Nonlinear Process Detection Based On Recurrence Pattern

Nonlinear phenomena are ubiquitous in daily production and life,which is of great significance to the research of the Non-linear process detection method.Especially with the development of modern technology,the production process is advancing towards integration,intellectualization and complication,which puts forward higher requirements for the reliability of production process,so it becomes more and more important to study the detection method of production process.The development of computer storage and sensor technology makes the process data easy to be collected and stored,how to effectively excavate the hidden process information from the data,can not only satisfy the performance monitoring of the non-linear process,but also can predict the process state.Therefore,the research of the Non-linear process detection method based on data driving has become one of the hotspots in the present research.In this paper,the method of recurrence pattern is used to study the detection of the non-linear process,which provides a new idea for the process detection in various fields.Firstly,this paper introduces the current situation of the research of nonlinear process detection methods at home and abroad,and the method based on data driven is undoubtedly the mainstream,and the application of recurrence plot and recurrence quantification analysis in different fields is emphatically introduced.Secondly,the basic theory and structure meaning of recurrence plot and recurrence quantification analysis are analyzed.And this method is used to simulate and verify some typical signals of different types.The results show that the characteristics of the original signal can be reflected on the recurrence plot.Then,the important parameters of the three recurrence plot are studied,such as delay time,embedding dimension and threshold value,then the embedding dimension d is determined by the false nearest neighbor method,the time delay parameter ? is determined by the mutual information method,and the multiple of the standard deviation is selected as the threshold of the recurrence plot.Finally,this paper applies the nonlinear process detection method based on recurrence pattern to the concrete example.By collecting the pulse wave signal of the human body,the recurrence plot and the recurrence quantification analysis of the pulse wave signals in the static and motion state are presented.The results show that this method can be used to identify the different state of the human body through the pulse wave signal.