Coal Identification And Heat Estimation Based On Flame Detection

It has been proven that to know the type of coal that the boiler is burning in time is very important to the thermal power plants. But the identification of the coals online is difficult, especially in China where blended coals are widely used. In this paper, research is done on the on-line identification of coal types and the estimation of unknown coals. The main innovation of this thesis is as follows.1) On-line coal identification in a Drop-tube furnace. Three flame features including the average value, the standard deviation and the oscillation frequency are firstly extracted by a flame detector. These extracted features are then used to form the feature matrixes. By calculating the correlation coefficients of the matrixes, the coals can be well identified. For the single coals, only by comparing the correlation coefficient of the average matrix the coals can be identified. For the blended coals, comparing the correlation coefficient of average matrix we can make sure which two single coals the blended coal is consisted of. And the mixing ratio can be known by further comparing the correlation coefficients of the standard deviation and oscillation frequency matrixes. As the method does not need complex estimation, it can be used to identify coals fast online. The experimentations has been done on the drop-tube furnace, where 28 kinds of single coals and many blended coals are tested, and the results show that the method works well for the identification of the coals.2) Estimation of the heat of an unknown coal on line. Through the experiments in drop-tube furnace, it is found that when the volatile content of the coal is constant, there is a linear relationship between the heat of the coal and the average response spectrum of the coal, and that when the fixed carbon of the coal is constant,, there is a linear relationship between the heat of the coal and the average response spectrum of the coal too. bringing the above two points and the traditional fuel recognition technology, the heat of an unknown coal can be calculated. At the same time, the method is given which is used to calculate the heat of coal by building BP Neural Networks although the technological parameters are unknown.