; in this embodiment, in the training set, when the liquid level height of molten steel is 500mm, the volume of molten steel is 3.25e 9mm3, and meanwhile, the polynomial fitting is performed on the training set by inputting the effective data corresponding to 550-850mm, the height unit mm, the volume unit m 3, the fitting constant a=2e -6, b=0.0056, and c= 0.0119, so that the fitting relation curve of the liquid level height of molten steel and the volume of molten steel is obtained as follows:
V(h)＝2e-6h2+0.0056h+0.0119
According to the fitting relation curve, adjusting parameters of the initial tundish molten steel volume statistical model to obtain a final tundish molten steel volume statistical model;
according to the tundish molten steel volume statistical model, a tundish molten steel weight mechanism model is built, and the number of the tundish is used as a matching label of the corresponding tundish molten steel weight mechanism model, and the method comprises the following steps:
according to the mechanism among the weight, the volume and the density of molten steel of the tundish, a tundish molten steel weight mechanism model is constructed, and the mechanism formula of the tundish molten steel weight mechanism model is as follows:
M＝ρ·V(h)
wherein M is a molten steel weight prediction result; ρ is the density of molten steel; v (h) is the volume of molten steel; h is the liquid level height of molten steel;
establishing a link relation between a tundish molten steel weight mechanism model and a corresponding tundish molten steel volume statistical model;
Taking the number of the tundish as a matching label of a corresponding tundish molten steel weight mechanism model;
According to a plurality of tundish molten steel weight mechanism models, a model database is established; the model database is established to store tundish molten steel weight mechanism models with different models and different molten steel capacities, the number of the tundish is used as a retrieval tag, the practicability and the working efficiency of tundish molten steel weight prediction are improved, and the intelligent degree of the method is improved by rapidly retrieving the corresponding models to predict the molten steel weight;
according to the number of the current tundish, matching a corresponding tundish molten steel weight mechanism model in a model database;
the method comprises the following steps of inputting the current liquid level data of the molten steel of the tundish into a corresponding tundish molten steel weight mechanism model, and predicting the molten steel weight to obtain a current tundish molten steel weight prediction result, wherein the method comprises the following steps:
acquiring the current liquid level data of the molten steel of the tundish, and inputting the liquid level data of the molten steel into a tundish molten steel weight mechanism model;
According to the liquid level data of the molten steel, a corresponding tundish molten steel volume statistical model is called, and molten steel volume prediction is carried out to obtain a corresponding molten steel volume;
And according to the molten steel volume returned by the tundish molten steel volume statistical model, predicting the molten steel weight by using a tundish molten steel weight mechanism model to obtain a molten steel weight prediction result.
According to the tundish molten steel weight prediction method, the tundish molten steel weight mechanism model is established through the tundish physical data of the tundish and the mechanisms among the tundish molten steel weight, the tundish molten steel volume and the tundish molten steel density, so that the intelligent automatic prediction of the tundish molten steel weight is realized, the molten steel weight acquisition mode is simplified, the corresponding molten steel weight can be obtained according to the input molten steel level height data, the intelligent degree and the accuracy and the reliability of the data are improved, meanwhile, a model database is established according to a plurality of tundish molten steel weight mechanism models, the method can be suitable for the tundish of different models, and the practicability of the tundish molten steel weight mechanism model and the applicability to actual scenes are improved.
