Continuous casting is one of the most important processes in steelmaking. There are many factors affecting the breakout accidents in continuous casting process, among which the sticker breakout is the most common, accounting for about 70% of the total breakout. In order to better solve the sticker breakout accident, clarify the cause and provide a theoretical basis for establishing a more comprehensive and accurate prediction model, the occurrence mechanism, influencing factors and prediction model of sticker breakout in continuous casting were summarized and analyzed. The influence of mold power performance, molten steel conditions, crystallizer parameters, personnel operation and other factors on the sticker breakout accident law was mainly discussed. For the current commonly used logical judgment breakout prediction model and machine learning-based breakout prediction model, the current research status was analyzed, and the current shortcomings were put forward. It is predicted that the breakout prediction system will continue to develop in the direction of intelligence and high-end in the future, which provides a reference for the research of sticker breakout and how to effectively prevent it.

In order to solve the problem of internal cracking of vanadium-containing steel in the continuous casting process, the solidification process analysis, organization prediction and surface temperature control of HRB500E rebar steel were carried out by using thermodynamic calculations, solidification organization simulation and the method of second-cooling water distribution control. It was found that the thermodynamic precipitation temperature of V(C,N) in the rebar steel was 1118 ℃ and the maximum precipitation amount was 0.06 wt%, which indicated that vanadium had the dual raction of precipitation strengthening and solid solution strengthening; the organization analysis of the as-cast billet showed that when vanadium was increased from 0 wt% to 0.089 wt%, the grain radius was reduced by 2 times, the number of grains was increased by 2.8 times, and the central equiaxial crystal was increased by 2.2%. It shows that the addition of vanadium can refine the grain, improve the organizational properties and gain metallurgical properties; at the same time, it is found that the reason for cracks and other defects is mainly due to the unreasonable water distribution of the second cold, resulting in the surface of the billet return temperature is too large. Based on this, a corresponding optimization of secondary cooling water distribution process has been proposed, along with the introduction of a three-section program for 2.9 m length of secondary cooling process, and ultimately designed a 1.4-2.7 m/min pulling rate of the water table, the surface of the billet return temperature was below 100 ℃ and the return temperature met the metallurgical guidelines. Industrial production has proved that the internal quality of billet has been greatly improved by the optimization of water distribution in the secondary cooling.

Taking 42CrMo4 round billet with a section size of ϕ800 mm as the research object, a large round billet model was established with ProCAST finite element method and CAFE method. The effects of superheat, drawing speed and specific water volume in secondary cooling zone on solidification behavior and grain structure were compared. The results show that the influence of the drawing speed on the solidification temperature, solidification end point and the thickness of the solidified billet is greater, while the influence of superheat on the grain growth is greater. When the drawing speed is increased by 0.02 m/min, the solidification end point moves back about 3 m, and the change of superheat and specific water volume in the secondary cooling zone has no obvious effect on the temperature and central solid fraction. When the superheat is increased from 10 ℃ to 55 ℃, the equiaxial crystal rate decreases from 43.2% to 9.2%, while the influence of pulling speed and secondary cooling zone specific water volume on the equiaxial crystal rate is not obvious, and the change is less than 8%. Based on the electromagnetic stirring position of the end, an optimization scheme was proposed for the existing process. The superheat was reduced to 10 ℃, the specific water volume was reduced to 0.16 L/kg, which could effectively improve the equiaxed crystal rate and refine the grains.

The technology of feeding steel strip to mold in continuous casting process is an effective method to control the temperature field distribution of molten steel and restrain or reduce the internal defects such as central segregation and central porosity. However, the continuous casting process is invisible at high temperature, so it is impossible to observe and measure the evolution of solidification structure in the mold online. A visual solidification experiment system was set up to investigate the effects of no cold strip, fixed cold strip and vibrational strip feeding on the microstructure evolution of NH₄Cl-70% H₂O solution during solidification. The results show that: firstly, feeding cold strip can significantly reduce the superheat in the center of mold, fuse dendrites, effectively increase the number of free grains, and then increase the volume ratio of equiaxed grains in solidification structure. Secondly, the superheat in the mold center decreases more rapidly after the cold strip is vibrated, and the nuclei formed by fusing are more and finer, and the proportion of equiaxed crystal volume increases further. The proportion of equiaxed crystal volume under three conditions of no cold strip, fixed cold strip and vibrational feeding strip is 26%, 31% and 36%, respectively.

