Dr. Jaroslaw Krzywanski

Jan Dlugosz University, Poland


Talk Title
Bio-inspired, intelligent and simple models for optimizing of energy systems
Talk Abstract

Different modeling approaches can be distinguished in the open literature. One of the most promising in recent years seems to be the neurocomputing approach, which belongs to the so-called artificial intelligence (AI) methods. These techniques allow reproducing an empirical problem from training samples. Artificial neural networks (ANN) are capable to deal with ill-defined, uncertain, incomplete, and imprecise, redundant and excess data and can generalize interactions between inputs and outputs. Since such kind of data is usually a feature of complex systems, they are sometimes considered to be the methods which can overcome the shortcomings of the laborious programmed computing approach and expensive, time-consuming experimental procedures.

The present work shows the experience in modeling of selected energy processes and devices focusing on the powerful abilities of shallow, smart ANNs structures. Intelligent, bio-inspired models, which have good capabilities of modeling various energy processes and devices with both fixed and moving solids beds applied, are discussed in the study.

Firstly, AI modeling methods of gaseous pollutants emissions: NO and SO2 from coal combustion in circulating fluidized bed (CFB) boilers, are shown in the work.

Air-firing combustion conditions, when combustion runs in air, oxygen-enriched (also so-called air enriched with oxygen or O2/N2 mode) and the oxy-combustion (oxygen-fired combustion) conditions, which means the mixture of oxygen with CO2 or recycled flue gas (RFG) with various fractions of oxygen (O2/CO2 mode and O2/RFG mode, respectively are taken into account.

Moreover, the modeling of cooling production processes from low-grade thermal energy are considered in the study. A conventional two-bed single-stage assembled into a complex hybrid cooling system and an advanced tri-bed twin-evaporator cooler as an innovative design in desalination-cooling systems are presented. The two main energy efficiency factors in cooling production, i.e. cooling capacity (CC) and coefficient of performance (COP) of the adsorption chillers were successfully predicted by the developed models.

Finally, selected parts of energy devices, such as large-scale CFB boilers and adsorption chillers are also analyzed. Membrane walls in the combustion chamber of a CFB boiler and a large falling film evaporator as the most promising evaporator in renewable, adsorption desalination-cooling systems, are taken into account.

These shallow artificial neural networks applied to perform the models turned out to be very effective and powerful tools. An interesting fact is, that even they form small structures, they are smart enough to describe such complex systems. 

The neural networks used in intelligent models allows achieving adequate response via non-iterative procedures, with a low processing time and small memory resources. Good performance of the developed models has been achieved, even for the new, testing data sets, not previously “seen” by the network.

The models can be considered as useful optimization tools of the considered energy processes and devices.

Short Biography

Jaroslaw Krzywanski is an Associate Professor at the Faculty of Science and Technology of the Jan Dlugosz University in Czestochowa, Poland. He serves as the head of the Division of Advanced Computational Methods. He received the M.Sc. degree from the Czestochowa University of Technology, Department of Mechanical Engineering and Computer Sciences, Institute of Thermal Machinery, Poland and a Ph.D. degree from the Silesian University of Technology, Faculty of Energy and Environmental Engineering, Poland. Last two years he obtained a D.Sc. degree (Doctor Habilitatus).

He has published more than 120 refereed works, including papers, monographs, conference proceedings and serves as an editorial board member of several international journals. He has participated in the scientific committee of numerous conferences and serves as a reviewer in a wide range of international scientific journals.

Area of expertise: modeling of energy devices and processes, including solid fuels combustion, gas emissions and hydrogen production from biomass combustion and gasification. He uses both programmed and AI methods (including artificial neural networks, genetic algorithms and fuzzy logic techniques, to predict e.g. heat transfer and pollutants emissions from coal and biomass combustion and co-combustion in large- and pilot-scale circulating fluidized bed (CFB) boilers, chemical looping combustion (CLC) and calcium looping combustion (CaL) in fluidized bed (FB) systems, performance of adsorption chillers in cooling-desalination classical and hybrid systems, modeling of heat exchangers as well as the hydrogen concentration in syngas during the H2 production via CaO sorption enhanced anaerobic gasification of sawdust in FB and CFB units.

Talk Keywords
Bio-inspired modeling; NOx; SO2; heat-exchangers; oxy-fuel combustion; hydrogen; cooling production.
Target Audience
Students, Post doctoral, Industry, Doctors and professors
Speaker-intro video

Sponsors & Partners