Under complex and harsh operating conditions, water-lubricated bearings inevitably experience a transition in lubrication state, leading to mixed lubrication regime. The micro-asperities contact occurs on journal-bearing interface. Lubrication, friction and wear characteristics are influenced by surface roughness and elastic deformation. A program to model water-lubricated bearing under mixed lubrication regime was developed. Effects of various working conditions, structural parameters, roughness peaks, and elastic deformation on static and dynamic characteristics were investigated. The influence of four interface conditions on lubrication performance was quantified. The obtained data could guide parameter optimization in subsequent work to mitigate operational risks and prevent adverse impacts, enhancing the reliability and durability of water-lubricated bearings.
Structure active vibration control is a significant problem in practical engineering and the placements of actuators and sensors is an essential step in the controller design. However, various indicators must be comprehensively considered when evaluating optimal placement schemes, making the judgement of placement schemes a multi-objective decision-making problem. Meanwhile there are uncertainties with insufficient data in practical engineering, leading to the fluctuations of these indicators. To solve these problems, a novel framework for multi-indicator optimal placement scheme judgement of the active vibration control system with interval uncertainties is proposed. The main novelties are that the interval uncertain entropy weights are used to measure the influence of all the indicators and TOPSIS is adopted to solve these problems. Taking the extensively used proportional-integral-differential (abbreviated as PID) controller as an example, firstly the controllability indicator, the observability indicator, the energy-consuming indicator, the safety indicator and distance indicator are selected to comprehensively judge the performance of the placement scheme. Then, the uncertainties are quantified with interval forms, and the boundaries of indicators are calculated with the vertex method. Next, considering the large fluctuation of the distance indicator, the uncertain information entropy weights of the first four indicators are calculated and TOPSIS considering interval uncertainties is adopted to obtain the closeness of the schemes. Finally, the closeness and distance indicator are weighted summed as the comprehensive indicator to judge the optimal placement scheme of the uncertain system. Two numerical examples demonstrate the practicability and rationality of the proposed framework.
Accurate identification of the stern pressure field holds significant engineering implications for controlling hull structure vibrations and propeller excitation effects. In order to address the accuracy defects of the traditional distributed load inversion method when applied to complex surface structures, a baseline neural network model has been introduced and improved. Furthermore, a method of intelligent identification of stern pressure based on the Bayesian optimization algorithm and the attention mechanism of the CNN- BiLSTM model has been proposed. Taking the KCS container ship as the research object, we constructed a multi-physical-field coupled dataset containing 65 stern pressure measurement points and 9 structural strain measurement points through CFD simulations under five propeller rotational speed conditions. CNN and BiLSTM are used to extract spatio-temporal features from the dataset. Dynamic feature weight allocation is achieved through an attention mechanism, while Bayesian optimisation determines hyperparameter values to reduce bias. Results demonstrate that compared with basic models (CNN-BiLSTM, GRU, etc.), BOA-CNN-BiLSTM achieves optimal performance across MAE, MSE, RMSE, and MAPE metrics, with R2 reaching 0.987. This method achieves high-precision and high-reliability reconstruction of multi-point pressure data at the stern through finite strain monitoring, providing an effective solution for distributed load identification in complex curved structures.
The safety and stability of dock mooring systems for Floating Production Storage and Offloading (FPSO) vessels are critical, as they directly impact dock structural integrity and cargo loading/unloading operations. However, existing research on mooring dynamics in complex marine environments faces several challenges, including inadequate multi-periodicity modeling, neglect of transient impact loads, and limited data fidelity. These challenges mainly arise from the scarcity of long-term field measurements and the uncertainty of environmental conditions, rather than from the limitations of numerical modeling itself. In this study, a high-fidelity dynamic mooring model based on ANSYS-AQWA—calibrated against the real configuration of the terminal project—is employed to generate reliable reference data for training and benchmarking of machine learning models. To address these issues, this study proposes a prediction model based on TimesNet, designed to explicitly capture the multi-periodic characteristics and dynamic coupling effects of mooring tension. A fully coupled numerical model of the FPSO dock mooring system is developed using ANSYS-AQWA (Advanced Quadrature Wave Analysis), generating a high-fidelity dataset that accounts for the combined effects of wind, waves, and currents. The prediction performance of six mainstream models—including Long Short-Term Memory (LSTM), Bidirectional Long Short-Term Memory (BiLSTM), and Complete Ensemble Empirical Mode Decomposition with Adaptive Noise-Convolutional Neural Networks-LSTM-Attention (CEEMDAN-CNN-LSTM-AT) is systematically compared. Results indicate that the TimesNet model outperforms others in ultra-short-term prediction (with a time step of 9s), achieving an R2 value of 0.982 for bow mooring line tension forecasting. Under high-tide conditions, increased mooring line tension leads to greater nonlinearity and fluctuation, resulting in higher prediction errors compared to low-tide conditions. Despite larger load fluctuations in single cables within the operational mooring layout, the model maintains high predictive accuracy. The findings of this study provide theoretical support for the adaptive control of FPSO dock mooring winches and preventive maintenance strategies while laying the groundwork for the development of digital twin systems.
Ship navigation in polar regions faces significant challenges due to sea ice, which not only increases navigation resistance but also poses risks of structural damage and vessel immobilization. In response to these challenges, various route planning methods integrating ship transit models with optimization algorithms have been proposed. However, few studies have addressed multi-objective Pareto-optimal route planning that accounts for the effects of drifting sea ice and time-varying environmental conditions on vessel navigation performance. This study proposes a comprehensive multi-objective Pareto-optimal route planning framework specifically designed for polar navigation. The framework integrates a physically grounded ship transit model-incorporating the influence of sea ice drift-with an iterative cost assessment algorithm. Four objectives, namely navigation distance, navigation time, fuel consumption, and navigation risk, are jointly optimized. To identify Pareto-optimal solutions across these conflicting objectives, the NSGA-III algorithm is employed and enhanced with an A*-based initialization strategy to improve convergence efficiency. The proposed method is validated through a case study involving the Xuelong 2 icebreaker navigating from Shanghai to the Arctic. The results demonstrate that the optimized route effectively balances the four competing objectives through sophisticated multi-objective tradeoff analysis. This study contributes a robust and extensible methodological framework for polar navigation route planning.