Increased Caster Productivity with the B.O.P.S. Breakout Prediction System

For new casting plants and the modernization of older ones, maximum productivity depends on reaching the maximum cast speed possible. And it’s here that a new problem arises. The shell of the strand, which is formed in the mold, or defective zones in the shell cannot heal out before the strand leaves the mold. A costly breakout then occurs, and production must be stopped for several hours, defective segments and cables must be changed, etc.

Typical causes of breakouts

Breakouts are typically caused by “stickers” and “cracks.”

Stickers are the result of high localized friction between the strand and the mold, due, for example, to a shortage of casting powder. At the point of increased friction, the strand sticks to the side of the mold more strongly than elsewhere, and its cooling speed at that point is reduced. This increases stress in the strand shell, which breaks open. Molten steel then gains access to the side of the mold and raises the temper-ature there.

Air cushions, or cracks, which form between the strand and the mold, are another cause of breakouts. The low thermal conductivity of the air greatly reduces the transmission of heat from the strand to the mold with the result that a very thin shell is formed on the strand. Together, stickers and cracks account for some 96 % of all breakouts. Both causes demonstrate typical temperature patterns, which are the basis of the breakout evaluation and are illustrated in Figure 1.

B.O.P.S. hardware

To meet the challenge of avoiding stickers and cracks, Siemens B.O.P.S. is implemented with the advanced SIMATIC S7 automation system, with visualization performed by SIMATIC WinCC running on an industrial PC. This means that breakout prediction can be easily integrated into the overall automation system of any type of continuous casting plant.

Operation and parameterization

Through the “Strand Online” and “Online Curves” display screens, the operator is kept informed about the actual temperatures within the mold and the calculated probability for a breakout (Fig. 2). In the event of an alarm, all temperature values of the sensors, their probability and casting parameters are stored under “Offline Curves” and “Alarm Report” for further evaluation by the production and maintenance team.

For a new plant, the screens are integrated into the overall visualization system, with at least one separate terminal provided for B.O.P.S. In existing plants, one terminal or a server/client system (depending on the desired working places) can be provided.

Siemens B.O.P.S. is extremely versatile. Thanks to a global database, which can be accessed over a special parameter screen (“Mold and Steel Grade Archives”), B.O.P.S. can be adapted to all types of slab caster and different steel grades. Parameterization is done during the commissioning and first optimization, but can easily be extended by the customer later on.

Temperature pattern recognition

The strand surface temperature is measured within the mold using up to three rows of the temperature sensors installed in the copper plate. Siemens B.O.P.S. relies on fuzzy logic within the internal feedback electronics for pattern recognition at each sensor. The feedback of breakout probability contains the compressed history of the measured trace. This method is faster and much more efficient than others that store and evaluate only the complete temperature pattern. Figure 3 depicts this procedure. Besides the safe detection of a breakout, false alarms are reduced to a minimum by taking into account the change of cast speed, mold level, ladle, and mold width. Moreover, defective thermoelements are automatically deselected so that B.O.P.S. can continue operation. These features promote the high availability and flexibility of the system, ensuring increased plant productivity.

B.O.P.S. at work

At Krupp Thyssen Nirosta (KTN) in Krefeld, Germany (modernization of a single-strand slab caster for stainless steel), the increase of casting speed required a breakout prediction system. Siemens B.O.P.S. was selected due to its flexibility and expansion capability. KTN provided the mold equipment itself, comprising three rows of 20 temperature sensors each. The third row of sensors is designed to a breakout when filling the mold and before starting the pinch rolls.

For visualization of the casting process, Siemens supplied a client/server system using three terminals, one at the caster control pulpit for the operator, a second at the electrical room for maintenance personnel, and the third for production personnel to perform further analysis. After an optimization period of approximately three months, which was mostly done by remote access via ISDN, the KTN breakout prediction system is running successfully. Besides avoiding some breakouts, other mechanical problems at the mold can now be detected by B.O.P.S.

At Saldanha Steel in South Africa (new plant involving single-strand thin slab caster with cast speed up to 5 m/min), the system was upgraded with advanced SIMATIC S7 control technology running WinCC and featuring fuzzy evaluation. Siemens also supplied the equipment on the mold consisting of two rows of thermoelements, each row containing 12 sensors. An online PC detects fast breakout, and a second offline PC supports further evaluation and data transfer to the Level 2 automation system.