Algorithms are the core of the Siemens process model that optimizes hot-strip cooling.

A decisive criterion for the quality of hot-rolled strip is that it must lie within the steel customer’s tolerance range for tensile and yield strength. To document this quality, the rolling mill operator traditionally has taken samples taken from ongoing production and subjected them to an extensive – and expensive – series of tests.

To overcome these shortcomings, a new Microstructure Monitor from Siemens Metals Technologies (www.siemens. com/metals) reportedly determines these quality parameters online, during production, reducing the need for laboratory measurements, and the time needed to carry out those tests.

Additionally, the Microstructure Monitor enables process parameters, such as coiling temperature, to be optimized with regard to target mechanical values. Siemens says the model has been successful in applications at ArcelorMittal, ThyssenKrupp Steel, and Salzgitter AG, all in Germany, at Rautaruukki in Finland, and SSAB in Sweden.

At a glance, Siemens lists these benefits:

• Cost savings through reduced sampling
• Cost savings through faster release of strip for further processing
• Improved strip quality through smaller spread of quality parameters
• Determination of technological windows
• Continuous quality control of the entire strip without an increase in costs
• Documentation of quality
• Support in the diagnosis of irregularities

Recently, a hybrid model for determining the mechanical properties of steel strip was developed to combine the technology of a physical-metallurgical process model with the learning ability and accuracy of an artificial neural network.

Supported by process parameters, including temperature, thickness reduction, strip velocity, process time, and chemical composition, the hybrid model determines the mechanical properties of hot-rolled steel. Meanwhile, the entry parameters are checked for plausibility and the microstructure is calculated.

The algorithms take into account strain hardening, static and dynamic recrystallization, and grain size, as well as the γ-α transformation in the cooling section. The model development takes place successively for different steel grades. As soon as the Microstructure Monitor has been adapted to a steel grade, such as low-CMn steels, Nb- or Ti-micro-alloyed steels, measurements can be saved and strip release can be accelerated.

Parallel to the physical model, a neural network refines the results (mainly the mechanical properties of the steel) and makes a significant contribution to greater precision.

A process protocol containing informative results is made available along with the coiled strip.

Not only does this reduce the need for laboratory tests and accelerate the release of the strip, it also ensures the accurate control of quality over the entire strip length. The Microstructure Monitor also acts as a diagnostics tool; regular deviations between forecast and sampled strain indicate irregularities in the mill train.

Besides the passive forecasting of mechanical properties, active intervention in the rolling and cooling process helps to ensure conformity to customer quality requirements. Direct optimization, for example, can reduce the need for expensive alloys, such as Nb and Mn.

A further possibility involves setting the coiling temperature as a function of desired steel properties. After the strip geometry has been established through rolling, the Microstructure Monitor uses data from the rolling process and the chemical composition of the steel to determine the actual properties of the strip.

Achieva’s signal processing algorithms will allow Steel Dynamics to estimate coating thickness during order transitions within one coating gauge scan.

The optimizer then employs this information to decide what should happen in the cooling section, in order to achieve mechanical targets for yield and tensile strength.

The system can yield signification cost savings. If, for instance, the system eliminates the need for 500 samples per month, Siemens officials calculate this translates into an annual saving of US$ 300,000.

The online capability of the system supports faster release of strips through the replacement of time-consuming sampling and analysis. Offline simulations, on the other hand, yield deeper knowledge of the dependencies in the microstructure and the mechanical characteristics of process parameters, which ultimately lead to a better quality product. In contrast to physical sampling, which is confined to a few locations on the strip, the Microstructure Monitor delivers results non-invasively over the entire strip length, laying the foundation for comprehensive documentation of strip quality.

Steel Dynamics Inc. placed an order last month with Industrial Automation Services (www.indauto.com) to design, supply, and commission three zinc-coating mass control systems. Two systems will be delivered to the SDI’s Butler plant and the third to the Jeffersonville plant, both in Indiana. According to IAS, engineering for the three systems will be executed in parallel with commissioning of the Butler systems occurring first.

The zinc-coating control systems are based on the IAS Achieva Coating Control system, which uses a physics-based coating-mass model to generate accurate references and control the knife to strip distances and stripping gas pressure.

Achieva’s signal processing algorithms will enable the controls to estimate coating thickness during order transitions within one coating gauge scan. Resulting control actions minimize the generation of out-of-spec material, and the minimized coating standard deviation reportedly allows for zinc savings. Typical returns on investment are gained within a few months, the developer says.

The open architecture control system will incorporate IAS’s easy-to-maintain graphical programming interface, and take advantage of IAS’s process know-how and local support. The new control system will operate in concert with existing Sentek coating gauges to provide improved strip coating performance.

IAS will execute engineering design, project management, installation assistance, commissioning and operator/maintenance training from its offices in Pittsburgh, and Newcastle, NSW, Australia.

In operation Achieva will control the coating thickness of the hot-dip coating lines by monitoring the feedback from coating thickness gauges and regulating jet pressure, horizontal jet-to-strip distance, and knife height. The kernel of the system is the adaptive setup and control system gain calculation.

The major functions of the system are:

• Jet stripping rig setup: This module is model-based and uses adaptive learning algorithms to determine rig setup references and control system gains.
• Dynamic coating mass control: Information from the coating mass gauge is used to control the stripping rig actuators (position, pressure, etc). The control system uses high-speed sampling and algorithms use time-delay compensated feedback, feed-forward, and coating-mass target optimization.
• Operator Interface: The interface supports displays for: line setup and trims, mimic, axonometric plots, controls selection, controls status, controls diagnostics, and system status.
• Diagnostic functions: These cover the system and controls and include a system status display. A second PC is located in the computer room for system development, data logging, and remote access and may also be used as a spare.