Today, everyday objects are becoming more and more intelligent and some-times even have self-learning capabilities. These self-learning capacities in particular also act as catalysts for new developments in the steel industry.Technical developments that enhance the sustainability and productivity of steel production are very much in demand in the long-term. The methods of Industry 4.0 can support the steel production process in a way that enables steel to be produced in a more cost-effective and environmentally friendly manner. This thesis describes the development of an opto-acoustic system for the early detection of slag slopping in the BOF (Basic Oxygen Furnace) converter process. The prototype has been installed in Salzgitter Stahlwerks, a German steel plant for initial testing. It consists of an image monitoring camera at the converter mouth, a sound measurement system and an oscillation measurement device installed at the blowing lance. The camera signals are processed by a special image processing software. These signals are used to rate the amount of spilled slag and for a better interpretation of both the sound data and the oscillation data. A certain aspect of the opto-acoustic system for slopping detection is that all signals, i.e. optic, acoustic and vibratory, are affected by process-related parameters which are not always relevant for the slopping event. These uncertainties affect the prediction of the slopping phenomena and ultimately the reliability of the entire slopping system. Machine Learning algorithms have been been applied to predict the Slopping phenomenon based on the data from the sensors as well as the other process parameters. / Idag blir vardagliga föremål mer och mer intelligenta och ibland har de självlärande möjligheter. Dessa självlärande förmågor fungerar också som katalysatorer för den nya utvecklingen inom stålindustrin. Teknisk utveckling som stärker hållbarheten och produktiviteten i stålproduktionen är mycket efterfrågad på lång sikt. Metoderna för Industry 4.0 kan stödja stålproduktionsprocessen på ett sätt som gör att stål kan produceras på ett mer kostnadseffektivt och miljövänligt sätt. Denna avhandling beskriver utvecklingen av ett opto-akustiskt system för tidig detektering av slaggsslipning i konverteringsprocessen BOF (Basic Oxygen Furnace). Prototypen har installerats i Salzgitter Stahlwerks, en tysk stålverk för första provning. Den består av en bildövervakningskamera på omvandlarens mun, ett ljudmätningssystem och en oscillationsmätningsenhet som installeras vid blåsans. Kamerans signaler behandlas av en speciell bildbehandlingsprogram. Dessa signaler används för att bestämma mängden spilld slagg och för bättre tolkning av både ljuddata och oscillationsdata. En viss aspekt av det optoakustiska systemet för släckningsdetektering är att alla signaler, dvs optiska, akustiska och vibrerande, påverkas av processrelaterade parametrar som inte alltid är relevanta för slöjningsevenemanget. Dessa osäkerheter påverkar förutsägelsen av slopfenomenerna och i slutändan tillförlitligheten för hela slöjningssystemet. Maskininlärningsalgoritmer har tillämpats för att förutsäga Slopping-fenomenet baserat på data från sensorerna liksom de andra processparametrarna.

A escória de aciaria LD, principal coproduto da conversão do ferro gusa em aço, se utilizada como agregado na produção de concreto, pode reduzir custos com a estocagem de resíduos e reduzir a extração predatória de agregados naturais. Para tal, este estudo utilizou a escória de aciaria LD em substituição ao agregado miúdo do concreto, a areia, em teores de 0, 25, 50 e 75 %, em massa. O desempenho mecânico do concreto se mostrou próximo ao de referência, com exceção do módulo de elasticidade, que foi significativamente reduzido. A durabilidade dos concretos foi avaliada quanto à resistividade elétrica superficial, absorção de água, índice de vazios permeáveis, resistência à carbonatação acelerada, resistência à migração de cloretos, envelhecimento acelerado e natural. Quanto à durabilidade, apesar de verificada a redução da resistividade do concretos com escória, a substituição parcial da areia não promoveu o aumento da migração de cloretos, o que se atribui a fixação de clores pela formação de sal de Friedel; no entanto, a instabilidade volumétrica da escória se mostrou um fenômeno que gera graves patologias ao material, de modo que os corpos de prova com maior teor de substituição, após 70 dias de envelhecimento acelerado, romperam por expansão. / The basic oxygen furnace (BOF) slag, the main co-product of the process pig iron conversion to steel, if used as an aggregate in the concrete production, can reduce costs with the waste storage and reduce the predatory extraction of natural aggregates. For this, in this study the BOF slag was used in substitution to the small aggregate of the concrete, the sand, in contents of 0, 25, 50 and 75% by mass. The mechanical performance of the concrete was similar to that of reference, with the exception of the elasticity modulus, which was significantly reduced. The durability properties of the concretes were evaluated for surface electrical resistivity, water absorption, permeable voids index, resistance to accelerated carbonation, resistance to chloride migration, accelerated and natural aging. Regarding the durability, despite the reduction of the resistivity of slag concrete, the partial replacement of the sand did not promote the increase of chloride migration, which is attributed to the fixation of chlorides by Friedel Salt formation. However, the volumetric instability of the slag was a phenomenon that causes serious pathologies to the material, so that the specimens with higher substitution content, after 70 days of accelerated aging, ruptured by expansion.