Example 2:
As shown in fig. 2, the present embodiment provides a tundish molten steel weight prediction system, which is applied to a tundish molten steel weight prediction method, and includes a factory server, a data server and a Web server, wherein the factory server, the data server and the Web server are sequentially connected, and the Web server is connected with a plurality of external query terminals;
The factory server is used for collecting tundish physical data and numbers of a plurality of historical tundish, and the current tundish molten steel level height data and numbers, and sending the collected data to the data server;
The data server is used for establishing a tundish molten steel weight mechanism model according to the historical tundish physical data and the serial numbers thereof of the tundish sent by the factory server, matching the corresponding tundish molten steel weight mechanism model according to the current tundish molten steel level data and the serial numbers thereof sent by the factory server, carrying out molten steel weight prediction, and sending the current tundish molten steel level data, the serial numbers thereof and the corresponding molten steel weight prediction results to the Web server;
the Web server is used for providing Web pages, and visually displaying the molten steel liquid level height data, the number and the corresponding molten steel weight prediction result to the query terminal through the Web pages;
The data server comprises a data processing unit, a model building unit, a database unit and a molten steel weight prediction unit, wherein the data processing unit is respectively connected with the factory server, the model building unit and the molten steel weight prediction unit;
The data processing unit is used for converting the data format of the tundish physical data and the serial numbers of the historical tundish sent by the factory server and the current tundish molten steel liquid level data and the serial numbers thereof, sending the tundish physical data and the serial numbers after the data format conversion to the model construction unit and sending the current tundish molten steel liquid level data after the data format conversion to the molten steel weight prediction unit;
The model construction unit is used for establishing a plurality of corresponding tundish molten steel weight mechanism models according to the historical tundish physical data and the serial numbers of the tundish physical data, and sending metadata of the tundish molten steel weight mechanism models to the database unit;
the database unit is used for establishing a model database according to the received metadata of the weight mechanism models of the plurality of tundish molten steels;
The molten steel weight prediction unit is used for matching the corresponding tundish molten steel weight mechanism model in the model database according to the number of the current tundish, inputting the current tundish molten steel level data into the corresponding tundish molten steel weight mechanism model, predicting the molten steel weight to obtain the current tundish molten steel weight prediction result, and sending the current tundish molten steel level data, the number thereof and the corresponding molten steel weight prediction result to the Web server.
According to the tundish molten steel weight prediction system provided by the invention, historical data and on-site real-time data are acquired through the factory server, the addition of external hardware equipment is avoided, the hardware cost investment is reduced, the influence of external factors such as on-site environment and the like on errors of sensor hardware is avoided, the continuous casting production is better served finally, the real-time on-line automatic prediction of the tundish molten steel weight is provided by the data server, the tundish molten steel weight mechanism model is used for the real-time on-line automatic prediction of the tundish molten steel weight, the intelligent degree of the system is improved, the molten steel weight acquisition process is simplified, the Web server provides Web pages for data visualization, the real-time inquiry and the monitoring of the molten steel weight prediction result by a query terminal are facilitated, and the practicability of the system is improved.
The invention is not limited to the alternative embodiments described above, but any person may derive other various forms of products in the light of the present invention. The above detailed description should not be construed as limiting the scope of the invention, which is defined in the claims and the description may be used to interpret the claims.
Claims (10)
1. A tundish molten steel weight prediction method is characterized by comprising the following steps of: the method comprises the following steps:
According to the tundish physical data and the serial numbers of a plurality of historical tundish, a plurality of corresponding tundish molten steel weight mechanism models are established;
According to a plurality of tundish molten steel weight mechanism models, a model database is established;
according to the number of the current tundish, matching a corresponding tundish molten steel weight mechanism model in a model database;
and inputting the current liquid level data of the molten steel of the tundish into a corresponding tundish molten steel weight mechanism model, and predicting the molten steel weight to obtain a current tundish molten steel weight prediction result.
2. The tundish molten steel weight prediction method according to claim 1, characterized by comprising the steps of: according to the tundish physical data and the numbers of a plurality of historical tundish, a plurality of corresponding tundish molten steel weight mechanism models are established, and the method comprises the following steps:
Constructing a corresponding tundish structure three-dimensional model according to the historical tundish physical data of the tundish;
Constructing a tundish molten steel fluid domain three-dimensional model according to the tundish structure three-dimensional model;
constructing a tundish molten steel volume statistical model according to the three-dimensional model of the tundish molten steel fluid domain;
and constructing a tundish molten steel weight mechanism model according to the tundish molten steel volume statistical model, and taking the number of the tundish as a matching label of the corresponding tundish molten steel weight mechanism model.
