Scholarlink

<p>Water-lubricated journal bearings play a critical role in marine applications, where operational efficiency and environmental protection are essential. This study focuses on optimizing the design parameters of herringbone groove journal bearings (HGJBs) to improve their static performance. A Taguchi L<sub>25</sub> orthogonal array is employed to analyze the influence of groove depth, groove angle, and the number of grooves on performance parameters, including load-carrying capacity (LCC), maximum pressure&#xa0;(P<sub>max</sub>), frictional force&#xa0;(FF), frictional coefficient&#xa0;(FC), power loss&#xa0;(PL), leakage rate (LR), attitude angle (AA), and frictional torque (FT). Sensitivity analysis using ANOVA identifies the most significant factors influencing performance. To address the computational challenges of multi-objective optimization, Taguchi-based gray relational analysis (GRA) is integrated with an artificial neural network (ANN)-based surrogate model. The ANN predicts the gray relational grade with high accuracy, significantly reducing computational time using three error minimization techniques: mean squared error&#xa0;(MSE), mean squared error with regularization&#xa0;(MSEREG), and sum of squared errors&#xa0;(SSE). The optimized HGJB parameters demonstrate a 43.57% reduction in LR and a 15.68% reduction in FT compared to conventional plain journal bearings&#xa0;(PJBs), albeit at the cost of a reduction in LCC at a relative eccentricity ratio of 0.7. This makes them particularly suitable for environmentally sensitive marine applications where conserving fluids and reducing energy losses are a higher priority than maximizing LCC. The novelty lies in combining ANN with Taguchi-GRA for rapid and accurate performance prediction, providing a robust framework for optimizing water-lubricated HGJBs while ensuring sustainability and efficiency.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A novel unified framework for multi-objective optimization, sensitivity analysis, and ANN-based performance prediction of herringbone-grooved water-lubricated journal bearings

  • 期刊论文

  • 发表日期：

    2025年9月2日

查看全文
  • Sumit Kumar Ohdar,
  • Suraj Kumar Behera

摘要

Water-lubricated journal bearings play a critical role in marine applications, where operational efficiency and environmental protection are essential. This study focuses on optimizing the design parameters of herringbone groove journal bearings (HGJBs) to improve their static performance. A Taguchi L25 orthogonal array is employed to analyze the influence of groove depth, groove angle, and the number of grooves on performance parameters, including load-carrying capacity (LCC), maximum pressure (Pmax), frictional force (FF), frictional coefficient (FC), power loss (PL), leakage rate (LR), attitude angle (AA), and frictional torque (FT). Sensitivity analysis using ANOVA identifies the most significant factors influencing performance. To address the computational challenges of multi-objective optimization, Taguchi-based gray relational analysis (GRA) is integrated with an artificial neural network (ANN)-based surrogate model. The ANN predicts the gray relational grade with high accuracy, significantly reducing computational time using three error minimization techniques: mean squared error (MSE), mean squared error with regularization (MSEREG), and sum of squared errors (SSE). The optimized HGJB parameters demonstrate a 43.57% reduction in LR and a 15.68% reduction in FT compared to conventional plain journal bearings (PJBs), albeit at the cost of a reduction in LCC at a relative eccentricity ratio of 0.7. This makes them particularly suitable for environmentally sensitive marine applications where conserving fluids and reducing energy losses are a higher priority than maximizing LCC. The novelty lies in combining ANN with Taguchi-GRA for rapid and accurate performance prediction, providing a robust framework for optimizing water-lubricated HGJBs while ensuring sustainability and efficiency.