数据驱动的船体大部件关键尺寸偏差分析预测

Data-Driven Analysis and Prediction of Critical Dimensional Deviations of Large Hull Block

针对船体大部件建造过程中数据分析效率低、精度管控依赖工程经验的现状,研究一种数据驱动的关键尺寸偏差分析预测方法。该方法首先对建造过程中的关键尺寸偏差数据进行分析,利用I-MR控制图判断工序是否处于受控状态,并采用高斯过程回归构建关键尺寸偏差预测模型。以双层底分段建造过程为例,通过I-MR控制图监控生产过程中的异常波动,并验证高斯过程回归模型的有效性。因此,所提出的船体大部件关键尺寸偏差分析与预测方法,可为现场开展定量化的精度管理工作提供指导。

Due to the inefficient data analysis and the reliance on engineering expertise for accuracy control in the construction process of large hull blocks,a data-driven method for analyzing and predicting critical dimension deviations is developed.The method first analyzed the deviation data critical dimensions during the construction process and applied the I-MR control chart to determine whether the process is under control.Additionally,Gaussian process regression was used to build a model for predicting critical dimensional deviations.Taking the double-bottom block construction process as a case study,abnormal fluctuations in the construction process were monitored through I-MR control charts,and the effectiveness of the Gaussian process regression model was verified.Therefore,the proposed analysis and prediction method of critical dimension deviations in large hull blocks can provide valuable guidance for quantitative precision management work on site.