The present thesis deals with the development of a new type of dynamic model for metallurgical reactors. It also covers some of the theoretical aspects of steelmaking that is necessary to include in such an application. The thesis consists of modeling work, high temperature experiments and cold model experiments. Two different aspects of slags in the oxygen steelmaking were investigated. In the first study, slag samples were equilibrated with copper at 1923K in order to study their capacities in capturing phosphorous. Some of the samples were liquid-solid mixtures. The solid phases in these samples were identified by SEM analysis. The identified phases were found to agree well with Thermocalc calculations while the amount of solid fractions didn’t. The phosphorous distribution between the different phases was examined. The phosphate capacities of the samples were evaluated. The MgO content didn’t show any appreciable impact on the phosphate capacity. Furthermore the activities of FeO in the liquid slag samples were calculated and were found to deviate positively from ideality. In the second study the foaming height of CaO-SiO2-FeO slags by the reaction with hot metal was investigated. It was found that the foaming height increased with increasing FeO content up to 20-25%. The foaming height was seen to decrease with increased viscosity. The present results indicated that simply using foaming index for converter slag might lead to wrong conclusion. Simulation experiments using cold model at room temperature were conducted. Cold model experiments were carried out in order to study the penetration depth due to an impinging gas jet on the surface of a liquid metal. The liquid alloy Ga-In-Sn was used to simulate steel. And an HCl solution was used to simulate the slag. A comparison with predictions of existing models was made and a new model parameter was suggested. The observation of the movement of metal droplets generated by the gas jet was also made. The low velocity of droplets suggested that the turbulent viscosity played important role and the droplets could have long resident time in the slag. Furthermore a study of the effect of gas flow rate on homogenization and inclusion removal in a gas stirred ladle was carried out. Both industrial trials and cold model experiments were conducted. As an auxiliary tool CFD was used to predict the mixing times and was found to agree well with both the model experiments and industrial data. The increase of flow rate of inert gas would not improve the mixing substantially at higher flow rates. The water model study showed also that the gas flow rate had negligible effect on the rate of inclusion removal. Both the experiments and CFD calculation strongly suggested that low gas flow rate should be applied in the ladle treatment. Lastly a new approach to a dynamic process model of 300 ton BOF converter was made. The main feature was to utilize the velocity vectors obtained by CFD simulation. In the standalone model, the steel melt domain was sliced into 1000 cells. Based on the imported velocity vectors from the CFD calculation, the mass transfer of carbon and phosphorus was calculated taking into account the slag metal reactions. The mass exchange between slag and metal was considered to be dominated by the metal droplet formation due to oxygen jet. The convergence of the model calculation and the promising comparison between the model prediction and the industrial data strongly suggested that the proposed approach would be a powerful tool in dynamic process control. However, more precise descriptions of other process aspects need to be included before the model can be practically employed in a dynamic controlling system. / <p>QC 20120829</p>

