Design And Implementation Of Coherent Laser Wind Radar Signal Processing

The development of clean energy is one of the important goals of the country’s "14th FiveYear Plan".Wind power,as a major component of clean energy,has also received more and more attention.The detection of wind field information plays an important role in selecting sites for wind power generation,reducing losses and increasing power generation.Compared with the traditional wind measurement technology,the coherent laser wind radar has the advantages of good real-time performance,high accuracy and strong resolution.In addition to wind power generation,it has also used in airport wind field detection,missile launch Angle correction and weather forecasting functions.At present,in the application scenario of wind power generation,pulsed coherent laser wind measurement radar still has problems such as insufficient effective measurement frequency,excessive consumption of computing resources,insufficient measurement accuracy,high price,and insufficient applicability in actual scenarios.These problems lead to the low popularity of wind measurement radar in wind power generation and insufficient improvement in power generation efficiency.In response to the above problems,this thesis takes wind power as the scene,researches and designs a coherent laser wind radar digital signal processing system based on FFT multiplexing and data pipeline processing structure,and makes customized optimization and multi-scenario configurable solutions for wind power generation scenarios..The digital signal processing system and the optical system were jointly debugged to complete the construction of the radar prototype.Finally,it was tested and compared with the radar in the same application scenario on the market.In order to improve the efficiency of radar processing,a new coherent laser wind measurement radar digital signal processing structure is constructed.This structure not only uses the method of multiplexing FFT modules to reduce the overall resource consumption,but also uses methods such as ping-pong RAM to break through data congestion to achieve data pipeline processing,greatly improving the processing speed,and the overall parameters can be configured.FFT is the module that consumes the most resources and time in digital signal processing.This thesis uses Chisel language to design a high-performance configurable FFT module,completes the design of three data types FFT module,and compares it with all aspects of Xilinx IP.The results show that the block floating-point data format FFT module designed in this paper is 100% faster than Xilinx IP,and has the best performance.On this basis,for the actual scene,the circuit structure of the real sequence FFT algorithm is designed to further improve the calculation efficiency.In order to improve the accuracy of wind speed measurement,it is necessary to improve the accuracy of frequency estimation.As a discrete operation,FFT needs to be corrected for discrete spectrum.After simulating the energy center of gravity method and the interpolation method,it is found that the accuracy is not enough in this data zero-filling environment.Therefore,this thesis proposes a discrete spectrum correction method for amplitude comparison,which also has good correction accuracy under zero-padding conditions,which improves the theoretical wind speed measurement accuracy to more than 200% of similar products.In other respects,in response to the problem of excessive non-linear computing resources and time consumption in the floating-point format complex number modulus in this design,this paper proposes the modulus approximation algorithm for the floating-point format complex number for the first time,which changes the traditional modulus approximation algorithm.Making full use of the characteristics of floating-point numbers for coefficient grouping,the maximum approximate error is 0.35%,which is more than 1.5 times more accurate than the traditional modulo approximation algorithm,and the resources and clock cycles consumed are only less than 50% of the direct calculation.On the ZYNQ7035-based radar signal processing board,Verilog HDL and Chisel language are used to complete the design of the overall hardware structure.After simulation and synthesis on the board,the joint debugging is carried out with the optical part in the laboratory environment to complete the continuous wind laser radar.Radar detection of water vapor and atmospheric measurement experiments in two environments.On this basis,the prototype of the pulse laser wind measurement radar was completed,and the comparison and measurement were carried out with similar radars in an external room temperature environment.The comparison results are consistent and the self-developed radar has better parameters.The main parameters are that the range resolution is : 15～75m configurable,cumulative times configurable for 4000,effective measurement frequency is up to 10 Hz,Doppler frequency shift accuracy is 0.06 MHz,and the corresponding radial wind speed accuracy is 0.045 m/s.