Research On The Online Identification Technology Of Coal Quality Based On The Characteristics Of Flame Emission Spectrum

In order to solve the problems caused by changes in coal combustion and coal types,domestic power plants have begun to adopt various boiler combustion optimization control technologies.The optimization system needs to obtain real-time coal quality parameters,so as to make self-adaptive adjustment according to the characteristics of coal quality.Therefore,it is urgent to develop on-line detection technology of coal quality parameters.However,the commonly used on-line detection technology of coal quality is mostly installed on the feeding belt,and there is a time difference from the time of feeding into the furnace.There is still a certain uncertainty in the identification results.At the same time,the installation cost is high,and the detection environment also has high requirements.If the relationship between the flame combustion spectrum and the coal type in the furnace can be explored,and the on-site DCS data can be used as a supplement,an online coal quality identification model can be established and developed,and real-time coal quality parameters can be obtained,which will help the combustion adjustment,and will be beneficial to the improvement of thermal power generation.It is of great significance to reduce the coal consumption of the unit boiler efficiency.In this paper,the flame emission spectrum of a coal-fired boiler is taken as the research object,and the correlation between the coal type and the flame emission spectrum is established through the atomic characteristic lines of Na,K,Li and other elements on the flame emission spectrum,and based on this,to realize the online identification of coal quality,and provide real-time coal quality information for the combustion adjustment of thermal power units.Firstly,a small flame spectrum identification device was established in the laboratory,and the flame emission spectra of three biomass fuels were analyzed and processed.The direction has good similarity and regularity,so the coal quality identification characteristic quantity is extracted from the flame spectrum in the wavelength direction;the characteristic spectral line intensities of the three alkali metal elements Na,K,and Li of the three biomass fuels have relatively large intensities.Differences,the use of sodium-potassium ratio and sodium-lithium ratio as identification features can well distinguish the three biomass fuels;Na,K atomic characteristic spectral line intensities will change with the change of fixed carbon radiation intensity and combustion temperature,so the radiation intensity Intensity and combustion temperature will also affect the ratio of sodium-potassium and sodium-lithium,which can be used as supplementary characteristic quantities for coal quality identification;at the same time,the contents of three alkali metal elements,Na,K,and Li in different coal types are indeed quite different.,and they have obvious excitation lines in the flame emission spectrum,which can be used as identification elements for coal quality identification.Secondly,the on-site test equipment was built,and the integrated optical fiber sensor was used,which eliminated the installation and debugging of coal detection probes,greatly reduced the difficulty of system installation and debugging,reduced the workload of transformation,and saved the system.At the same time,the spectrometer and the industrial computer were installed in the control cabinet of the electronic room;the on-site spectral data was collected,and the data was preprocessed,including the subtraction of the dark spectrum and the use of wavelet decomposition and reconstruction to separate the characteristic spectral lines and spectral baselines;Due to the large disturbance of the spectral data due to various factors in the field,various statistical parameters such as sodium-potassium ratio,sodium-lithium ratio,radiation intensity,and combustion temperature were selected as the characteristic quantities for identification,and a preliminary identification was established using these characteristic quantities.model,and obtained reliable accuracy results.Finally,the parameters of the identification model are optimized,and the grid optimization algorithm and K-fold cross-validation are used to optimize the adjustable parameters of the support vector machine,and finally the optimal parameters are obtained,which improves the accuracy of the identification model.Sensitivity analysis of the characteristic parameters of the The potassium ratio plays the most prominent role,followed by the sodium-lithium ratio and combustion temperature,and the radiation intensity can be selected according to the specific identification effect;because the DCS data can reflect the operating status of the unit,and the operating status of the unit will affect the combustion status of the flame.Therefore,the DCS data was added to the identification model as an auxiliary identification feature quantity,and it was found that the load,total coal volume,total air volume,and flame temperature of the furnace measurement point could improve the accuracy of the identification model,while the furnace measurement point The flame temperature responds more directly to the combustion conditions,and the optimization effect is also the best.