Coal-fired Boiler Furnace Flame Temperature Field And Concentration Field Measurements And Combustion Diagnosis

The national environment codes have been enforced stricter regulations that demand cleaner environments, such as the reduction of by-product of the fossil fuel flame. If the combustion of the flame inside a boiler could be optimized, the usage of fuel and the amounts of pollution produced would be significantly reduced, and many of the common boiler tube can be avoided. This would result in a major costing saving to the utility industry, and would provide a cleaner environment to the society.The combustion flame image grabbed by CCD camera is processed to transform the luminous light signal into the spectral radiation signal. This luminous emission sensible to measuring system comes from tiny char particles, which are almost pure carbon and are generated during the combustion process. Based on the radiation image of coal-boiled flame, we have developed a set of measurement technique, including measurement of project temperature, measurement of intersection temperature distribution etc.Strict definition of project temperature is given in this article. Consideration of BP arithmetic of neural network has good character in function approach, we use it to ameliorate the colorimetric method to calculate the project temperature.After dividing the measuring section into minor grids, intersection temperature distribution measurement model > three-dimensional temperature measurement model and char concentration measurement model is established. According to the model, feasible solving algorithm is designed and numerical simulation analysis is applied. Finally, we use the new measurement technique to do experiments on coal-burned flame and receive satisfied results.This paper also introduces the application of the developed flame image processing system in the boiler of one 300MW power plant in detail. The main function guiding ideology and general situation of the whole system are explained. From the grabbed flame image inside the furnace, Character value is distilled from the original image. Kohonen Neural Network is designed andtrained to discern combustion state at the present time and BP Neural Network is designed and trained to forecast combustion state on hand. At last, a two-step fuel control system, which takes the radiation temperature as an intermediate controlled parameter is designed to improve the control character of boiler. Control model and respective control strategy is introduced and simulation calculation result is demonstrated in this article.