The aim of this thesis work is to increase the knowledge of phenomena taking place during the initial stage in a top blown converter. The work has been done in a few steps resulting in four different supplements. Water model experiments have been carried out using particle image velocimetry (PIV) technology. The system investigated was a fundamental top blown converter where an air jet was set to impinge on a water surface. The flow field of the combined blown case, where an air jet was introduced through a bottom nozzle, was also captured by the PIV. The work clearly showed that the flow field caused by an impinging top blown jet alone could not match that of the bottom blown case. The main re-circulation loop (or vortex) was investigated with respect to position and it was found that an increased flow rate pushes the center of the re-circulation loop downwards into the bath. However, for the top-blown case there is a point when the flow rate is too large to cause a distinguishable re-circulation loop since the jet becomes more plunging (i.e. penetrates deep into the bath) than impinging, with large surface agitation and splashing as a result.A numerical model with the same dimensions as the experimental system was then created. Three different turbulence models from the same family were tested: standard-, realizable- and a modified-(slight modification of one of the coefficients in order to produce less spreading of the air jet) k-ε turbulence model. It could be shown that for the family of k-ε turbulence models the difference in penetration depth was small and that the values corresponded well to literature data. However, when it comes to the position of the re-circulation loop it was shown that the realizable k-ε model produced better results when comparing the results to the experimental data produced from the PIV measurements, mentioned earlier.It was then shown how the computational fluid dynamics (CFD) model could be coupled to thermodynamics databases in order to solve for both reactions and transport in the system. Instead of an air-water system, a gas-steel-slag system was created using the knowledge obtained in the previous simulation step described above. Reactions between gas-steel, gas-slag, steel-slag and gas-steel-slag were considered. Extrapolation of data from a few seconds of simulation was used for comparison to experimental data from the literature and showed reasonable agreement. The overall conclusion was that it is possible to make a coupling of the Thermo-Calc databases and a CFD software to make dynamic simulations of metallurgical processes such as a top-blown converter.A parametric study was then undertaken where two different steel grades were tested; one with high initial carbon content (3.85 mass-%) and one with lower carbon content (0.5 mass-%). The initial silicon content was held constant at 0.84 mass-%. Different initial temperatures were tested and also some variation in initial dissolved oxygen content was tried. It was found that the rate of decarburization/desiliconization was influenced by the temperature and carbon concentration in the melt, where a high temperature as well as a high carbon concentration favors decarburization over desiliconization. It was also seen that the region affected by a lower concentration of alloys (or impurities) was quite small close to the axis where the impinging jet hits the bath. Add the oscillating nature of the cavity and it was realized that sampling from this region during an experiment might be quite difficult. / QC 20100720

The purpose of the current study is to increase the understanding of different steelmaking processes at the decarburization stages by use of mathematical modeling. More specifically, two De-Laval nozzles from a VOD (Vaccum Oxygen Decarburization) process, which is used for producing stainless steels with ultra-low carbon grades, was investigated for different vessel pressures. Moreover, the post combustion phenomena in a BOF or LD (Linz-Donawitz) process as well as an AOD (Argon Oxygen Decarburization) process were studied focusing on the decarburization stage. Two industrial VOD nozzles were numerically studied and compared at different temperatures and ambient pressures. Flow patterns of the oxygen jet under different ambient pressures were predicted and the flow information at different positions from the nozzle was analyzed. In addition, the effects of different ambient temperatures on the jet velocity and the dynamic pressure were compared. The predictions revealed that a little under-expansion is somewhat helpful to improve the dynamic pressure. The jet dynamic pressure and its width for the specific nozzle geometry were also studied. It was observed that a variation in the ambient pressure can influence the jet momentum and its width. In addition, a high ambient temperature was found to have a positive effect on the improvement of the jet dynamic pressure. Furthermore, it was found that a change in ambient pressure has a stronger effect on the jet force than a change in the ambient temperature. In addition, it was proved that the profiles of the dynamic pressure at a certain blowing distance fit well to Multi-Gaussian distribution. Post combustion in a BOF/LD and an AOD process during decarburization was also studied. Two mathematical models were created to show the post combustion phenomenon inside the converters, respectively. For the CFD modeling of the two processes, the realizable k-ɛ model, the species transport model and the discrete ordinate were adopted to calculate the turbulence, gas reaction and radiation present in the gas phase in the converter. For the BOF/LD modeling, a series of plant tests were conducted to collect data, which were used in the current model. These include the off-gas information, emissivity data, oxygen blowing parameters and the chemical composition of steel. After the simulation, the predicted flow pattern and detailed information of the gases taking part in the post combustion were compared to plant data. Specifically, the off-gas data from the plant was used for the model verification. The measured CO2 concentration was 15-20 wt% and the predicted value from the modeling was 16.7 wt%. For the AOD converter of interest in the current work, a fan is installed in the end of the AOD flue to help extract the off-gas from the converter. The influence of different fan gauge pressures as well as temperatures of the gas mixture, containing the generated CO and argon, on the post combustion in the whole AOD system was studied. It was indicated from the modeling results that the post combustion was only present in the flue for the present modeling conditions. Moreover, a critical fan gauge pressure (approx.. -550 Pa) was found which could yield a maximum post combustion in the flue gas. For both two models (BOF/LD and AOD), simulations indicated that a change of the converter temperature from 1500 to 1700 °C did not influence the post combustion reaction to a large degree. In addition, these two models can be regarded as the first step for a future more in-depth modeling work of the post combustion. / <p>QC 20130913</p>

