Answers are grounded in this paper and cite their sources.
The objective of this study is to create an effective prediction model that can quickly and accurately predict tunneling characteristics, enabling early-warning and intelligent performance monitoring of shield machine systems. Typically, an efficient predictive model usually depends on the characteristics of the data. However, the worldwide time-varying variations and inflection points in the data cannot be captured by current prediction methods. It is suggested that a prediction model incorporate change identification and Winsorization techniques to address the current problem. The proposed forecasting framework is constructed on integration of the stationary wavelet transform (SWT), the autoregressive moving average (ARMA) model, and the recursive least squares (RLS) algorithm. First, SWT is utilized to dissect cutterhead torque data, which effectively reduces the impact of outliers while obtaining global trend features. Second, the ARMA model, complemented by change detection and Winsorization approaches, is used to successfully identify time-varying variations and inflection points in cutterhead torque data. Finally, the efficacy of the proposed method across different prediction intervals is validated using cutterhead torque data examples. The findings indicate that the proposed method explains approximately 992% of the fluctuations in cutterhead torque. Across different prediction horizons, the average R 2 increases from 0.38 to 0.90, while the average root mean square error and mean absolute error decrease by 41.15% and 33.79%, respectively.
No takes yet. Share an insight, caveat, or question.
Yuan et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: