Breakout is one of the major accidents that often arise in the continuous casting shops of steel slabs in Bokaro Steel Plant, Jharkhand, India. Breakouts cause huge capital loss, reduced productivity, and create safety hazards. The existing system is not capable of predicting breakout accurately, as it considers only one process parameter, i.e., thermocouple temperature. The system also generates false alarms. Several other process parameters must also be considered to predict breakout accurately. This work has considered multiple process parameters (casting speed, mold level, thermocouple temperature, and taper/mold) and developed a breakout prediction system (BOPS) for continuous casting of steel slabs. The BOPS is modeled using an artificial neural network with a backpropagation algorithm, which further has been validated by using the Keras format and TensorFlow-based machine learning platforms. This work used the Adam optimizer and binary cross-entropy loss function to predict the liquid breakout in the caster and avoid operator intervention. The experimental results show that the developed model has 100% accuracy for generating an alarm during the actual breakout and thus, completely reduces the false alarm. Apart from the simulation-based validation findings, the investigators have also carried out the field application-based validation test results. This validation further unveiled that this breakout prediction method has a detection ratio of 100%, the frequency of false alarms is 0.113%, and a prediction accuracy ratio of 100%, which was found to be more effective than the existing system used in continuous casting of steel slab. Hence, this methodology enhanced the productivity and quality of the steel slabs and reduced substantial capital loss during the continuous casting of steel slabs. As a result, the presented hybrid algorithm of artificial neural network with backpropagation in breakout prediction does seem to be a more viable, efficient, and cost-effective method, which could also be utilized in the more advanced automated steel-manufacturing plants.

Mold breakout is one of the significant problems in a continuous casting machine (caster). It represents one of the key areas within the steel production facilities of a steel plant. A breakout event on a caster will always cause safety hazards, high repair costs, loss of production, and shutdown of the caster for a short while. In this paper, a logic-judgment-based mold breakout prediction system has been developed for a continuous casting machine. This system developed new algorithms to detect the different sticker behaviors. With more algorithms running, each algorithm is more specialized in the other behaviors of stickers. This new logic-based breakout prediction system (BOPS) not only detects sticker breakouts but also detects breakouts that takes place due to variations in casting speed, mold level fluctuation, and taper/mold problems. This system also finds the exact location of the breakout in the mold and reduces the number of false alarms. The task of the system is to recognize a sticker and prevent a breakout. Moreover, the breakout prediction system uses an online thermal map of the mold for process visualization and assisting breakout prediction. This is done by alerting the operating staff or automatically reducing the cast speed according to the location of alarmed thermocouples, the type of steel, the tundish temperature, and the size of the cold slab width. By applying the proposed model in an actual steel plant, field application results show that it could timely detect all 13 breakouts with a detection ratio of 100%, and the frequency of false alarms was less than 0.056% times/heat. It has the additional advantage of not needing a lot of learning data, as most neural networks do. Thus, this new logical BOPS system should not only detect the sticker breakouts but also detect breakouts taking place due to variations in casting speed and mold level fluctuation.

In continuous casting of steel, liquid steel is continuously solidified into a finished product. Bokaro Steel Limited have slab caster and during continuous casting, one of the major operational irregularities encountered is breakout. A breakout is a hazard whereby the thin shell of the strand breaks, allowing the still-molten metal inside the strand to leak out from the mould, thus necessitating an expensive machine shutdown. Hence the breakout has a profound influence on the caster availability; affecting the productivity, cost of production, damage of machinery, safety hazards, etc. Breakouts occur due to the presence of a primary crack before solidification of liquid steel. Controlling the primary cracks lead to reduce chances of breakouts. Breakout prediction system of Bokaro Steel Limited uses only thermocouples temperature as an input of logic based algorithm controlled by strand PLC to predict breakout and generate alarm. This system sometimes generates lots of false alarm and sometimes even fails to generate an alarm before breakout. In this paper, two back-propagation neural networks are developed to predict the presence of primary crack that might lead to breakout of liquid steel. One is neural network1 train only with one input parameter (temperatures of thermocouples attached to the mould) and the other is neural network2 train with three input parameters (temperatures of thermocouples, casting speed and mould level). Both the neural network1 and network2 have only one output. Output of neural networks is either 0 for no alarm (no primary crack) or 1 for alarm (presence of primary crack) in breakout conditions. Comparison of the network1 with network2 showed that network2 has less output error, better performance and identifies primary crack more accurately. Performance of neural networks are done using Levenberg–Marquardt training algorithm, and also log sigmoid transfer function is used for both the hidden and output layer. The neural networks model training, validation and testing is done by simulating in Mat-Lab/Simulink.

— Continuous casting of steel is a process in which liquid steel is continuously solidified into semi-finished or finished product (slabs, blooms or billets). There are many problems associated with continuous casting shop which affect the casting process. A major problem is associated in continuous casting shop is breakout of molten steel. Breakout leads to temporary shutdown of caster, damage of machinery due to splash of molten steel, capital loss, safety hazards etc. In Bokaro steel plant a logical based breakout prediction system is used to predict the breakout. This system sometimes generates false alarm and sometimes even fail to generate an alarm before breakout. Also the logical model has lot of dependence on specific equipment, process and calibration. Neural network can be implemented for a better breakout prediction system. So, in this paper a back-propagation neural network model is developed for predicting the existence of primary cracks that might lead to a breakout. The network gets its input temperatures from thermocouples which are attached to the wide and narrow sides of the mould. The output of the neural network is either logic 1 (for presence of crack) or logic 0 (for no crack). Testing the network shows excellent result as evident from the confusion matrix and performance plot. The neural network model is validated by simulating in MatLab/Simulink. The developed network may be used effectively in predicting breakouts during continuous casting. Such effective prediction can go a long way in reducing production losses in steel manufacturing.