Electric steelworks are a paradigmatic representation of the concept of circular economy, as it recycles steel components at the end-of-life products. Moreover, its importance is foreseen to grow according to the increasing demand of decarbonizing steel production to meet the ambitious goals of the European Green Deal. The electric arc furnace-based route is still characterized by a limited diversification of energy supply sources thus, managing the three factors of smart energy management, energy prices, and production planning can be jointly considered as a crucial leverage for reducing production costs while ensuring satisfaction of energy demand coming from the different processes and developing digital approaches and tools to implement the fast adaptation to power grid behaviour. The article describes a novel energy management system based on innovative components and a flexible infrastructure, which uses machine learning and an optimization approach to minimize electricity consumption and level trends by matching intelligent production planning and power grid offer and related energy costs. The developed solution and the set of neural networks-based models estimating electricity consumption in electric arc furnace and ladle furnace based on production information are described. The models were trained and validated using production and process data from a real steelworks.

Although electric arc furnace (EAF)-based steelworks produce steel from recycled ferrous scrap and inherently implement the concept of circularity, they are challenged to reduce their overall environmental impact, reduce CO2 emissions, and maximize energy and resource efficiency. The paper exemplary shows how advanced digital technologies, including artificial intelligence-based techniques, can support decarbonization and sustainability improvement of electric steelmaking. The paper presents computationally efficient machine learning models estimating sterile content in different types of scrap reaching the scrap yard as well as steel chemical composition and temperature at the exit of the Ladle furnace. The models are designed to be included in an innovative software platform based on federated learning (FL) helping industrial staff in decision-making by estimating energy consumption and other parameters affecting the environmental impact according to the material mix fed to the EAF. The paper describes the rationale behind models' design, the approach for selecting their hyperparameters, and the results achieved on data gathered from two different steelworks, the first one exploited as reference for models' first setup, the second one used to assess models' usability in the considered FL context. The performances are satisfactory in both cases, and key issues for implementation and further improvement are discussed.

In the context of electric steelworks, energy management is a key factor to reduce production costs while ensuring satisfaction of energy demands of all the different processes. Energy consumption optimization in the steel production chain can only be achieved by jointly considering individual processes as a network of users, in which each process is already close to the optimal operating point. This results in a large number of energy consumers to be managed, and to this aim a scheduler is usually adopted. The European project EnerMIND aims at effective and efficient steel production and a high utilization rate of production facilities through the implementation of a software demonstrator of a new energy management system based on a new energy management model exploiting a flexible infrastructure. This demonstrator, which covers the whole production chain, considers the areas of the steelworks showing the highest energy demands, i.e. Electric Arc Furnace, secondary steelmaking and the reheating furnaces that feed the rolling mills, in terms of total and peak values.

Reliable prediction models for smart agricultural system needs methods that stay stable, easy to understand, and can be reproduced across different sites, seasons, and cultivars. FIELD is a compact procedure for selecting a regression model that highlights stability and practicality, not only raw accuracy. The method combines six different performance metrics for describing the effectiveness of each model. It measures the average absolute error on an external test set with truly unseen data. It checks how concentrated or spread out those test errors are to guarantee consistency across the modelling case. It records the total time needed to fit the model and make predictions, so efficient models are valued. It looks at cross validation and considers how much the losses change from split to split and what the average normalized error across folds is. It also records how spread out that cross validation error is. After cleaning and standardizing the data, FIELD trains five families of regressors, namely a linear model, a quadratic polynomial, a decision tree, a Gaussian process, and a small feed forward neural network. Each performance metric is mapped to a common normalized scale and is then combined with a Mamdani fuzzy inference system whose rule base is short and easy to audit, using simple if-then statements. Final scores are ranked to select the best one, and if two models tie the more interpretable one is preferred. The proposed procedure increases the stability of model selection without long and complex hyperparameter searches. It is practical for agronomic datasets of moderate size, with mixed signal to noise ratios and with seasonal shifts, because its robustness to changes in site, season, or cultivar is an explicit goal and increases the model transparency and explainability.

