The On-line Monitoring System Of Fouling And Slagging Based On NN In Power Plant

The on-line monitoring system of fouling and slagging on heating surface, which changes blind sootblowing in power plant, is based on NN(Neural Network) in this paper. The system can help to finish suitable sootblowing. The mechanism, function and characteristic is set forth in the paper. And then introduce the improved BP(Back-Propagation) arithmetic. Combined the actual performance,the character of boiler units with RPROP ,a nonlinear model is designed. After trained,the model can predicts heat absorption of clean heating surface in every condition, meanwhile, the actual heat absorption can also be got. Finally the dirty coefficient of dust is calculated. The coefficient can reflect the degree of fouling and slagging effectively. The design, training and predict are all complished in Matlab because it has a strong NN function. The real time communication between Matlab and the on-line monitoring system of fouling and slagging is finished by DLL,which is written by VC.The local data-base is SQL Serve and interview by ODBC in order to optimize the speed of the system.