Methods Of Risk Data Knowledge Discovery Of Railway Freight Operation

With the rapid development of China’s economy,the continuous innovation and reform of the railway transport industry have been promoted,and the technical complexity has been strengthened.The existing safety management model is facing severe challenges correspondingly.The traditional risk management method of railway freight transport operation is generally based on a single linear relationship between the risk sources and the accidents,to effectively identify and control the risk sources of accidents that have occurred.But it is impossible to identify and correlate potential risk sources,or to accurately manage the risks in the rail transportation.The method satisfies the intelligent service of railway freight operation risk management,and realizes the value-added demand of informatization and intellectualization through real-time and dynamic data mining.Therefore,it is of great significance to study the knowledge-based method of railway freight operation risk data for further accurately mining risk sources and improving the level of railway freight operation risk management.Aiming at the characteristics of railway freight operation risk accidents,this dissertation probes into the knowledge-based problem of structured and unstructured data of railway freight operation risk based on ontology,rough set and artificial neural network,establishes knowledge base of risk data knowledge and scenario integration,and provides knowledge-based and management of railway freight operation risk data.It provides a reference for the knowledge and management of railway freight operation risk data.The main contents of this paper are as follows.(1)Establishment of index system for influencing factors of railway freight transport risk and screening of the key influencing factorsFrom the perspectives of human,equipment,environment and management,the index system of influencing factors for the railway freight operation risk is established.The hazard source,fault and accident data are selected as conditional attribute sets,and corresponding risk grade information is taken as decision attribute sets.A knowledge system of influencing factors of railway freight operation risk based on human,equipment,environment and management is constructed and applied.The rough set theory is used to screen the key risk factors affecting railway freight transport safety,calculate and compare the weight differences of the key risk factors.(2)Ontology construction and association pattern recognition of railway freight operation risk managementBased on the unstructured data of railway accident cases,the accident is described and analyzed,the railway freight operation risk sources are excavated,and the cause mechanism is analyzed.Several typical accident cases with complex causes and high accident levels are selected to complete the knowledge extraction of scenario cases,and an accident-risk ontology model based on accident scenario is established.The ontology of freight operation risk is formalized to identify the risk association model,and a knowledge-based method of railway unstructured data is proposed.(3)A knowledge-based method for risk management data of railway freight transport equipment is proposedIn view of the current situation of knowledge-based structured data for railway equipment condition detection,a knowledge-based method of railway freight equipment data based on neural network is proposed,and the track irregularity is taken as an example to verify it.The BP neural network is used to predict the cycle of the track.At the same time,the validity of the model is validated by seven test data,such as gauge,left track,right track,left high and low,right high and low,triangle pit and level.(4)Intellectualization of railway freight operation risk data and integration of the scenario knowledgeTo map the safety influence index system of railway and the related concept of risk source in the accident risk ontology correspondingly,and intellectualize the weight of the impact index of the railway freight transportation.To establish 4 sets of mapping rules that impact the indicators of railway freight operation,which are personnel mapping rules,management mapping rules,environmental mapping rules and equipment mapping rules.Taking the data mining and calculation process as an example,updated the conceptual part of the risk accident ontology accordingly,and the new ontology concept is mapped to risk,equipment,infrastructure,fixed equipment,track and line,use the Protege tools to establish a new concept in the risk accident ontology and construct the new ontology map.According to the type of risk source,it can be divided into four categories:human,equipment,environment and management.According to the factors of the accident situation,the cause of the risk and the accident,the analysis of the group relationship between the risk sources is defined as three types of risk group relationship:upgrade,relevance and causality.Through the analysis of the accident risk source and the participation behavior in the accident scenario,the corresponding relationship between the two parties is searched and the two groups are linked up to form the dynamic reasoning chain from participation behavior to risk group,and the rules of extracting,producing,catalyzing and dealing with the abnormal behavior risk chain are established,and the logic table of reasoning mapping is formed.The main innovations of this paper are as follows:(1)The knowledgeable method of the railway freight operation risk data is put forward.Based on the analysis of the cause mechanism of accident scenario,the risk sources are identified and the association rules among the risk factors affecting the safety of railway freight transport operation are excavated.The accident-risk ontology model is constructed based on risk association knowledge,and the ontology knowledge reasoning and updating mechanism of railway freight operation risk is studied.(2)The railway freight operation risk ontology model is constructed.Based on the ontology-based structural description and reasoning method of railway freight operation risk model,the reasoning mechanism of railway freight operation risk association knowledge is excavated and put forward,and the risk identification method based on management data and equipment data is established.Using ontology description of railway freight operation accident scenario,ontology integration of different types of railway freight operation risk is carried out,and the intellectualization of railway freight operation risk data is realized.(3)A method of knowledge modeling and reasoning research for railway freight operation risk data is put forward.This dissertation establishes an index system of influencing factors of railway freight operation risk,builds a knowledge system of influencing factors of railway freight operation risk based on human,equipment,environment and management,and screens the key risk factors affecting railway freight transport safety by the rough set theory.Through the railway freight accident scenario analysis and modeling,the causes of each accident chain and accident chain are analyzed and decomposed,and the interaction between unstructured railway freight operation risk data and accident data is explored.