3. The tundish molten steel weight prediction method according to claim 2, characterized by: the tundish physical data comprise a structural drawing, an actual size and a three-dimensional scanning video of the tundish;
according to the physical data of the tundish, constructing a three-dimensional model of the tundish structure, comprising the following steps:
Carrying out three-dimensional modeling according to the structural drawing of the tundish, and constructing an initial three-dimensional model of the tundish structure;
Performing size adjustment on the initial three-dimensional model of the tundish structure according to the actual size of the tundish to obtain the three-dimensional model of the tundish structure with the actual size;
and carrying out model correction on the actual-size three-dimensional model of the tundish structure according to the three-dimensional scanning video of the tundish to obtain a final three-dimensional model of the tundish structure.
4. The tundish molten steel weight prediction method according to claim 2, characterized by: according to the three-dimensional model of the tundish structure, constructing a three-dimensional model of the tundish molten steel fluid domain, comprising the following steps:
extracting a plurality of side rings of a molten steel area of the three-dimensional model of the tundish structure by using a volume extraction tool;
extracting a plurality of source surfaces and a plurality of sealing surfaces corresponding to the molten steel area according to a plurality of side rings of the molten steel area;
extracting a molten steel fluid domain according to a plurality of source surfaces and a plurality of sealing surfaces of the molten steel region;
and constructing a three-dimensional model of the molten steel fluid domain of the tundish according to the molten steel fluid domain.
5. The tundish molten steel weight prediction method according to claim 2, characterized by: according to the three-dimensional model of the liquid steel fluid domain of the tundish, a statistical model of the liquid steel volume of the tundish is built, and the method comprises the following steps:
Acquiring a plurality of molten steel liquid level height data and a plurality of molten steel volume data corresponding to the molten steel liquid level heights one by one according to the three-dimensional model of the molten steel liquid domain of the tundish;
Carrying out data preprocessing on a plurality of molten steel liquid level height data and a plurality of molten steel volume data to obtain an effective data pair training set;
establishing an initial tundish molten steel volume statistical model by using IAFSA-AdaBoost algorithm;
Inputting the effective data into an initial tundish molten steel volume statistical model for optimization training to obtain a fitting relation curve of molten steel liquid level height and molten steel volume, wherein the formula of the fitting relation curve is as follows:
V(h)＝Ah2+Bh+C
Wherein V (h) is a molten steel volume function; h is the liquid level height of molten steel; A. b, C are fitting constants;
And adjusting parameters of the initial tundish molten steel volume statistical model according to the fitting relation curve to obtain a final tundish molten steel volume statistical model.
6. The tundish molten steel weight prediction method according to claim 5, characterized by: the data preprocessing comprises the following steps:
taking the liquid level height data of the molten steel as an independent variable, taking the liquid volume data of the molten steel as an independent variable, and taking the liquid level height data of the molten steel and the corresponding liquid volume data of the molten steel as an initial data pair;
deleting the repeated data pairs from the plurality of initial data pairs to obtain a first data pair training set without repeated data;
Abnormal data pair deletion is carried out on the first data pair training set by using a 3 sigma principle, and a second data pair training set without abnormal data is obtained;
deleting the missing data pair of the second data pair training set to obtain a third data pair training set without missing data;
and carrying out normalization processing on the third data pair training set to obtain an effective data pair training set.
