Coal Quality Analysis Based On Nanosecond And Femtosecond Laser-induced Breakdown Spectroscopy

As the main primary energy source in China,coal is also the major fuel of thermal power generation.The coal quality parameters are related to the safe and efficient operation of thermal power generation equipment and the realization of energy conservation and emission reduction goals.Therefore,there is an urgent need for a real-time online detection method of coal quality.Laser-induced breakdown spectroscopy(LIBS),as a new spectroscopic analysis technology,can realize realtime online detection of sample composition and content,and has great potential and application prospects in the field of coal quality analysis.But at present,LIBS still has the problems of low detection accuracy and poor repeatability in coal quality analysis.Femtosecond laser is widely used in many fields because of its narrow pulse width and high spatial resolution.It also provides a new direction for the development of LIBS.In this paper,nanosecond laser-induced breakdown spectroscopy(ns-LIBS)and femtosecond laser-induced breakdown spectroscopy(fs-LIBS)are used to detect coal quality.The differences and advantages of plasma spectra between them are analyzed in detail.The two systems are combined with two multivariate regression analysis models for quantitative analysis of ash,volatile matter and calorific value of coal,which verifies the feasibility of fs-LIBS in the field of coal quality analysis.The main work is as follows:(1)In this paper,the formation mechanism of laser-induced plasma and the basic principle of spectral quantitative analysis are investigated.Four kinds of multivariate regression analysis models are introduced.The basic principles and advantages and disadvantages of partial least squares(PLS)and extreme learning machine(ELM)are emphasized.(2)Two systems of ns-LIBS and fs-LIBS were built,and the timings of the two systems were set separately.The influence of laser energy,delay time between laser pulse and the beginning of spectrometer,the integration time of spectrometer and the distance between laser focus and sample surface on the spectral signal are analyzed,and the experimental parameters are optimized.(3)The effects of ns-LIBS and fs-LIBS on the ablation pits of standard coal samples,the time evolution characteristics of plasma spectra,the differences of spectra under the optimal delay time,the time evolution characteristics of plasma electron temperature and the stability of spectral signals are compared.The results show that fs-LIBS plasma spectra have smaller ablation pits,shorter plasma lifetime,smaller continuous background noise,lower plasma electron temperature and faster decline trend,and more stable spectral signal.However,there is a problem of weak overall spectral intensity,which results in some elements with lower content being masked by continuous background noise.(4)In order to solve the problem of large fluctuation of spectral data and interference of background noise in LIBS experiment,spectral preprocessing of abnormal data removal,multiple measurements averaging and baseline correction is carried out.The calibration results show that the linear fitting degree of calibration curve is improved obviously after pretreatment.Two multivariate analysis models,partial least squares(PLS)and particle swarm optimization extreme learning machine(PSO-ELM),were introduced to quantitatively analyze ash,volatile matter and calorific value of standard coal samples.The results show that the plasma spectrum produced by fs-LIBS can better represent the actual composition and content of standard coal samples in LIBS system,and the prediction accuracy of particle swarm optimization extreme learning machine is higher in prediction model.