During the continuous casting process of steel, the microsegregation and macrosegregation will caused by the uneven distribution of solute elements. It should be noted that the macrosegregation in the billet cannot be removed by the subsequent heat treatment or rolling process, and it affects the product properties. Therefore, it is of great significance to simulate the macrosegregation by considering the microsegregation behavior, which aims at enchancing the simulation accuracy of macrosegregation, and improving the quality of billet and the performance of products. In this work, a macrosegregation model that couples microsegregation behavior in order to compare macrosegregation behavior with different microsegregation models was established. The best solute microsegregation will be chosen, and accurate prediction of solute macrosegregation behavior in continuous casting billet will be realized. Besides, the Won-Thomas microsegregation model was used to investigate the effects of secondary dendrite arm spacing on the solidification end point and macrosegregation defects for 82B cord steel billets, when the spacing of the secondary dendrite arms is reduced from 200 mm to 40 mm, the solidification end point is shifted forward by 1.4 m, and the carbon mass fraction in the center of the continuous casting billet decreases from 0.154% to 0.138%, indicating that the larger the spacing of the secondary dendrite arms, the slower the cooling and the more serious the segregation, and the rationality and reliability of the model were verified through a carbon sulfur analyzer.

In order to improve the internal quality of 42CrMoA steel in a factory and reduce the casting of billet center segregation, the ProCAST software was used to simulate the casting of billet solidification structure, and the effect of continuous casting parameters on the distribution of solidification organization, the rate of equiaxial crystal and equiaxial crystalline area density was studied. When the casting speed was increased from 0.8 to 1.1 m/min, the proportion of equiaxial crystal zone has increased by 5.26%; When the cooling water flows increases from 0.144 to 0.192 L/kg, the proportion of equiaxial crystal zone has decreases by 3.34%; The superheat increases from 5 to 35 ℃, and the proportion of isometric crystal zone decreases by 7.98%. It is concluded that it is possible to reduce the macroscopic segregation of 42CrMoA steel by appropriately increasing the casting speed, decreasing the amount of water in the second cooling ratio and reducing the superheat of the steel.

In order to study the main factors affecting electromagnetic field distribution of FTSC thin slab funnel mold, by measuring the electromagnetic field pattern distribution of funnel mold, a mathematical model of electromagnetic field inverse problem of funnel mold electromagnetic braking was constructed, which is based on the electromagnetic theory, and then the electromagnetic braking technology was studied in detail, and the effect of cooling water medium in copper plate water channels on the magnetic induction intensity of the funnel mold was investigated for the first time. The results show that the five pairs of magnetic poles form a geometric magnetic field distribution, which similar to the shape of "ω" on the wide surface of the mold. When the magnetic pole current intensity increases from 300-200 A to 1 100-800 A, the maximum magnetic induction intensity in the mold increases from approximately 70 mT to 200 mT; under the same coil current intensity, the larger the coil turns, the stronger the magnetic field generated; the cooling water medium in the copper plate water channels, which has little influence on the electromagnetic field in the mold.

To improve the solidification microstructure of CuNiCoSi rectangular continuous casting billets, a CuNiCoSi rectangular billet cross-section solidification microstructure model was established using the CA-FE method. The effects of different superheats and pulling speeds on the solidification microstructure were investigated using this model. The results indicate that reducing superheat can increase the proportion of equiaxed crystal zone in the solidification microstructure and improve the uniformity of the equiaxed crystal zone. Keeping other conditions constant, decreasing superheat from 30 ℃ to 15 ℃ resulted in an increase in the equiaxed crystal zone area from 42.90% to 47.32%, while the maximum equiaxed grain area decreased from 42.06 mm² to 33.94 mm². Increasing the pulling speed can increase the proportion of equiaxed grains in the solidification structure and obtain finer equiaxed grains. Increasing the pulling speed from 100 mm/min to 400 mm/min resulted in an increase in the equiaxed crystal zone area from 42.48% to 50.14%, while the average equiaxed crystal area decreased from 10.59 mm² to 9.38 mm². After optimizing the superheat and pulling speed, the quality of the billet solidification microstructure has significantly improved.

