Research On Impulse Noise Suppression And Symbol Synchronization For Smart Grid OFDM System

In recent years,China’s requirements for energy conservation,emission reduction and sustainable development are increasing.As one of the important means of energy conservation and emission reduction in the new era,smart grid has also ushered in unprecedented development opportunities and challenges.As the basic equipment of smart grid data acquisition,the corresponding data transmission mode of smart meter has always been one of the key research directions.Due to the existence of impulse noise in the transmission channel,the synchronization algorithm designed for Gaussian white noise channel is difficult to maintain the original performance,which poses a great challenge to the accurate synchronization of the receiver.Based on IEEE 802.15.4g,thesis builds a Simulink model around the mr-ofdm physical layer,and studies the impulse noise suppression and symbol synchronization technology on this basis.Traditional symbol synchronization algorithms are treated separately from noise suppression as a separate part.Thesis adds a simple means of noise suppression on the basis of symbol synchronization to further improve the synchronization accuracy.For impulse noise suppression,thesis analyzes and compares the impulse noise suppression effects of three common noise suppression methods:compressed sensing,iteration and nonlinear processing based on threshold;For symbol synchronization,thesis studies and simulates the symbol synchronization algorithm based on training sequence and the synchronization algorithm independent of the training sequence,and analyzes the synchronization accuracy.Through the above research,thesis proposes a symbol synchronization scheme based on the frame structure specified in IEEE 802.15.4g protocol.Firstly,the nonlinear noise suppression technology is used to perform simple noise suppression on the received signal.Then,Short Training Field（STF）is used to perform coarse synchronization and coarse frequency offset estimation.Finally,Long Training Field（LTF）is used to complete precise synchronization and accurate frequency offset estimation of received signals.For rough synchronization,this thesis designs five rough synchronization schemes based on STF by using the inverse part of frame structure and the idea of weighting,and carries out corresponding simulation experiments.The results show that the schemes 1 and 3proposed in thesis have better frame synchronization performance under the condition of multipath fading and low SNR.In the improved fine synchronization algorithm,the timing measurement function of the second stage can be obtained by optimizing the timing measurement of the first cross-correlation stage according to the optimization peak obtained in advance by LTF.This algorithm increases the synchronization accuracy by balancing the sub peak.Compared with other symbol synchronization algorithms that can be applied to the scene studied in thesis,the improved algorithm proposed in thesis has better MSE performance under low signal-to-noise ratio and lower algorithm complexity.It is suitable for battery powered simple communication devices such as smart meters.