Study on Hydrodynamic Derivative Identification of Ship Simplified Modular Model Based on Support Vector Regression (SVR)

ObjectivesIn recent years, system identification has demonstrated favorable outcomes in the mathematical modeling of hydrodynamic coefficients for whole-ship models. However, its application in modular models remains limited, even though these models play a unique and significant role in the broad-speed range prediction and control of ship motions. Methods This paper proposes a white-box modeling method based on Support Vector Regression (SVR) for MMG-type model. A data preprocessing strategy is introduced to enhance the effectiveness of the sample data.Further,introducing Lasso regression to select the most influential hydrodynamic coefficients and alleviate multicollinearity. Subsequently, a regression model for hydrodynamic derivatives identification is derived for MMG model. Data centralization and differencing method are employed to reconstruct the regression model, mitigating the impact of parameter drift on hydrodynamic derivatives identification errors. ResultsSimulation experiments demonstrate that the proposed method achieves high precision in identifying hydrodynamic coefficients. The established model exhibits favorable predictive capability and ?generalization performance.ConclusionsThe SVR algorithm successfully identifies hydrodynamic derivatives of the modular models.