Slag entrapment in the continuous casting bloom has a significant effect on the sub-skin temperature, shell thickness, first principal stress and equivalent stress. Based on the temperature inheritance algorithm, a two-dimensional transient thermal conductivity model of steel solidification was established by using ANSYS finite element software to analyze the behavior of slag entrapment on bloom heat transfer and stress. In common 1/4 position from the center of the continuous casting bloom, the 10 mm wide and 4 mm deep protective slag is usually found. The study shows that with the slag entrapment, the sub-skin temperature rises by 60 ℃ in average, the shell thickness becomes 3.3 mm, the first principal stress distribution is uneven and locally smaller, and the equivalent stress becomes smaller.

Aiming at the problem of whether there are certain key characteristic shape subsequences in high-frequency time series in industrial control processes and the specific location of this subsequence in the time series, identification and positioning algorithm for key shape characteristic subsequences in industrial time series is proposed based on shapelet. Shapelet are the most discriminative continuous subsequences in time series. The shapelet set can be applied to the similarity calculation of subsequences of different lengths, and the sequence identification results are interpretable. In order to improve the speed and accuracy of identifying and positioning key shape characteristic subsequences in time series, shapelet sets with specific shape are first extracted and screened from the time series data set based on genetic algorithms. Secondly, the method of data standardization and sliding Euclidean distance is used to calculate the similarity measurement value between the shapelet and the subsequence in the time series, which is used to evaluate the similarity of shape characteristic. Then, the concepts of adaptive similarity threshold and lag time are defined to achieve accurate identification and positioning of characteristic shape subsequences existing in time series and improve the recognition accuracy of key shape subsequences. Finally, the feasibility and accuracy of the method were verified using public standard data sets and time series data of casting speed during continuous casting process.

Mold flux is an important functional material in continuous casting. In order to accurately, quickly, and low-costly obtain the physical and chemical properties of mold flux, a model was established for predicting the physical and chemical properties of the mold flux (composition, melting point, melting rate, and viscosity data) using BP neural network combined with particle swarm optimization (PSO) algorithm based on the testing data from laboratory. Thirteen untrained test samples were selected to test the prediction accuracy of the PSO-BP model. The results showed that compared with the BP neural network prediction model, the average absolute errors of melting point, melting rate, and viscosity were reduced from 8.9 ℃, 4.7 s, and 0.012 Pa·s to 8.1 ℃, 2.8 s, and 0.010 Pa·s, respectively. Moreover, the error fluctuations of individual samples were reduced, and the overall prediction accuracy was improved. Based on this model, the influence of single or multiple changes in the composition of mold flux on the physical and chemical properties was studied. By controlling other components to remain unchanged, when the basicity increased from 0.8 to 1.2, the viscosity value decreased from 0.23 Pa·s to 0.18 Pa·s. In addition, the effects of single variable adjustment and simultaneous variation of Al₂O₃ and MgO on the viscosity performance of mold flux were demonstrated. The model calculation results were consistent with actual theoretical laws, indicating that the predictive model of mold flux based on PSO-BP neural network can be applied to the development and research of mold flux, shorten the research cycle, and reduce costs.

Based on the full process control of steelmaking-hot rolling-cold rolling (heat treatment) process, a digital platform for full process quality control was built using modern data communication and database (data cloud), and an online quality rating system for continuous casting slab was developed based on this platform. This digital platform for quality control throughout the entire process monitors and tracks the parameters of raw materials, production, processes, equipment, and other aspects in the steelmaking process in real time. The tracking results are fed back to each furnace and slab in the form of process events. By formulating corresponding rules, continuous casting slabs are graded and evaluated to achieve online real-time prediction and evaluation of continuous casting slab quality, in order to determine the grading and repair treatment method of continuous casting slabs. The quality inspection and judgment results of the subsequent rolling process show that the quality rating system can classify and screen continuous casting slabs with different quality risk levels more intuitively and accurately, reducing the proportion of slab defects and improving the quality of continuous casting slabs.