基于频域分解和多测点联合的土压力计测值异常检验方法OA

In the identification of operational behavior of earth pressure cells and structural safety assessment,mea-surement gross errors often coexist with abrupt changes reflecting actual structural responses,thereby compromising the accuracy of safety evaluation.To address this issue,this paper proposes an anomaly detection method for earth pressure cell measurements based on frequency domain decomposition and multi-point joint analysis.Through decom-position and reconstruction in the frequency domain,frequency components representing long-term trends and short-term fluctuations are derived.Subsequently,by integrating correlation analysis and the Isolation Forest algorithm,joint anomaly detection and diagnosis across multiple measurement points are achieved.Case studies based on actual engineering monitoring data demonstrate that the proposed method can accurately distinguish between gross errors and abrupt changes caused by structural anomalies or environmental factors,with high detection accuracy.The method enhances the diagnostic capability of earth pressure cell measurements and improves structural safety monitor-ing.