Neural Network Combustion Optimisation in Naantali Power Plant
An optimisation system giving advice to the operators is installed to unit 2 of a coal-fired power plant located in the city of Naantali, Finland.
The basis of the optimisation is a model of the steady state behaviour of the boiler. Due to the complexity and severe non-linearity of the process, a neural network approach was chosen.
Control inputs to the boiler model include oxygen in the flue gas, the percentage of burning air fed as over fire air, the coal/air -ratio and the tilting angle of the burners. Outputs of the model are the boiler efficiency, the level of NOxemissions, the unburned carbon in the fly ash and the maximum material temperature in the reheater tubes of the boiler.
The optimiser operates in an advisory mode allowing the operators to acquire optimal setpoints for control variables. It can either optimise any of the outputs while specifying limits for the other outputs and allowable values for the control variables. Or with the what-if simulation, the operator can study the effect of changes in the control variables in advance.
KeywordsSteady State Behaviour Unburned Carbon Boiler Efficiency Boiler Model Reheater Tube
Unable to display preview. Download preview PDF.
- Mokhtar S. Bazaraa, Hanif D. Sherali, and C. M. Shetty. Nonlinear Programming; Theory and Algorithms. John Wiley & Sons, second edition, 1993.Google Scholar
- Howard Demuth and Mark Beale. Neural Network Toolbox User’s Guide. The Math Works Inc., 1998.Google Scholar
- Simon Haykin. Neural Networks; A Comprehensive Foundation. Prentice Hall, 1994.Google Scholar