7. The tundish molten steel weight prediction method according to claim 5, characterized by: introducing a Circle chaotic sequence initialization and dynamic reverse learning strategy and improving a traditional artificial fish swarm algorithm based on a Cauchy distribution self-adaptive artificial fish visual field to obtain IAFSA optimizing algorithm, and optimizing network parameters of an AdaBoost neural network by using the IAFSA optimizing algorithm to obtain a IAFSA-AdaBoost algorithm;
The Circle chaotic sequence initialization formula is:
Wherein x i+1,j+1 is the initial position of the circular chaotic mapping artificial fish school; x i,j is the initial position of the artificial fish school generated randomly; mod (-) is a mod function; i is an indication quantity of artificial fish; j is a dimension indicating quantity;
The formula of the dynamic reverse learning strategy is:
x′ij(t)＝k(aj(t)+bj(t))-xij(t)
Wherein x' ij(t)、xij (t) is the reverse position and the forward position of the j-th dimension of the i-th artificial fish respectively; a j(t)、bj (t) is the upper bound and the lower bound of the j-th dimension of the current artificial fish school respectively; k is a decreasing inertia factor, and k=0.9-0.5D/D max;D、Dmax is the current iteration number and the maximum iteration number, respectively; t is a time indication quantity;
the formula for updating the visual field range of the self-adaptive artificial fish based on the Cauchy distribution is as follows:
Wherein v (x i) is a visual field range updating function of the artificial fish; v is the original field of view of the artificial fish; v C(xi) is the visual field range of the artificial fish in the latter half of iteration after the Kexiong distribution transformation; x i (t) is the position of the ith artificial fish; D. d max is the current iteration number and the maximum iteration number respectively;
the formula of the field of view after the Kexiong distribution transformation is:
Wherein v C(xi) is the visual field range of the artificial fish after the Cauchy distribution transformation in the latter half of the iteration; f (x i) is the fitness value of the ith artificial fish; f worst is the worst fitness value of the artificial fish; gamma is a scale parameter, and the smaller the value thereof, the steeper the cauchy distribution probability density curve, and gamma=1/(pi·v origin);vorigin is an initial field value.
8. The tundish molten steel weight prediction method according to claim 2, characterized by: according to the tundish molten steel volume statistical model, a tundish molten steel weight mechanism model is built, and the number of the tundish is used as a matching label of the corresponding tundish molten steel weight mechanism model, and the method comprises the following steps:
according to the mechanism among the weight, the volume and the density of molten steel of the tundish, a tundish molten steel weight mechanism model is constructed, and the mechanism formula of the tundish molten steel weight mechanism model is as follows:
M＝ρ·V(h)
wherein M is a molten steel weight prediction result; ρ is the density of molten steel; v (h) is the volume of molten steel; h is the liquid level height of molten steel;
establishing a link relation between a tundish molten steel weight mechanism model and a corresponding tundish molten steel volume statistical model;
and taking the number of the tundish as a matching label of a corresponding tundish molten steel weight mechanism model.
9. The tundish molten steel weight prediction method according to claim 8, characterized by: the method comprises the following steps of inputting the current liquid level data of the molten steel of the tundish into a matched tundish molten steel weight mechanism model, and predicting the molten steel weight, wherein the method comprises the following steps of:
acquiring the current liquid level data of the molten steel of the tundish, and inputting the liquid level data of the molten steel into a tundish molten steel weight mechanism model;
According to the liquid level data of the molten steel, a corresponding tundish molten steel volume statistical model is called, and molten steel volume prediction is carried out to obtain a corresponding molten steel volume;
And according to the molten steel volume returned by the tundish molten steel volume statistical model, predicting the molten steel weight by using a tundish molten steel weight mechanism model to obtain a molten steel weight prediction result.
10. A tundish molten steel weight prediction system, applied to a tundish molten steel weight prediction method as set forth in any one of claims 1 to 9, characterized in that: the system comprises a factory server, a data server and a Web server, wherein the factory server, the data server and the Web server are sequentially connected, and the Web server is connected with a plurality of external query terminals;
The factory server is used for collecting tundish physical data and numbers of a plurality of historical tundish, and the current tundish molten steel level height data and numbers, and sending the collected data to the data server;
The data server is used for establishing a tundish molten steel weight mechanism model according to the historical tundish physical data and the serial numbers thereof of the tundish sent by the factory server, matching the corresponding tundish molten steel weight mechanism model according to the current tundish molten steel level data and the serial numbers thereof sent by the factory server, carrying out molten steel weight prediction, and sending the current tundish molten steel level data, the serial numbers thereof and the corresponding molten steel weight prediction results to the Web server;
