Research On Spatio-Temporal Association Comprehensive Diagnosis Of Complex Power Quality Disturbance Events

With the extensive use of large capacity nonlinear,impact,and fluctuating loads,and the high penetration of new energy power generation equipment in power grid,power quality(PQ)problem of modern power grid is becoming more and more prominent.Effective analysis and handling on PQ monitoring data by using large scale intelligent monitoring management system is the key to improve the PQ of power grid.In complicated power grid environment,multiple PQ disturbances detected by monitoring system may be related in causal space-time and constitute complex disturbance events.A comprehensive analysis of the complex disturbance event evolution and their internal relationship not only provides a basis for analyzing the event consequence,investigating the incident responsibility,and controlling the PQ pollution,but also is beneficial to improve PQ analysis level of monitoring system.In this paper,a comprehensive analysis is given that includes complex PQ disturbance event source location,complex disturbance pattern mining,and complex disturbance event diagnosis,the details are as follows.Firstly,the concept of complex PQ disturbance model is proposed to establish methods for mining possible complex perturbation patterns in power grid.According to the formation mechanism of perturbation,four basic event association types are given.A PQ disturbance extension model is established.On the basis of the extension analysis method of Extenics,four kinds of methods for complex disturbance patterns mining including event development network,perturbed conduction implication system,causal chain and spatio temporal hierarchical deductive graph are constructed,as well as the mining steps of complex perturbation patterns are given.Moreover,the direction location methods of complex PQ disturbance event source are proposed.Based on the formation mechanism and the temporal and spatial characteristics of complex disturbance events,the reason that causes the difference of quantities and locations of disturbance sources of complex disturbance events in different periods is analyzed.The particle swarm optimization algorithm is used for time-frequency atom decomposition algorithm to achieve the detection on disturbance signals.According to the start and stop time of each basic event,disturbance interval division method is established.The components of three-phase instantaneous power during the disturbance occurs are analyzed based on the superposition theorem.A direction location method of complex disturbance event sources based on disturbance power is proposed and the direction location flow of disturbance event sources in different time intervals is given.The numerical simulation verification is conducted.Furthermore,a location method of complex PQ disturbance source is studied on the basis of the disturbance direction location.A framework for the section location of complex PQ disturbance event source is constructed.A method for determining the branch of suspicious disturbance event sources based on power transmission access is established.A measurement concept is introduced in the initial logical state representation of proposition,the measured weighted fuzzy Petri net ten tuple is defined,and the location model of perturbed event source weighted fuzzy Petri nets is established.The state measurement is used to represent the credibility degree of disturbance direction of various monitoring devices.A method for state measurement extraction is established.A method of parameter determination of Petri network is given according to the accuracy of disturbance information at monitoring positions.The numerical simulation verification is conducted.Finally,the synthetic diagnosis method of complex PQ based on complex event processing(CEP)is studied.A hierarchical polymerization principle is used to construct a hierarchical complex PQ comprehensive diagnosis system model.PAPIDE-EPL description language is used to model the underlying data layer,vertical and horizontal layer,and integrated diagnosis layer.The disturbance characteristic information and topology information are used to model the basic events in the underlying data layer.The correlated characteristic of basic events is used to model the vertical and horizontal correlation layer events.The communalism and the development characteristic of basic events included by complex events of vertical and horizontal correlation layer are used to model integrated diagnosis layer events.According to the mode information base of complex disturbances,the characteristic and the association relationship of various layer events is used to design the aggregate matching rule of vertical and horizontal correlation layer events and integrated diagnosis layer events.The numerical simulation verification is conducted.