Data for: The prediction of ship added resistance at the preliminary design stage by the use of an artificial neural network
收藏资源简介：
AddedResistance (MAC/WIN) program was developed for added resistance estimation on head waves by the use of an MLP artificial neural network. In order to create a reliable model, only experimental data determined through model tests measurements was used to train this neural network. The following model tests results of these vessels were used in the development of this software: • Van der stel (Gerritsma and Beukelman , 1972; Journee 1976), • S175 containership (Nakamura, 1976), • Series 60 CB=0.6, CB=0.65, CB=0.7, CB=0.75, CB=0.8, Strøm-Tejsen (1973), • WILS II containership (Söding et al., 2014), DTC containership (Sprenger et al. , 2015), • Cruise ship (Ley et al., 2014), Bulk carrier (Kadomatsu, 1988), • RoPax (Sprenger et al., 2015), • KVLCC2 tanker (Guo & Steen, 2010; Sadat-Hosseini et al., 2013), • KRISO containership (Simones et al. 2014). AddedResistance program could have practical application at the preliminary design stage. But this computer program might have two limitations. Firstly, the program may be inaccurate in the design of an innovative ship. This results from the data used to train the network which was measured on standard ship hulls. Secondly, the neural network was developed using the model test data with limited parameter ranges. Therefore, AddedResistance might only be used to estimate added wave resistance for ships with the following design characteristic ranges: • length between perpendiculars LBP from 90 m to 335 m, • breadth B from 16.25 m to 58 m, • draught d from 4.2 to 20.8 m, • block coefficient CB from 0.503 to 0.829, • Froude number Fn from 0.087 to 0.3, • LBP/B ratio from 5 to 7.5, • B/d ratio from 2.5 to 4.5. The application of this computer software to calculate ship resistance with design characteristics outside these ranges might be associated with the risk of low accuracy estimations. The total mean added resistance of sailing ship in irregular waves is calculated by the use of the wave energy spectrum. A common ITTC spectrum based on Bretschneider wave energy spectrum was used here to calculate the mean added resistance in irregular wave. Lazarus IDE software (Lazarus 2.0.0, 2019) with a Free Pascal compiler was used to develop the AddedResistnace program.
AddedResistance（MAC/WIN）程序是为利用多层感知器（MLP）人工神经网络开展迎浪附加阻力估算而开发的。为构建可靠的预测模型，本程序仅采用模型试验实测数据对该神经网络进行训练。本软件开发过程中所用的船舶模型试验结果如下： • 范德斯特尔船型（Gerritsma与Beukelman，1972；Journee，1976） • S175型集装箱船（Nakamura，1976） • 系列60型船（CB=0.6、CB=0.65、CB=0.7、CB=0.75、CB=0.8）（Strøm-Tejsen，1973） • WILS II型集装箱船（Söding等，2014）、DTC型集装箱船（Sprenger等，2015） • 邮轮（Ley等，2014）、散货船（Kadomatsu，1988） • 滚装客船（RoPax）（Sprenger等，2015） • KVLCC2型油轮（Guo与Steen，2010；Sadat-Hosseini等，2013） • 韩国造船海洋研究所（KRISO）集装箱船（Simones等，2014） 该程序可应用于船舶初步设计阶段，但存在两处局限性：其一，针对创新型船型的设计场景，程序可能存在精度不足的问题，根源在于训练网络所用数据均来自标准船型的实测结果；其二，本神经网络的训练数据参数范围有限。因此，AddedResistance程序仅可用于估算符合以下设计参数范围的船舶的附加波浪阻力： • 垂线间长LBP：90 m 至 335 m • 船宽B：16.25 m 至 58 m • 吃水d：4.2 至 20.8 m • 方形系数CB：0.503 至 0.829 • 弗劳德数Fn：0.087 至 0.3 • 垂线间长与船宽比LBP/B：5 至 7.5 • 船宽与吃水比B/d：2.5 至 4.5 若将本软件用于估算超出上述参数范围的船舶阻力，可能存在估算精度偏低的风险。不规则波中航行船舶的总平均附加阻力通过波浪能谱进行计算，本文采用基于布雷切施耐德波浪能谱（Bretschneider wave energy spectrum）的国际船池会议（ITTC）通用谱来计算不规则波中的平均附加阻力。本程序采用Lazarus IDE软件（Lazarus 2.0.0，2019）搭配Free Pascal编译器开发完成。
