SOM-Based Method for Process State Monitoring and Optimization in Fluidized Bed Energy Plant

Self-organizing maps (SOM) have been successfully applied in many fields of research. In this paper, we demonstrate the use of SOM-based method for process state monitoring and optimization of NOx emissions. The SOM was trained using a dataset from a fluidized bed energy plant. Reference vectors of the SOM were then classified by K-means algorithm into five clusters, which represented different states of the process. One neuron in each cluster was defined optimal based on the NOx emission of the process. The difference between reference vectors of the optimal neuron and the neuron in each time step could be used for determination of reasons of non-optimal process states. The results show that the SOM method may also be successfully applied to process state monitoring and optimization of NOx emissions.