the Web server is used for providing Web pages, and visually displaying the molten steel liquid level height data, the number and the corresponding molten steel weight prediction result to the query terminal through the Web pages;
The data server comprises a data processing unit, a model building unit, a database unit and a molten steel weight prediction unit, wherein the data processing unit is respectively connected with the factory server, the model building unit and the molten steel weight prediction unit are both connected with the database unit, and the molten steel weight prediction unit is connected with the Web server;
The data processing unit is used for converting the data format of the tundish physical data and the serial numbers of the historical tundish sent by the factory server and the current tundish molten steel liquid level data and the serial numbers thereof, sending the tundish physical data and the serial numbers after the data format conversion to the model construction unit and sending the current tundish molten steel liquid level data after the data format conversion to the molten steel weight prediction unit;
The model construction unit is used for establishing a plurality of corresponding tundish molten steel weight mechanism models according to the historical tundish physical data and the serial numbers of the tundish physical data, and sending metadata of the tundish molten steel weight mechanism models to the database unit;
the database unit is used for establishing a model database according to the received metadata of the weight mechanism models of the plurality of tundish molten steels;
The molten steel weight prediction unit is used for matching the corresponding tundish molten steel weight mechanism model in the model database according to the number of the current tundish, inputting the current tundish molten steel level data into the corresponding tundish molten steel weight mechanism model, predicting the molten steel weight to obtain the current tundish molten steel weight prediction result, and sending the current tundish molten steel level data, the number thereof and the corresponding molten steel weight prediction result to the Web server.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311837118.5A CN117910620A (en) | 2023-12-28 | 2023-12-28 | Tundish molten steel weight prediction method and system thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311837118.5A CN117910620A (en) | 2023-12-28 | 2023-12-28 | Tundish molten steel weight prediction method and system thereof |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117910620A true CN117910620A (en) | 2024-04-19 |
Family
ID=90696057
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311837118.5A Pending CN117910620A (en) | 2023-12-28 | 2023-12-28 | Tundish molten steel weight prediction method and system thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117910620A (en) |
-
2023
- 2023-12-28 CN CN202311837118.5A patent/CN117910620A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111353656B (en) | Steel enterprise oxygen load prediction method based on production plan | |
| CN108256260A (en) | A kind of continuous casting billet quality Forecasting Methodology based on extreme learning machine | |
| CN110066895B (en) | Stacking-based blast furnace molten iron quality interval prediction method | |
| JPH0315902A (en) | Method and system for supporting process operation | |
| CN103942422B (en) | Granular-computation-based long-term prediction method for converter gas holder positions in metallurgy industry | |
| CN112633584B (en) | River sudden water pollution accident water quality prediction method based on improved LSTM-seq2seq model | |
| CN111258984B (en) | Product quality end-edge-cloud collaborative forecasting method under industrial big data environment | |
| CN106845826B (en) | PCA-Cpk-based cold continuous rolling production line service quality state evaluation method | |
| CN108460481A (en) | Unmanned plane spot development law prediction technique based on Recognition with Recurrent Neural Network | |
| CN103971133A (en) | Automatic steel plate surface defect recognition method based on case-based reasoning | |
| CN108265157A (en) | Intelligent arc furnace steelmaking system | |
| CN101285816A (en) | Copper matte air refining procedure parameter soft sensing instrument and its soft sensing method | |
| CN113807596B (en) | Management method and system for informatization project cost | |
| CN117910620A (en) | Tundish molten steel weight prediction method and system thereof | |
| CN109648057A (en) | Packet level stability control method and system in a kind of continuous casting | |
| CN117689256A (en) | Aluminum alloy casting product quality tracing method | |
| CN114581859B (en) | Converter slag discharging monitoring method and system | |
| CN112835392A (en) | Pneumatic pressure automatic control system | |
| CN112541634A (en) | Top layer oil temperature prediction and false fire alarm discrimination method, device and storage medium | |
| CN114969068B (en) | Urban pressure pipe network real-time flow monitoring data analysis method and system | |
| CN112182854B (en) | Intelligent monitoring method and system for abnormal furnace conditions of blast furnace | |
| CN113059184B (en) | Optimization method for parameters of ingot blank spray forming process | |
| CN110174409B (en) | Medium plate periodic defect control method based on real-time detection result | |
| CN101093522A (en) | Computer aided scheduling model for guiding rhythm of steel making, and continuous casting steel supply | |
| CN100371938C (en) | Quality design method under minute new aluminium sample |
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 |