Optimization of a Coal Fired Boiler Using Artificial Immune System
Łukasz Śladewski , Konrad Świrski
AbstractOn-line optimization of power boilers is a very important and challenging issue in research and implementation, particularly in the context of increasing environmental requirements. Combustion process is a complex MIMO process (Multi-Input-Multi-Output). A large inertia and non-linearity of the combustion combined with frequent changes of disturbance signals like boiler load and fuel quality necessitated the searching for new solutions in the field of control and optimization. There are plenty of algorithms used for combustion optimization, ranging from MPC through neural network etc. Most of the algorithms ae inspired by phenomenon that could be observed in the nature. The newest algorithm, that has been successfully applied in combustion optimization project is artificial immune system. The artificial immune system like real immune system has ability of adaptation. Once a body is infected by a known pathogen (bacteria or virus) the immune response - antibody production is much faster and the illness is less painful to the body, compering to infection by a new pathogen. Artificial immune system is a self-learning algorithm - it searches and remembers most effective solutions for specific process conditions. The advantage of IT system with artificial immune algorithm is ability of fast adaptation to continuously changing conditions with multi-criteria optimization. The artificial immune algorithm has been applied combustion optimization projects in miscellaneous hard coal, lignite, gas and oil fired power boilers with capacity ranging from 300 t/h to 2300 t/h. This paper presets example combustion optimization results.
|Publication size in sheets||0.5|
|Book||2019 IEEE 6th International Conference on Energy Smart Systems (ESS), 2019, IEEE, [978-1-7281-2160-4]|
|Score||= 5.0, 09-06-2020, ChapterFromConference|
|Publication indicators||= 0; = 0|
* presented citation count is obtained through Internet information analysis and it is close to the number calculated by the Publish or Perish system.