Research On Optimization Of Voltage Sag Monitoring Points And Signal Identification

With the development of social economy and science and technology in China,more and more sensitive loads such as power electronic equipment and computers will be connected to the power system,and the power quality requirements of power grid power supply will be more and more strict.In the power quality problems,voltage sag has become the focus of engineering and academic fields because of its high frequency and serious economic loss.So it is very important to study the related problems of voltage sag.The optimization of voltage sag monitoring points and the identification of disturbance sources are taken as the starting points.The main work of this research is as follows:This article first introduces the current definition of voltage sag and briefly describes the causes and possible harms of voltage sag.Mainly carried out detailed analysis and research on the three causes of voltage sag,and established a model to simulate and verify various causes,and analyzed the voltage change trend and recovery ability before and after the sag occurred.Secondly,it introduces several detection algorithms commonly used at present,including effective value method,fundamental component method,defect voltage method,instantaneous d-q transform method,wavelet analysis method,S transform method,and validates the test results of some algorithms,and analyzes the principles of these algorithms and their advantages and disadvantages in the application level.Then it analyzes the layout method of voltage sag monitoring points.Because of the huge power system and the complex number of nodes,it is not suitable to install monitoring devices in each node.At the same time,the storage and transmission of monitoring data will be more complex.Taking IEEE-39 node as an example,according to its network structure and parameters,a linear programming algorithm is used to optimize the location of monitoring points,which can meet the requirements of online monitoring of all nodes and all kinds of faults in the system under the limited number of monitoring points.Finally,a voltage sag feature extraction and identification method based on deep confidence network(DBN)is proposed.The feature extraction capability of DBN is used to perform feature self-extraction on measured waveform data,which solves the problem that manual feature extraction relies too much on expert experience and is affected by unknown features Larger non-general problems.The model integrates the feature extractor and the classifier,which improves the efficiency of sag source identification,and proves the superiority of the method in feature extraction and sag source identification,which is suitable for practical engineering.The paper has 26 pictures,6 tables,and 73 references.