Three-section advanced online identification method for dam safety monitoring data exception

Abstract

The invention discloses a three-segment advanced online identification method for dam safety monitoring data exception, which comprises the following steps: 1) data exception identification: dividingdata types, constructing a data identification model cluster, matching the data types with identification models; 2) measurement error reduction, including single accidental error reduction and instrument troubleshooting; and 3) environment quantity change induced abnormal change reduction, including data sequence construction for reducing the environment quantity, data sequence early warning threshold construction for reducing the environment quantity and environment response identification. According to the invention, the problems of misjudgment of normal measured values and missed judgmentof abnormal measured values easily occurring in a traditional single data anomaly identification method are solved; and meanwhile, classified online identification of non-structural variation inducedby accidental errors, monitoring instrument faults, environmental quantity changes and the like and structural variation induced by structural state deterioration is realized.

Description

Technical field [0001] The invention relates to the field of dam safety monitoring, in particular to a three-stage advanced online identification method for abnormal dam safety monitoring data. Background technique [0002] At present, there are many methods to identify anomalies in dam safety monitoring data, including Rait criterion method, statistical regression model, catastrophe theory, and fuzzy cluster analysis. The statistical regression model method based on the Rait criterion is the most commonly used in online identification of abnormal dam safety monitoring data because it can comprehensively reflect the impact of environmental variables, convenient calculation, low programming difficulty, and high reliability. This method uses the Rait criterion for the residual sequence to set an abnormal early warning threshold. For data sequences with large sample size, normal distribution, and moderate magnitude, the effect of online abnormality recognition is very good, but it i...

AI Technical Summary

Problems solved by technology

Traditional methods cannot identify the causes of data anomalies online, and cannot classify and identify non-structural changes induced by monitoring instrument failures, changes in environmental quantities, etc., and structural changes induced by deterioration of structural behavior. Screening and abnormal identification have low recognition and poor timeliness, which is not conducive to real-time monitoring and evaluation of dam safety performance

Images