Development Test Platform For Fault Diagnosis Module Of Active Distribution Network

More than 80% of the faults in the power system originate from the distribution network.The rapid diagnosis,isolation and recovery of distribution network faults are critical to ensuring electrical energy supply and the safety of people and equipment.Therefore,a large number of distribution network fault diagnosis algorithms have been researched and applied.A large number of distributed generations are connected to the distribution network,which significantly changes the flow and fault characteristics.The effectiveness of traditional fault diagnosis algorithms in the distribution network and newly developed fault diagnosis algorithms in the active distribution network require systematic testing by various scenarios.At present,various test methods for fault diagnosis algorithms of distribution networks can meet requirements to varying degrees,but the methods are relatively independent and the fault scenarios covered are different.There is no comprehensive and unified testing scheme for fault diagnosis algorithms of the distribution network.Therefore,this thesis designs and implements a development test platform for fault diagnosis module of active distribution network.The information model for active distribution network fault scenarios and batch generation method for fault scenarios are proposed.The specific work and research content are as follows:Design and implementation of development test platform for fault diagnosis module of active distribution network.The test platform is mainly divided into three phases:data preparation phase,fault data processing phase,and fault diagnosis algorithm evaluation phase.The large-scale fault scenario datasets are used to test and evaluate the correctness and accuracy of the fault diagnosis algorithm under test.The test platform is designed to meet four functional requirements:standardized fault data formats,batch generation of fault scenario data,automated testing process,and modularity and scalability of fault diagnosis algorithms.The information model for active distribution network fault scenarios.Power,network and load together constitute the information model for active distribution network fault scenarios.Probability models and statistical characteristics are used to describe the uncertainties of variables such as distributed power output,line faults,and load fluctuations.The joint distribution model considering the correlation of multidimensional variables is discussed.Through the probability model of each variable and the joint distribution model of multi-dimensional variables,the failure scenario simulation dataset is sampled,which provides the basis for generating the failure scenario dataset of the test platform.Batch generation method for active distribution network fault scenarios.Call the PSCAD automation library through Python scripts to achieve model control and parameter modification.Large-scale fault simulation scenario sets are generated in batches,avoiding the tedious process of manually modifying model parameters,which greatly improves the efficiency of generating fault scenarios.To shorten the total time of PSCAD simulation,a method for improving the efficiency of PSCAD based on parallel computing is introduced.In the analysis and verification part of the test case,the fault scenarios dataset obtained from a variety of sources such as numerical simulation,physics laboratory simulation,and field tests is introduced.Use MATLAB Coder to generate code and compile the fault diagnosis algorithm module.The effectiveness of the fault diagnosis algorithm module is verified based on RTDS,which prepares for its application to actual power systems.