Flame Measurement And Coal Identification Based On Digital Image

Coal is the most important energy source of our country. The coal consumption in coal-fired power stations occupies half of the total coal consumption in our country. Therefore, reducing the coal consumption in coal-fired power stations is the main method for achieving sustainable energy development of our country. Nowadays, there are many problems in the operation of boilers in coal-fired power stations, which leading to low energy utilization efficiency and environmental pollution. To improve these phenomena, more comprehensive flame measurement system and online coal detection system should be added on the existed furnace safeguard supervisory system in coal-fired power stations.Currently, in the study of flame measurement based on digital image, great development has been achieved, the research results also have some practical application value. In the study of online coal identification, there are not many researches and the reliability of current technologies in practical applications can not be proven by the research results.For the research status mentioned above, a series of studies have been carried out in this thesis:First, a set of portable flame measurement and coal identification system has been developed, including the hardware for flame monitoring and the software for data processing. The system can capture flame images and light intensity signals, get the spatial and temporal distribution characteristics of the flame to achieve flame measurement, use feature extraction and SVM to achieve coal identification. In addition, the flame measurement system was calibrated by blackbody furnace, making it capable to measure the flame temperature by the two-color method.Second, flame measurement experiment was carried out by the developed system on a one-dimensional coal combustion furnace. The influence of excess air ratio, ratio of primary air and secondary air, coal feed rate and coal quality on the energy release and stability of the coal combustion process was studied.Third, coal identification experiment was also carried out on the one-dimensional coal combustion furnace to assess the identification accuracy. The averaged identification accuracy was up to99.1%. Meanwhile, the influence of each flame feature, the number of samples acquired in each operation case and the number of operation cases for sampling on the identification accuracy was researched.Fourth, the flame measurement experiment for blended coals was carried out on the one-dimensional coal combustion furnace. The blended coals mixed by different coal qualities and proportions were studied, focusing on the relevance of combustion characteristics between the blended coal and the single coals which compose the blended coal.Finally, the identification accuracy for blended coals was tested. The accuracy and the difficulty for blended coal identification were studied by the definition of similarity. The evaluated identification accuracy was up to96.0%, which could meet the requirements for industrial applications.