Research And Application Of Vibration And Noise Data Compression Algorithm For Auxiliary Equipment Of Generator

Monitoring and diagnosis of auxiliary equipment of generator are important condition to ensure the normal operation of the whole unit.Extraction of auxiliary equipment data information is very important for the management of the plant and the reliability and safety of the generator unit.As sampling of the auxiliary equipment data is in high frequency and uninterrupted,how to store and transmit the data at least cost without losing the information in the data becomes a key problem to be solved.At present,many algorithms have been used to compress the power quality and fault data,but the research for data compression of the auxiliary equipment of the power plant is in the primary stage.In this paper,it collected and analyzed the existing data compression algorithms in power systems and make some improvement in the algorithm for the auxiliary equipment data.The proposed method has achieved good results after computing and analyzing auxiliary equipment data from three aspects(lossless compression,denoising,lossy compression),and the simulation result has verified the effectiveness of the method.Firstly,this paper uses the lossless compression algorithm to compress the data,the fast Fourier transform is used to find the fundamental frequency component of the signal which is used as the signal's prediction component,and then encoding the residual of the signal.The proposed LZW algorithm is used in residual coding.In this algorithm,the dictionary length is used to map the prefixes and it eliminates the redundancy and improves the perform compare with the exist algorithms,and achieves better results than the entropy of the data.Simulation results proved the effectiveness of the LZW algorithm.Secondly,considering that signal contains different kinds of noise,this paper uses the wavelet transform to denoise the signal,and stores the wavelet domain coefficient to realize the data compression.A smoother threshold function is proposed in the signal denoisin,it is more conducive to preserve the high frequency information of the signal than the soft threshold function,and more conducive to reduce the shock than the hard threshold function.After simulated,the function testified feasible.Finally,wavelet packet is employed in this paper for signal lossy compression.Due to the periodic characteristics of the power system data,we reshape the original signal into two-dimensional signal for compression.After using the wavelet packet to respectively compress the original signal and the reshape signal,the reshaped signal achieved better compression performance.In this paper,the SVD algorithm is used to compress the reshape signal,and a compression method which combining SVD and wavelet packet is proposed.The unitary matrix of SVD is reorganized by the new method,and the reorganized matrix is compressed by the new method.The SVD algorithm achieves a larger compression ratio than wavelet packet decompose algorithm after tested by simulation,but the reconstruction error is bigger.The proposed method can achieve a large compression ratio under a small reconstruction error cost compare to the SVD.