Research On Echo Characteristics And Wind Velocity Estimation Algorithm Of Coherent Doppler Wind Lidar

The Coherent Doppler wind Lidar(CDWL)is an important tool for marine atmospheric wind field measurement.It can continuously,real-time and accurately obtain wind field data.It has the advantages of high measurement accuracy and strong real-time performance.It is used in offshore wind energy utilization and environmental monitoring and other aspects have important applications.When the CDWL measures the atmospheric wind field,it receives the backscattered signal of the aerosol.Because the composition and characteristics of the aerosol will be affected by factors such as region,climate,environment,etc.,the received scattered echo signals are also different.Therefore,this thesis studies the characteristics of the echo signal of the CDWL for different marine aerosols under marine atmospheric conditions.Meantime,in view of the low signal-to-noise ratio of the echo signal,the performance of the wind velocity signals processing algorithm is analyzed in detail.This has important reference significance for improving the accuracy of wind velocity estimation and determining the parameters of the CDWL in different aerosol environments.In this thesis,based on the principle of the CDWL,the three heterodyne detection methods of Doppler signals are introduced,focusing on the analysis of the signal-to-noise ratio of heterodyne detection.For the velocity measurement system,there is an optimal local oscillator optical power to maximize the signal-to-noise ratio of the detection system,and the value of the optimal local oscillator optical power is not related to the echo signal optical power.Thus,when analyzing the echo characteristics of different aerosols,the foundation for the value of the local oscillator optical power is laid.Subsequently,the aerosol data built in the OPAC(Optical Properties of Aerosols and Clouds)software package and the MaexPro(Marine Aerosol Extinction Profiles)model data are used to compare the physical properties of aerosols under three different environmental factors(wind speed,fetch,humidity),and the parameters of the existing CDWL are used to compare the physical properties of aerosols.The study found that the extinction coefficient and backscattering coefficient decrease with the increase of wind speed when the fetch is small(coastal conditions);when the fetch is large(marine conditions),they increase with the increase of wind speed.When the wind speed increases to 15m/s,compared with the length of the small fetch(2km),the echo signal power has increased by about an order of magnitude at the large fetch(60km),and the signal-to-noise ratio has increased by about 10dB.In order to accurately extract wind velocity information from the echo signal with low signal-to-noise ratio,it is necessary to use a wind velocity estimation algorithm to estimate the wind velocity.The wind velocity estimation algorithms based on the ideal Gaussian emission pulse shape of the emitted laser pulse are commonly used.However,for the CDWL,the laser pulse shape also satisfies the Lorentz,rectangular and hyperbolic secant distributions.Based on the frequency domain mathematical model and the Kolmogorov turbulence mechanism model,the echo signal of the CDWL under different pulse shapes are simulated.First,the signal model and a real correlation random wind field algorithm are explained.The Cramer-Rao lower bound(CRLB)of the performance of the maximum likelihood(ML)discrete spectral peak(DSP)algorithm under different pulse shapes is analyzed and derived,and compared with the exact CRLB.In addition,the ML DSP algorithm for coherent laser Doppler velocimetry system is applied to simulation signal and summarized by employing the Monte Carlo simulations.The relationships between the SNR,estimation precision,and detection probability under different signal models are simulation and summarized.According to the analysis results of simulated,the CRLB is smallest for Hyperbolic-secant distribution and Rectangle distribution is the largest;the ML DSP algorithm performance is the best for Hyperbolic-secant distribution,while Lorentz distribution has the worst performance.To satisfy the 80%detection probability,the turbulence model of Lorentz requires maximum signal-to-noise ratio(-1 6dB)in very strong turbulence intensity.The simulation results can provide a reference for selecting the appropriate pulse shape and signal-to-noise ratio for the actual CDWL.