NO_x Emission Prediction Of Coal-fired Boiler Based On Image Thermometric Method

Coal-fired power plant is one of the important sources of global NO_x emissions. To predict NO_x emissions from coal-fired power plant not only has guiding role to combustion adjustment, but also is of great significance to the protection of the environment. The furnace temperature obtained through the image thermometry and together with other main operational parameters is used to predict the NO_x emissions. This is the purpose of the thesis. Then the effective instruction information of operation can be provided for the power plant personnel. This article takes the image thermometry and the forecast and analysis of NO_x emissions as the primary coverage, and the main content is specifically as follows.A study on NO_x formation mechanism and NO_x emissions laws was presented. According to the data collected, a variety of operating parameters and changes in coal that impacted NO_x emission and the main operating parameters were also mainly studied. The result showed that one of the most important parameters to the emissions of NO_x is the level of the furnace temperature uniformity and distribution. Therefore, it was necessary to carry out the reconstruction of temperature field.Image noise reduction, local enhancement and edge detection were carried on flame digital image, so the characteristics of the image were highlighted. The temperature was measured and the temperature field was reconstructed though colorimetric method and colorimetry mechanism. The trends and pseudo-color for temperature were displayed by Programming with VB, and the temperature was used one of the main parameters to predict NO_x emission.The prediction of NO_x emissions based on principal components analysis and Bayesian Regularization BP neural network. Aiming at the complex framework of power plant NO_x emissions prediction model of neural network because of excessive size of input factors, which leads to decrease the prediction precision, the redundant data was removed through principal component analysis. And Bayesian regularization method was used to improve the traditional BP neural network. Taking the data experimented on 300MW coal-fired boiler in different working conditions as an example, the NO_x emissions were predicted by the improved Neutral network. The result was contrasted with the synthesis haze analyzer's observed value. The results showed that the method could effectively reduce the size of the model and the generalization capability of the model was better than the traditional neural network. The effective means that the characteristics of NO_x emissions was forecasting by the improved Neural network could reduce NO_x emission and increase boiler efficiency by adjusting combustion condition. What's more, the influential factors on NO_x emission were studied by means of Grey Relational Analysis, the results can be a reference for the adjustment of combustion conditions for reducing NO_x emission.Finally, the prediction system interface and module development of image temperature measurement and NO_x emissions developed by Matlab and VB mixed programming.