New approach for the determination of the blast furnace cohesive zone

The cohesive zone of the blast furnace acts as a gas distributor, which has a remarkable impact on the regularity of blast furnace operation and operating ratios such as productivity, hot metal quality and regularity, or reducing agent consumption. Many techniques have been tried in the past to determine the cohesive zone characteristics, but either they cannot be used on a routine basis, or they lack of accuracy, and are at least hardly verifiable. The objective of the project was then to propose a reliable methodology for the determination of the cohesive zone shape on the basis of the real time process data of the fumace. With this aim, three different tools have finally been built and/or evaluated in the course of the research, which can be classified with regards to their complexity and sophistication degree. In a first stage, ACERALIA concentrated on the construction of a Neural Network System (NNS) that will help to characterise the shape of the cohesive zone by making use of some relevant process parameters (temperature profiles, pressure measurements, characteristics of the operating point of the furnace,...). The idea was to classify the operating situations in different patterns based on the available process information, and to link these patterns with the ones corresponding to the different temperature profiles measured with the in-burden probes. By doing this, it was supposed that these in-burden probe temperature profiles were directly linked to the shape of the cohesive zone. Unfortunately, the low availability of the in-burden probe in the course of this project didn't enable to conclude this approach. Finally, ACERALIA developed a relatively simple model, which proposes an indirect estimation of the cohesive zone shape through the determination of the upper limit of the thermal reserve zone. This method makes use of the temperature information of the above and in-burden probes, for the calculation of a temperature gradient in the upper shaft, from which the upper limit of the thermal reserve zone is derived. This indirect estimation of the cohesive zone can be partly justified by the results of some quenched blast furnaces, which show that its general shape was similar to the one of the temperature isotherms corresponding to the thermal reserve zone (examples of Kukioka BF 4 and Hirohata BF 1). A graphic user interface, which shows the estimation of the position of the top of the thermal reserve zone and other relevant variables, has been developed. The tests performed confirmed that this operating guidance system offers the ACERALIA's operator a valuable description of the profile of the top of the thermal reserve zone, and of its evolutions with time. IRSID's objective within this project has been to develop a specific methodology able to produce, at a given frequency, typically each 24 hours, a synthetic and consistent picture of the cohesive zone characteristics, based on the available on-line information. With this aim, IRSID has constructed a specific model, called ZAP, for the estimation of the shape and position of the cohesive zone. The general principles of the models, issued from former experiences, are that the softening line shape should be linked to the CO efficiency profile measured at the above burden probes, while the C/(C+O) ratio profile at the throat will impact the cohesive zone thickness, and then enable to derive the melting line from the softening line. After the evaluation of a first version of the model, which has not been conclusive, this basic vision has been completed by the integration of the results of two existing models. The first one enables to evaluate the average position of the melting line from the blast conditions, while the second one provides ZAP with an estimation of the cohesive zone volume, fitted on the shaft pressure taps information. This last version of the ZAP model has been evaluated by computing off-line the evolution of SOLLAC Fos BF 1 cohesive zone during