Slawomir Golak

Application of neural network for the prediction of eco-efficiency

This paper presents the application of neural networks in the design process of new technologies taking into account factors such as their influence on the environment and the economic effects of their implementation. The use of neural networks allowed eco-efficiency assessment of technologies based on highly reduced number of descriptive design parameters, which are very difficult to collect at the conceptual design stage. The great diversity of technologies involved along with the small number of available examples made difficult to construct a neural model and demanded careful data preprocessing and network structure selection.

Homogenization of Electromagnetic Force Field During Casting of Functionally Graded Composites

The paper presents the analysis of the process of casting functionally graded composites under alternating electromagnetic field. The process utilizes the effect of electromagnetic buoyancy on nonconductive particles of the reinforcement in the conducting liquid (melted metal). The authors analyzed the possibilities of homogenizing the distribution of electromagnetic field in such a way as to obtain the desired direction of electromagnetic buoyancy and at the same time minimize the stirring of the molten metal, which makes it difficult to receive the optimal distribution of reinforcement in the casting. The suggested solution was to use the conductive elements of the mold to move the nonuniformity of the electromagnetic field outside the casting and a parabolic inductor to smooth the field distribution in the casting.

Fabrication of functionally graded composites using a homogenised low-frequency electromagnetic field

A characteristic feature of functionally graded materials is the continuous spatial change of their mechanical, thermal or electrical properties. Deliberate controlling of the spatial distribution of reinforcement in the metal matrix composite is one way of these materials. The article describes the theoretical basis and experimental verification of a new method using a homogenised low-frequency electromagnetic field to produce particle-reinforced metal matrix composites with a controlled distribution of reinforcement. The method developed allows for obtaining local reinforcement of the composite casting, analogous to radial centrifugal casting. The article describes the methodology for designing a casting system, which allows such control of the electromagnetic field distribution in the liquid casting, which makes it possible to obtain the desired reinforcement distribution. Experimental verification of the method developed was carried out using the example of a sleeve made of AlSi12CuMg alloy reinforced with SiC particles at the outer wall.

Neural network committees optimized with evolutionary methods for steel temperature control

This paper presents regression models based on an ensemble of neural networks trained on different data that negotiate the final decision using an optimization approach based on an evolutionary approach. The model is designed for big and complex datasets. First, the data is clustered in a hierarchical way and then using different level of cluster and random choice of training vectors several MLP networks are trained. At the test phase, each network predicts an output for the test vector and the final output is determined by weighing outputs of particular networks. The weights of the outputs are determined by an algorithm based on a merge of genetic programming and searching for the error minimum in some directions. The system was used for prediction the steel temperature in the electric arc furnace in order to shorten and decrease the costs of the steel production cycle.

Shaping inductor geometry for casting functionally graded composites

Purpose – The attractiveness of functionally graded composites lies in the possibility of a gradual spatial change of their properties such as hardness, strength and wear resistance. The purpose of this paper is to discuss the use of electromagnetic buoyancy to separate the reinforcement particles during the casting process of such a composite. Design/methodology/approach – The basic problem encountered in the process of casting composites is to obtain electromagnetic buoyancy and simultaneously to avoid a flow of the liquid metal which destroys the desired composite structure. In this paper the authors present the methodology of numerical optimization of inductor geometry in order to homogenize the electromagnetic force field distribution. Findings – The optimization method based on searching the solution subspace created by applying knowledge of the modelled process physics proved better than the universal local optimization methods. These results were probably caused by the complex shape of the criteri...

A MLP solver for first and second order partial differential equations

A universal approximator, such as multilayer perceptron, is a tool that allows mapping of any multidimensional continuous function. The aim of this paper is to discuss a method of perceptron training that would result in its ability to map the functions constituting the solutions of partial differential equations of first and second order. The developed algorithm has been validated by means of equations whose analytical solutions are known.

IMPURITIES REMOVAL PROCESS FROM THE VACUUM INDUCTION FURNACE CHARGE: A NUMERICAL STUDY

The technology of mental melting in a vacuum induction furnace enables the efficient removal of impurities and provides an opportunity to melt refractory metals, such as titanium. These materials can be applied in cutting edge technologies, such as aviation (turbine blades) and biotechnology (prosthesis and implants). To control metallurgical heat and mass processes within an induction furnace, measurements and a numerical analysis can be conducted. In this paper, numerical approaches are discussed. Simulation requires the development of vacuum induction furnace coupling between fluid dynamics and electromagnetic fields. The proposed numerical domain was modelled as a threedimensional slice with a properly defined periodic boundary condition. To define the analysed electromagnetic problem, a set of Maxwell differential equations was specified. A fluid dynamics sub-model was composed of the mass and momentum conservation equations using the volume of fluid multiphase formulation, two-equation k- turbulence model and species transport to track the inclusion position within the melt. The main purpose of this study was an examination of the impurities removal process via the free surface of the melt within an induction furnace. The coupled computations were performed for five operating conditions, including different power inputs of the inductor. The results indicated a strong influence of the inductor power on the free surface area and therefore on the purification process intensity.

Modeling Acoustic Effects During Casting Nanocomposites Under Electromagnetic Field

Casting is one of the most cost-effective methods of producing metal matrix nanocomposites. However, it is extremely difficult to disperse nanoparticles uniformly in a metal matrix due to their large surface-to-volume ratio and their low wettability in liquid metal, which cause their agglomeration and clustering. This paper presents a model of the process of nanocomposite casting where an alternating electromagnetic field is used to induce a standing acoustic wave in the molten metal. The model requires a coupling of electromagnetic field, metal flow field, and acoustic field. It allows optimization of the parameters of the coil supply so as to produce a cavitation phenomenon in the metal, which breaks up the agglomerations of nanoparticles.

Model of Infiltration of Spent Automotive Catalysts by Molten Metal in Process of Platinum Metals Recovery

This paper presents the model for the washing-out process of precious metals from spent catalysts by the use of molten lead in which the metal flow is caused by the rotating electromagnetic field and the Lorentz force. The model includes the coupling of the electromagnetic field with the hydrodynamic field, the flow of metal through anisotropic and porous structure of the catalyst, and the movement of the phase boundary (air-metal) during infiltration of the catalyst carrier by the molten metal. The developed model enabled analysis of the impact of spacing between the catalysts and the supply current on the degree of catalyst infiltration by the molten metal. The results of calculations carried out on the basis of the model were verified experimentally.

Evolutionary optimized forest of regression trees: application in metallurgy

A forest of regression trees is generated, with each tree using a different randomly chosen subset of data. Then the forest is optimized in two ways. First each tree independently by shifting the split points to the left or to the right to compensate for the fact, that the original split points were set up as being optimal only for the given node and not for the whole tree. Then evolutionary algorithms are used to exchange particular tree subnodes between different trees in the forest. This leads to the best single tree, which although may produce not better results than the forest, but can generate comprehensive logical rules that are very important in some practical applications. The system is currently being applied in the optimization of metallurgical processes.