This thesis presents a framework for automatic recognition of human actions in uncontrolled, realistic video data such as movies, internet and surveillance videos. In this thesis, the human action recognition problem is solved from the perspective of local spatio-temporal feature and bag-of-features representation. The bag-of-features model only contains statistics of unordered low-level primitives, and any information concerning temporal ordering and spatial structure is lost. To address this issue, we proposed a novel multiscale local part model on the purpose of maintaining both structure information and ordering of local events for action recognition. The method includes both a coarse primitive level root feature covering event-content statistics and higher resolution overlapping part features incorporating local structure and temporal relationships. To extract the local spatio-temporal features, we investigated a random sampling strategy for efficient action recognition. We also introduced the idea of using very high sampling density for efficient and accurate classification. We further explored the potential of the method with the joint optimization of two constraints: the classification accuracy and its efficiency. On the performance side, we proposed a new local descriptor, called GBH, based on spatial and temporal gradients. It significantly improved the performance of the pure spatial gradient-based HOG descriptor on action recognition while preserving high computational efficiency. We have also shown that the performance of the state-of-the-art MBH descriptor can be improved with a discontinuity-preserving optical flow algorithm. In addition, a new method based on histogram intersection kernel was introduced to combine multiple channels of different descriptors. This method has the advantages of improving recognition accuracy with multiple descriptors and speeding up the classification process. On the efficiency side, we applied PCA to reduce the feature dimension which resulted in fast bag-of-features matching. We also evaluated the FLANN method on real-time action recognition. We conducted extensive experiments on real-world videos from challenging public action datasets. We showed that our methods achieved the state-of-the-art with real-time computational potential, thus highlighting the effectiveness and efficiency of the proposed methods.

This thesis presents a framework for automatic recognition of human actions in uncontrolled, realistic video data such as movies, internet and surveillance videos. In this thesis, the human action recognition problem is solved from the perspective of local spatio-temporal feature and bag-of-features representation. The bag-of-features model only contains statistics of unordered low-level primitives, and any information concerning temporal ordering and spatial structure is lost. To address this issue, we proposed a novel multiscale local part model on the purpose of maintaining both structure information and ordering of local events for action recognition. The method includes both a coarse primitive level root feature covering event-content statistics and higher resolution overlapping part features incorporating local structure and temporal relationships. To extract the local spatio-temporal features, we investigated a random sampling strategy for efficient action recognition. We also introduced the idea of using very high sampling density for efficient and accurate classification. We further explored the potential of the method with the joint optimization of two constraints: the classification accuracy and its efficiency. On the performance side, we proposed a new local descriptor, called GBH, based on spatial and temporal gradients. It significantly improved the performance of the pure spatial gradient-based HOG descriptor on action recognition while preserving high computational efficiency. We have also shown that the performance of the state-of-the-art MBH descriptor can be improved with a discontinuity-preserving optical flow algorithm. In addition, a new method based on histogram intersection kernel was introduced to combine multiple channels of different descriptors. This method has the advantages of improving recognition accuracy with multiple descriptors and speeding up the classification process. On the efficiency side, we applied PCA to reduce the feature dimension which resulted in fast bag-of-features matching. We also evaluated the FLANN method on real-time action recognition. We conducted extensive experiments on real-world videos from challenging public action datasets. We showed that our methods achieved the state-of-the-art with real-time computational potential, thus highlighting the effectiveness and efficiency of the proposed methods.

