Analysis of Offshore Pile–Soil Interaction Using Artificial Neural Network DOI Creative Commons
Peiyuan Lin, Kun Li, Xiangwei Yu

et al.

Journal of Marine Science and Engineering, Journal Year: 2025, Volume and Issue: 13(5), P. 986 - 986

Published: May 20, 2025

Offshore wind power is one of the primary forms utilizing marine green energy in China. Currently, near-shore predominantly employs monopile foundations, with designs typically being overly conservative, resulting high construction costs. Precise characterization interaction mechanisms between piles and surrounding soils crucial for foundation design optimization. Traditional p-y curve methods, simplified fitting functions, inadequately capture complex pile–soil behaviors, limiting predictive accuracy model uncertainty quantification. To address these challenges, this research collected 1852 empirical datasets offshore interactions, developing horizontal displacement prediction models using artificial neural network (ANN) expressions comprehensive statistical analysis. The constructed ANN demonstrates a simple structure satisfactory performance, achieving average error margins below 6% low to moderate dispersion (26%~45%). In contrast, traditional show 30%~50% biases substantial near 80%, while conventional finite element methods exhibit approximately 40% dispersion. By strictly characterizing probability cumulative function factors, provided reliability-based design. Through case verification, it demonstrated that ANN-based has significant advantages terms computational efficiency foundations.

Language: Английский

Analysis of Offshore Pile–Soil Interaction Using Artificial Neural Network DOI Creative Commons
Peiyuan Lin, Kun Li, Xiangwei Yu

et al.

Journal of Marine Science and Engineering, Journal Year: 2025, Volume and Issue: 13(5), P. 986 - 986

Published: May 20, 2025

Offshore wind power is one of the primary forms utilizing marine green energy in China. Currently, near-shore predominantly employs monopile foundations, with designs typically being overly conservative, resulting high construction costs. Precise characterization interaction mechanisms between piles and surrounding soils crucial for foundation design optimization. Traditional p-y curve methods, simplified fitting functions, inadequately capture complex pile–soil behaviors, limiting predictive accuracy model uncertainty quantification. To address these challenges, this research collected 1852 empirical datasets offshore interactions, developing horizontal displacement prediction models using artificial neural network (ANN) expressions comprehensive statistical analysis. The constructed ANN demonstrates a simple structure satisfactory performance, achieving average error margins below 6% low to moderate dispersion (26%~45%). In contrast, traditional show 30%~50% biases substantial near 80%, while conventional finite element methods exhibit approximately 40% dispersion. By strictly characterizing probability cumulative function factors, provided reliability-based design. Through case verification, it demonstrated that ANN-based has significant advantages terms computational efficiency foundations.

Language: Английский

Citations

0