Alarm Flood Analysis And Management Of Large Industrial Process

The alarm system is an important part of the modern industrial production monitoring system and plays an important role in ensuring the safe,stable and efficient operation of industrial production.With the widespread application of computer technology in industrial control systems,the setting of alarm variables has become easier and the alarm status set in the system has increased significantly.If not handled properly,the alarm system may generate a large number of alarms in a short period of time.The operator cannot process all alarm information in a limited time,and some key faults cannot be handled in time,resulting in serious safety production accidents.In response to the above problems,it is necessary to develop an intelligent alarm management strategy based on the analysis of historical alarm data.The analysis and processing of alarm flood is an important research content in large-scale industrial process.The root cause of alarm flood is that the same fault triggers the alarm state of multiple variables,which forms the propagation of abnormal states between devices.The time information of the alarm flood sequence needs to be discovered and processed to find the propagation path of the abnormal fault.Alarm flood analysis can be divided into two parts,offline and online.The main purpose of the offline part is to find out the sequence template of the alarm flood,and the main purpose of the online part is to perform online matching of the alarm flood and dynamically manage the alarm information.The work of this paper focuses on the offline analysis of the alarm flood.After clustering the historical alarm flood sequence,the pattern within the cluster is mined to obtain the alarm template of the current cluster.The innovation of this thesis is to propose a multi-sequence alignment algorithm based on alarm priority.This algorithm uses the dynamic programming idea to align multiple alarm sequences to obtain the template sequence of this kind of alarm flood.Compared with existing algorithms,this algorithm adds the important information of alarm priority,so that it is more sensitive to high-priority alarms.In addition,this paper also proposes an improved progressive multiple sequence alignment algorithm,which mainly includes three steps of generating distance matrix,creating guide tree and progressive alignment.The algorithm can reduce the calculation time complexity and shorten the calculation time.The algorithm can be regarded as an effective supplement to the algorithm based on the idea of dynamic programming.