Application Of Reversed Modeling Method In CFB Bed Temperature Parameters Modeling

The reversed modeling method was given serious attention because of its adaptability in complex thermodynamic system modeling and its convenience in application of engineering. The advantages of circulating fluidized bed (CFB for short) boiler are environmental protection and fuel adaptability, which depend on the controlling of bed temperature. It will discharge excessive pollutants while temperature is too high and will be fire extinction while temperature is too low. So it is very meaningful of modeling of bed temperature.The reversed modeling method was introduced briefly, and application examples in CFB boiler parameters modeling were given. Examples concluded the modeling of monitoring and forecasting of bed temperature, by the use of BP and Elman neural networks algorithm.It was as possible as not to make assumptions or simplifying to reduce the influence of human factors. In data selecting, strong correlation was chosen by.grey relational grade.The results of modeling of monitoring the bed temperature suggest that the neural networks algorithm has better results in approximation of target parameter and BP algorithm is better than Elman algorithm. In the modeling of short-term forecast of bed temperature, it is very important to make sure the optimum time step. In the time interval of 5.4 minutes, Elman neural network shown strong predictive power and accuracy to maintain a continuous ability and predicted more than 40 minutes.Finally, the algorithm of continuous short-term forecast of CFB bed temperature was given. The reverse modeling method applications of some parameter can be extended to other parameters.