Research And Application Of Widely Linear Minimum Variance Distortionless Response

The traditional frequency estimation based on frequency estimation has good performance in balancing power systems,but when used in unbalanced power systems,the performance of estimation bias,estimation variance,and real-time performance are significantly deteriorated.In this paper,the frequency estimation algorithm based on the wide linear MVDR spectrum in unbalanced power system is proposed and improved.The main work of this paper is as follows:The research background and significance of frequency estimation under unbalanced power system are introduced.The research status of frequency estimation under unbalanced power system both at home and abroad are briefly outlined,and the development of wide linear MVDR spectrum is discussed.Then some basic techniques of frequency estimation based on broad linear MVDR spectrum are studied and analyzed.The concept of broad linearity,wide linear modeling and mathematical representation of unbalanced power system are discribed.At the same time,the lower bound of Cramers' frequency of the frequency estimation is deduced,and several types of voltage sag are compared.Finally,the widely linear MVDR spectrum is briefly presented,that is,the minimum variance spectrum estimation.The basic MVDR spectrum is studied in detail.At first,a wide linear MVDR spectrum model is established and a frequency estimation method based on it is derived.The matrix inversion problem encountered in the process of frequency estimation using the wide linear MVDR is proposed several methods,as well as some approaches to obtain the inverse matrix during the whole simulation were employed.Secondly,the optimal wide linear MVDR spectrum estimation model is posed,and the frequency estimation for unbalanced power system based on the Newton iterative method and the optimal wide linear MVDR spectral estimation model is derived.In addition,this paper demonstrates the original Capon minimum variance distortionless response estimation method and several other estimation algorithms derived from original MVDR spectrum.The system frequency estimation based on these spectrum estimation algorithms is deduced and the iterative MVDR estimation algorithm,augemented MVDR algorithm and approximate maximum likelihood estimation algorithm were analyzed.In this paper,the frequency estimation algorithm based on the wide linear MVDR spectrum,the optimal wide linear MVDR spectrum and the improved wide linear MVDR spectrum are simulated in unbalanced power system.The simulation results show that the performance of the proposed method based on the wide linear MVDR spectrum and the optimal wide linear MVDR spectrum estimation algorithm and the traditional iterative MVDR estimation algorithm are not obvious when used in the unbalanced power system.When the inverse algorithm of the ill-conditioned matrix in the algorithm is improved by adding coefficients,the frequency estimation performance based on the improved wide linear MVDR spectrum is better than iterative MVDR algorithm and ACLMS algorithm in terms of estimation error,variance and online tracking.Especially in the case of frequency modulation,the improved wide linear MVDR algorithm(block estimation)is superior to iterative MVDR,ACLMS algorithm and augmented MVDR algorithm in terms of both tracking delay and estimation error.