基於船舶減阻之高速排水型單體船水翼設計優化

Abstract

本研究旨在優化高速排水型單體船的水翼設計，藉此減少船舶靜水航行總阻力。本研究目標船形為NPL開放船形，船長約42公尺，排水量為247.18噸；安裝於船體的水翼則使用NACA0015翼形。本研究選定弦長、展弦比，縱向與垂向安裝位置以及安裝角度為水翼設計參數，使用計算流體力學軟體STAR-CCM+預測船速23.7節（Fn=0.6）具水翼的靜水阻力和姿態，利用MATLAB軟體以阻力預測結果訓練深度神經網路模型。本研究使用的深度神經網路模型具五層隱藏層，各層分別有8、7、6、5和5個神經元，平均絕對百分比誤差值為0.40%。本研究接著使用深度神經網路模型預測水翼的優化設計參數，透過考慮目標阻力和設計參數範圍，深度神經網路模型會進行各參數的排列組合，找出符合目標阻力的最佳設計參數。計算流體力學和深度神經網路預測具優化水翼設計的船舶靜水航行總阻力相差3.0%。具優化設計水翼船舶的阻力較裸船減少24.9%。本研究雖然針對Fn=0.6進行阻力優化，在Fn=0.3到0.9，安裝優化水翼的NPL船形預測減少58.4%到113.9%的俯仰角變化，7.6%到77.1%的動態下沉量以及14.1%到34.5%的阻力值。

This study aims to obtain an optimized hydrofoil design for the calm-water resistance reduction of a high-speed displacement monohull vessel at Fn=0.6. The target hull form is the NPL hull form with length of about 42 meters and displacement of 247.18 tons. The NACA0015 airfoil is selected for the section design of the hydrofoil installed on the hull. The design parameters of the hydrofoil are the chord length, aspect ratio, longitudinal and vertical installation positions, and installation angle. The computational fluid dynamics (CFD) software STAR-CCM+ is used to predict the calm-water ship resistance with the hydrofoil installed at the ship speed of 23.7 knots (Fn=0.6). The CFD resistance prediction is adopted to train a deep neural network (DNN) model with the help of MATLAB. The proposed DNN model consists of five hidden layers with 8, 7, 6, 5, and 5 neurons, respectively, and has a mean absolute percentage error (MAPE) of 0.40%. The proposed DNN model is then used to predict the optimized design parameters of the hydrofoil. With the target resistance and the given range of design parameters, various parameter combinations are explored to identify the best design parameters that meet the target resistance. The CFD and DNN predictions of the ship resistance for the optimized hydrofoil design differ by 3.0%. After the introduction of the optimized hydrofoil design, the ship resistance is reduced by 24.9% when compared to the bare hull. Although the ship resistance is optimized for Fn=0.6, the optimized hydrofoil design is able to reduce the running trim angle by 58.4% to 113.9%, the dynamic sinkage by 7.6% to 77.1%, and the ship resistance by 14.1% to 34.5% for Fn=0.3 to 0.9.