Research on the determination method of key events in process system fault evolution
The purpose of this paper is to study a method for determining key events in the system fault evolution process,and to analyze the objects measured at various times during this process.A basic data matrix is established consisting of two dimensions:factors and objects.The set of neighboring objects of each particular object needs to be considered to determine a characteristic object.The event that occurs when the characteristic object is measured and can be defined as the key event.With this consideration in mind,a key event determination method based on Kernel Local Preserving Projections(KLPP)is proposed.The system fault evolution process,key events,and factor space are then discussed,after which a key event determination method is proposed.Finally,an example is analyzed.This research shows that the system fault evolution process has a complex structure and hierarchy:the experienced event is the object of evolution measurement,while the key event is the decisive one.Key events are the basis for describing a process of evolution and can be determined by a distribution of objects in the factor space.Through the KLPP method,the distribution characteristics of objects can be studied,as well as the characteristic objects obtained by analyzing the nearest neighboring objects.The experienced events corresponding to these characteristic objects are defined as key events.The requirement is to arrange the characteristic objects in ascending order based on measurement time,serving as nodes for a space fault network that describes the evolution process.The use of electrical systems as an example is influenced by six factors.The evolution process measures 100 objects and a process of algorithm analysis obtained 16 characteristic values and characteristic object sets.The key events of the evolution process are then obtained by arranging the characteristic objects in the corresponding measurement order.Finally,the research significance and characteristics of this method are presented.
basic discipline of safety science and technologysystem faultevolution processkey eventscharacteristic objectKernel Local Preserving Projections(KLPP)
万方数据
维普
