Batch process fault detection method based on sample timing sequence and neighborhood similarity information

Abstract

The invention relates to a batch process fault detection method based on the sample timing sequence and the neighborhood similarity information. The method comprises steps that the training sample data is acquired; clustering of the training sample data is carried out; training samples after clustering are integrated to acquire different sub stages; a T2 statistical magnitude and an SPE statistical magnitude of each sub stage and a control limit of the two are calculated; the fault sample data is acquired, which sub stage of the training samples the fault sample belongs to is determined; a t2 statistical magnitude and an spe statistical magnitude of the fault sample are calculated; whether the statistical magnitudes of the calculated fault sample exceed the control limit is determined, if the two statistical magnitudes do not exceed the control limit, no fault is generated, otherwise, a fault is generated. Compared with the prior art, the method is advantaged in that accuracy and detection performance are high, the method is in accord with the practical situation, and the false alarm rate and the alarm missing rate are reduced.

Description

technical field [0001] The invention relates to the field of modern industrial manufacturing, in particular to a batch process fault detection method based on sample timing and neighbor similarity information. Background technique [0002] Batch processes play an increasingly important role in modern industrial manufacturing. In order to ensure the safety of batch process and the quality of products, the fault detection of batch process has attracted more and more attention of researchers. Due to the non-linear and multi-stage nature of the batch process, fault detection for the batch becomes more difficult. The application of multivariate statistical methods, such as Principal Component Analysis (PCA) and Partial Least Squares (PLS), has brought a new direction for fault detection in batch processes. How to extract useful information from a large amount of data in the production process, so that the fault detection based on multivariate statistical methods has higher accu...

AI Technical Summary

Problems solved by technology

However, this kind of clustering method has two main shortcomings in the clustering process. One is that it does not take into account the timing information of the production process, which will cause disorder in the clustering results; the other is that it does not make full use of the similarity between adjacent samples. Sexual information will make the clustering results inaccurate

After clustering, the data in the sub-stage cannot correctly reflect the characteristics that the stage should have, which makes the detection results unsatisfactory

Images