Research On Integrated Case-based Reasoning For Response Planning In Power Grid Emergency

In recent years,extreme weather and geological disasters occur frequently,resulting in great threat to the security of the power grid in our country.The freeze disaster in southern China in early 2008,the big earthquake happened in Wenchuan in May 2008,the Typhoon HAIKUI happened in August 2012,the Typhoon Fitow in October 2013,the Typhoon Mujigae in October 2015 and the Typhoon Meranti in September 2016 resulted in transmission lines broken,tower damage and severe power off,causing a wide range of damages to the social and economic development and the security of peoples’ life and property.The research for power grid emergency management has been a vital problem to be solved urgently.Most of the research at the present is focused on the system construction of emergency management for power grid,the assessment for emergency capability,countermeasures,emergency platform architecture,special technology application and so on,paying less attention to emergency response and emergency rescue decision-making.The incidents of power grid are characterized by complex,urgency,severe damage and incomplete information,making the the traditional method of emergency management and decision-making failed.This paper aims to do the research of Case-Based Reasoning of emergency management for power grid oriented by scenario-response in the guidance of the emergency model of scenario-response.In order to provide decision-supporting for emergency response and rescue,this paper proposes a co-operation system prominent by Case-Based Reasoning by using theories,models and technologies from Case-Based Reasoning,Rule-Based Reasoning,genetic algorithm,grey correlation analysis,neural network,and D/S evidence theory,which has important theoretical value and practical significance.For desired purpose,the paper has carried out the following research:First of all,the research of Case-Based Reasoning system for power grid emergency management oriented by scenario-response has been done in the thesis paper.On the basis of the construction of emergency cases ontology model based on ABC ontology,the scenario features,emergency subject and emergency resources in the structure of power grid incidents are presented comprehensively.After the grasp of scenario features and response process,the demand,the process and the advantages of Case-Based Reasoning for power grid incidents are proposed.At the same time,the system structure of Case-Based Reasoning is designed and the critical technologies relating to the system structure are identified,laying a solid foundation for further research.Secondly,the case structured method for power grid incidents under thecondition of incomplete information is proposed.On the basis of case representation,the frame representation method for power grid emergency case under the condition of incomplete information is proposed.At the same time,the basic probability assignment from D/S device theory is introduced into frame representation in this method to meet the structured and quantitative requirements.Then,the retrieval method for power grid emergency cases under the condition of incomplete information is proposed.Finally,the case structured method under the condition of incomplete information is tested and the results show that this method can meet the structured and quantitative requirements of Case-Based Reasoning for power grid emergency incidents with incomplete information.Thirdly,the case adaptation method based on improved backward propagation neural network is proposed.One-output neural network with each scenario feature and all the decision features is built and all one-output neural networks are paralleled to form the parallel structure for case adaptation.At the same time,for the overcoming of randomness and tentative in the connection weights adjustment in the backward propagation neural network based on simulated annealing algorithm,the objective function is used to adjust learning speed automatically,avoiding network paralysis or in a local optimum.Next,the case base expanding method for power grid emergency case based on Rule-Based Reasoning and improved multi-object genetic algorithm is proposed.On the basis of case base building method based on similarity.In this method,Rule-Based Reasoning is responsible for the selection of the corresponding decision features to the scenario features in the power grid emergency cases;the improved multi-object genetic algorithm is designed to compute the similarity of the scenario features with similarity function and compute the correlation of the scenario features and the decision features,ensuring a high similarity and a high correlation.In order to overcome shortcomings of premature and slow convergence of the simple genetic algorithm,the multi-object genetic algorithm is improved by adopting improved adaptive genetic operators and population generation method based on the fitness.At the same time,D/S evidence theory is introduced into the algorithm,solving the problem of incomplete information calculation and enriching theories and methods for case base expansion and case adaptation.Finally,the system of Case-Based Reasoning for power grid emergency management is developed with VC + + 6.0,Microsoft Access2003 and Matlab2012 a and the practicability and the validity of key technologies,models and methods proposed in this paper are tested.