Water is a fundamental steelworks additional resource; its efficient management is crucial for process reliability, product quality and environmental sustainability. Within steelworks, water is exploited mainly for direct or indirect cooling and is usually reused and recycled after cooling and treatments to eliminate contaminants. However, bottlenecks often exist, limiting water management efficiency and increasing water consumption. These issues are mainly related to water treatments efficiency, lack of water parameters monitoring and the manual/semi-manual management of water networks. Furthermore, these aspects are generally associated with the plant’s service life; brownfield sites are mostly affected. In these cases, improving sensor circuits coupled with decision support tools can support human decisions and lead to significant advantages. The paper discusses a potential application of such tools after new sensors installation in a use case concerning the minimization of the use of high-quality make-up-water for the indirect cooling system of a wire-rod mill in electric steelworks. The effectiveness of the described tool is shown, and the advantages are highlighted in terms of potential savings that can reach 95% and 4% of the current consumption of well and osmotic water in the considered circuit, respectively, corresponding to a saving of about 9400 m3/year of high-quality water.

Ovarian cancer (OC) is the most lethal of all gynecological cancers. Due to vague symptoms, OC is mostly detected at advanced stages, with a 5-year survival rate (SR) of only 30%; diagnosis at stage I increases the 5-year SR to 90%, suggesting that early diagnosis is essential to cure OC. Currently, the clinical need for an early, reliable diagnostic test for OC screening remains unmet; indeed, screening is not even recommended for healthy women with no familial history of OC for fear of post-screening adverse events. Salivary diagnostics is considered a major resource for diagnostics of the future. In this work, we searched for OC biomarkers (BMs) by comparing saliva samples of patients with various stages of OC, breast cancer (BC) patients, and healthy subjects using an unbiased, high-throughput proteomics approach. We analyzed the results using both logistic regression (LR) and machine learning (ML) for pattern analysis and variable selection to highlight molecular signatures for OC and BC diagnosis and possibly re-classification. Here, we show that saliva is an informative test fluid for an unbiased proteomic search of candidate BMs for identifying OC patients. Although we were not able to fully exploit the potential of ML methods due to the small sample size of our study, LR and ML provided patterns of candidate BMs that are now available for further validation analysis in the relevant population and for biochemical identification.

Within the design of a machine learning-based solution for classification or regression problems, variable selection techniques are often applied to identify the input variables, which mainly affect the considered target. The selection of such variables provides very interesting advantages, such as lower complexity of the model and of the learning algorithm, reduction of computational time and improvement of performances. Moreover, variable selection is useful to gain a profound knowledge of the considered problem. High correlation in variables often produces multiple subsets of equally optimal variables, which makes the traditional method of variable selection unstable, leading to instability and reducing the confidence of selected variables. Stability identifies the reproducibility power of the variable selection method. Therefore, having a high stability is as important as the high precision of the developed model. The paper presents an automatic procedure for variable selection in classification (binary and multi-class) and regression tasks, which provides an optimal stability index without requiring any a priori information on data. The proposed approach has been tested on different small datasets, which are unstable by nature, and has achieved satisfactory results.

Data pre-processing plays a significant role in data mining in ensuring good quality data especially when dealing with industrial datasets. Dimensionality reduction and outliers detection are significant tasks in machine learning and data mining applications. Many industrial real-world datasets are affected by several errors and present a huge number of process variables. This circumstance can cause problems when the data are exploited to develop efficient machine learning-based models. In this paper some automatic pre-processing procedures are proposed considering both detection of anomalous data and informative variables selection to be used before models training in order to improve their performance. The paper proposes three industrial case studies coming from the context of the steelmaking industry and all the obtained results prove the efficiency of a correct and appropriate pre-processing stage.

Variable selection is an essential tool for gaining knowledge on a problem or phenomenon, by identifying the factors that shows the highest influence on it. It is also fundamental for the implementation of machine learning-based approaches to modelling and classification tasks, by improving performances and reducing computational cost. Furthermore, in many real-world applications, such as the ones in the medical field, a relevant number of variables are jointly observed, but the number of available observations is quite limited. In these cases, variable selection is clearly essential, but standard variable selection approaches become "unstable", as the high correlation among different variables or their similar relevance with respect to the considered target lead to multiple solutions leading to similar performances. In machine-learning based classification, the stability of variable selection, namely its robustness with respect variations in the classifier training dataset, is as important as the performance of the classifier itself. The paper presents an automatic procedure for variable selection in classification tasks, which ensures excellent stability of the selection and does not require any a priori information on the available data.

This paper faces the problemVannucci, Marco of variables selection through the use of a genetic algorithm based metaheuristic approach.Colla, Valentina The method is based on the evolution of a population of variables subsets, which is led by the genetic operators determining their selection and improvement through the algorithm generations. The impact of different genetic operators expressly designed for this purpose is assessed through a test campaign.Cateni, Silvia The results show that the use of specific operators can lead to remarkable improvements in terms of selection quality.

