Design of low frequency disturbance filter based on data fusion and its application to radiant signal filtering

Due to the big-lagged and big-inertial characters of the load control object in thermal power units, it´s difficult for conventional feedback signals to guarantee the rapidness and precision simultaneously. A low frequency disturbance filter is designed based on data fusion theory for filtering the measured radiant energy signal in boiler furnace. The low frequency disturbance of the signal is gotten rid of without any phase lags by using general observer configuration and frequency domain cross-complementary knowledge. The observer has additional degrees of freedom for design, and the cross-complementary knowledge is measured from different sources, which are the radiant energy and the heat signal in the paper. The application to the measurement of radiation signal in a boiler furnace shows that the method can remove the low frequency disturbances effectively and is robust against model perturbation.