Industrial process state monitoring method based on spectral radius-interval principal component analysis

Abstract

The invention relates to an industrial process state monitoring method based on spectral radius-interval principal component analysis, which comprises the following steps of: aiming at data which is acquired in an industrial process and contains measurement noise, measurement error and uncertainty, converting the data into a data conversion method based on kernel density estimation; converting process data collected in the industrial process into interval data; based on the obtained interval data, establishing a spectral radius-interval principal component analysis model, performing feature extraction on a complex industrial process containing inaccurate process data, and projecting high-dimensional interval data to a low-dimensional space; introducing offline monitoring statistics, and determining a control limit of the statistics based on a kernel density estimation method; and based on the obtained control limits of the four monitoring statistics, analyzing the relationship between the on-line monitoring statistics and the control limits, and realizing on-line monitoring of the process state.

Description

technical field [0001] The invention relates to the technical field of industrial process state monitoring, in particular to a complex industrial process state monitoring method with inaccurate measurement data. Background technique [0002] Since the 21st century, with the rapid development of science and technology and the global economy, as well as a new round of industrial revolution and a major shift in the paradigm of global industrial competition, the pace of intelligent manufacturing has been accelerating. In the tide of rapid global economic development, product quality and production safety have become the top priority for enterprises to gain a foothold and develop. Adopting the correct process status monitoring method can improve the safety of industrial equipment operation, prevent catastrophic accidents, and reduce product quality fluctuations, thereby improving the competitiveness of enterprises. With the improvement of science and technology and the populariz...

AI Technical Summary

Problems solved by technology

The addition of the above noises, errors and uncertainties will often lead to false positives and false positives in commonly used data-driven condition monitoring methods. In severe cases, it is even impossible to distinguish between normal working conditions and abnormal working conditions [4]

Scholars at home and abroad have made some explorations on the state monitoring of complex industrial processes with inaccurate measurement data, but there are still deficiencies in the following aspects: (1) Most of the current data-driven multivariate statistical process state monitoring methods are based on Single-valued process data containing measurement noise and measurement errors are used for condition monitoring [5], or process data in the form of intervals are used for condition monitoring [6], without considering the existence of single-valued data in industrial processes There is also the case of interval data; (2) In the current research on process state monitoring based on interval data, there are problems of large amount of calculation, complex calculation, and insufficient mining of internal information of the interval, such as the vertex principal component proposed by Cazes et al. Analysis method [7]; (3) Most of the existing methods of converting single-value data into interval data are based on the idea of ​​"packaging" [8][9], simply using interval domains for single-value data at adjacent time points Expressed in the form of , ignoring the important information of the original single-valued data and the intrinsic relationship of the original data attributes

Images