Variable selection is an important task in machine learning and data mining applications. In many real world problems a huge volume of data is often available which corresponds to a large number of variables. When developing a model for classification, clustering or other applications, the search of an optimal subset of relevant input variables is crucial. In this paper an automatic variable selection method, which combines Genetic Algorithms and Self Organizing Maps, is proposed for classification purposes. The Genetic Algorithm is used to select the most relevant input variables and to set some relevant parameters of a classifier implemented through a Self Organizing Map. This method has been tested with several datasets belonging to the UCI repository. The results of the tests are presented and discussed in this paper. The proposed approach provides a good classification accuracy and contributes to the comprehension of the phenomenon under consideration. Keywordsvariable selection, feature selection, genetic algorithms, Self Organizing Maps, classification.

The paper presents a machine learning-based system aimed at improving the homogeneity of tensile properties of steel strips for automotive applications over their strip length in the annealing and hot dip galvanizing lines. A novel modular approach is proposed exploiting process and product data and combining smart data pre-processing and cleansing algorithms, an ensemble of neural networks targeted to specific product classes and an ad-hoc developed iterative procedure for identifying the variability ranges of the most relevant process variables. A decision support concept is implemented through a software tool, which facilitates exploitation by plant managers and operators. The system has been tested on site. The results show its effectiveness in improving the control of the thermal evolution of the strip with respect to the standard operating practice.

In the last years, the artificial neural networks have been effectively applied to several industrial problems in order to improve knowledge and get a deeper insight into correlations among different factors which affect production processes. In many applications Neural Networks are applied to predict the relationship between available input process variables and the target to be forecasted. Data pre-processing is an important step in developing a neural network application, which could affect the accuracy and the results of the developed models and applications. In the present paper an approach is proposed for data pre-processing, concerning a particular application related to the steel production. Such approach is tested on one row dataset coming from a real industrial context and the obtained results demonstrate the effectiveness of an accurate and appropriate pre-processing phase.

Echo-State Neural Networks represent a very efficient solution for modelling of dynamic systems, thanks to their particular structure, which allows faithful reproduction of the behavior of the system to model with a usually limited computational burden for a training phase. This aspect favors the deployment of Echo-State Neural networks in the industrial field. In this paper, a novel application of such approach is proposed for the modelling of industrial processes. The developed models are part of a complex system for optimizing the exploitation of process off-gases in an integrated steelwork. Two models are presented and discussed, where both shallow Echo-State Neural Networks and Deep Echo State Neural networks are applied. The achieved results are presented and discussed, by comparing advantages and drawbacks of both approaches.

Variable reduction is an essential step in data mining, which is able effectively to increase both the performance of machine learning and the process knowledge by removing the redundant and irrelevant input variables. The paper presents a variable selection approach merging the dominating set procedure for redundancy analysis and a wrapper approach in order to achieve an informative and not redundant subset of variables improving both the stability and the computational complexity. The proposed approach is tested on different datasets coming from the UCI repository and from industrial contexts and is compared to the exhaustive variable selection approach, which is often considered optimal in terms of system performance. Moreover the novel method is applied to both classification and regression procedures.

In this chapter we present some results of the first European research project dealing with the utilisation of Big Data ideas and concepts in the Steel Industry. In the first part, it motivates the definition of a multi-scale data representation over multiple production stages. This data model is capable to synchronize high-resolution (HR) measuring data gathered along the whole flat steel production chain. In the second part, a realization of this concept as a three-tier software architecture including a web-service for a standardized data access is described and some implementation details are given. Finally, two industrial demonstration applications are presented in detail to explain the full potential of this concept and to prove that it is operationally applicable. In the first application, we realized an instant interactive data visualisation enabling the in-coil aggregation of millions of quality and process measures within seconds. In the second application, we used the simple and fast HR data access to realize a refined cause-and-effect analysis.

This paper presents an analysis of the occurrence of ripple defects during Hot Deep Galvanising of flat steel products, with a focus on the study on thick coils having low zinc coating. Although skilled personnel can manage ripples defects through particular operations, for instance wiping nitrogen instead of air in air blades, the real effects of each process parameter variation is unknown. Therefore, the study of these phenomena can improve the quality of coils, by decreasing reworked or scrapped material and reducing costs related to a redundant use of nitrogen. An accurate pre-processing procedure has been performed and then the analysis focused on the possible causes of ripples occurrences. In particular, the attention is focused on the development of a model capable to identify process variables with a stronger impact on the presence or absence of ripples, by expressing such effect through an appropriate relationship.