Normal and abnormal data partitioning method and system for multivariable alarming system

Abstract

The invention relates to normal and abnormal data partitioning method and system for a multivariable alarming system. The method comprises steps: the minimum duration and the minimum deviation are selected, and a variation direction combination relation matrix of variables under a normal condition is constructed; time sequences of variables in data collection are standardized to obtain a standardized matrix; a time dimension is added to the standardized data matrix, primary clustering is carried out in the time dimension, the influence that local mutation caused by strong noise is recognized as a local trend can be eliminated, and the noise can be effectively reduced; and clusters obtained through the primary clustering are merged to recognize adjacent similar data segments; and secondaryclustering is carried out in a variable dimension, and similar data segments at a far time interval can be recognized. The error report rate can be greatly reduced while a high calculation speed is kept. Compared with the traditional method, the method and the system have huge advantages in recognizing the stage and the variation trend of the system.

Description

technical field [0001] The invention relates to the technical field of industrial alarms, in particular to a method and system for dividing normal and abnormal data of a multivariable alarm system. Background technique [0002] The efficient operation of many types of industrial alarm systems relies on a large amount of rich and labeled (such as "normal", "abnormal") historical data. Manually selecting such historical data from massive industrial data is inefficient and only applicable for smaller datasets. Therefore, it is particularly important to develop an algorithm that automatically divides normal / abnormal data segments. The qualitative trend analysis (QTA) method in the existing literature is one way to solve it. It contains three main methods, namely sliding window method (sliding windows), top-down method (top-down) and bottom-up method (bottom-up). However, there are still two unresolved issues: [0003] 1. The data segmentation method should not only be able t...

AI Technical Summary

Problems solved by technology

[0003] 1. The data segmentation method should not only be able to identify adjacent similar data segments, but also should be able to identify similar data segments that are far apart in time; the methods in the prior art cannot meet the above requirements

[0004] 2. Most of the existing methods focus on the local fitting of the signal

Local mutations caused by strong noise are easily identified as local trends, so such methods may fail when the real signal has strong noise

Images