The influence of ladle-slag additions on the BOF-process performance were investigated in plant trials. The aim of the study was to recycle ladle slag from secondary steelmaking to the LD-converter to save lime and improve the slag formation. More specifically, two plant trial campaigns covering in total 83 heats, whereof 47 with ladle-slag additions and 36 without ladle-slag additions, were performed.  Slag and steel sampling of the process were performed at tapping as well as during blowing at 15, 35, and 65% of the total blowing time. During the first campaign, ladle slag was added through the chute and lime reductions were made manually to correct for the ladle-slag addition.  In the second campaign, a development of the approach was made to suite a normal production practice. More specifically, the ladle slag was added through the weight-hopper system and implemented in the process-control system. In this way, the lime additions were reduced automatically by approximately 260 kg per heat. Moreover, the heat balance was compensated with a reduction in the iron-ore consumption. Additionally, the lance program was modified and the lance was lowered in the initial stages of the blow.  On the positive side, it was found that no demerits in the metallurgical performance of the process occur when ladle slag is recycled to the BOF-process. Furthermore, only slight affections on the slag composition were found, mainly with respect to the Al2O3 and FeO-content. In addition, the ladle slag was shown to melt during the initial stages of the blow. This contributed to an increased slag weight both during the blow and at tapping. However, a negative effect on the blowing time was experienced during the trials. Although, this effect was more pronounced during the first campaign and could be reduced with a controlled heat balance during the second campaign. / QC 20110503

<p>The slag metal emulsion may play a significant role in the global furnace decarburization kinetics in oxygen steelmaking. In recent years, the important interaction between droplet generation rate, droplet residence time in slag and droplet decarburization rate has become more evident in the literature. The decarburization kinetics of Fe-C-S droplets in CaO-SiO2-MgO-FeO slags were investigated for highercarbon droplets (approximately 4.2 % C). The effect of slag FeO, droplet mass and Sulphur content on decarburization rate were evaluated. The limit between external and internal nucleation of CO gas was investigated. A model was developed for prediction oftime to the onset of carbon boil.</p> / Master of Science (MSc)

Slopping in BOS-processes poses many problems, most significantly to work environmentand process effectiveness. Due to the current weaknesses in slopping prediction systems afoaming slag suppression system with immediate effect is needed to prevent slopping.This project aims primarily to be a proof of concept for pressurized oxygen gas blowing as amean of foam steady-state height suppression to prevent slopping and secondarily toappreciate the viability of this concept economically.Five nozzles were designed and used to blow pressurized air onto foam made of silicon oil, atdifferent air flows. It was then determined which nozzle was the most effective by comparingheight difference with airflow. The airflow was compared to a life scale scenario to determinethe real flow rate which was used to determine economic viability.Results show that a nozzle with a circular small hole is the most effective nozzle requiring20 [ln min-1] to reduce the foam height by 69.6%. The real flow to achieve this would be0.605 [m3 s-1], however due to the cold models limitations this is not the true value. Theslopping suppression technique shows promise as a concept both economically andpractically. / Överkok i BOF-processen skapar många problem, mest i arbetsmiljön och minskar så väl produktionens effektivitet som takt. Eftersom de kontrollsystem som finns att tillgå idag     har vissa begränsningar så behövs det ett system för att motverka skumtillväxt med direkt inverkan för att hindra överkokning.Det här projektet ämnar huvudsakligen till att bevisa att ett pålagt flöde av syrgas kan trycka ned det skummande slaggets höjd och på så sätt förhindra överkokning och sekundärt till att bedöma konceptets ekonomiska rimlighet. Fem  munstycken designades  och användes till  att  blåsa tryckluft, med olika flöden, på skum bestående av   silikonolja.   Effektiviteten   hos   munstyckena   utvärderades   genom  att jämföra höjdskillnaden mot det pålagda luftflödet. Luftflödet jämnfördes sedan mot ett scenario i industriell skala och det verkliga luftflödet kunde därefter beräknas. Med detta  som  bakgrund gjordes en ekonomisk analys. Resultat visar att munstycket med ett litet cirkulärt håll är mest effektivt då det krävdes ett flöde på 20 [ln min-1] för att reducera skumhöjden med 69,6%. Det verkliga flödet beräknades till 0,605 [m3 s-1], men eftersom den kalla modellen har vissa begränsningar så är detta värde inte  sant.  Den  här  tekniken  för   att  förhindra   överkokning  ser   lovande   utbåde  ur  ett   ekonomiskt men också ur ett praktiskt perspektiv.