基于数字孪生技术的复合参量计量研究进展

    Research Progress in Composite Parameters Metrology Based on Digital Twin Technology

    • 摘要: 针对大型基础设施结构在复杂环境下的复合几何参量计量难题,传统计量技术在可控环境、单一参数、离散点上的局限性,使其难以满足“测得全、测得准、溯源评价可控”的需求。将数字孪生技术引入计量领域可以有效解决复合参量动态连续稳定和智能化数据处理的问题,完善多维度、多参量、多环境耦合作用下的几何量计量技术,实现全样本、多场景的数字化计量校准。首先,以室内80 m标准基线为对象开展实验,建立标准尺度下多维参量数据库,构建热-力-流-固多场耦合机理模型,揭示复合参量虚实映射机制。其次,建立三维多体动力学数字化模型,并将机理模型、数据驱动模型、统计模型、经验知识等集成到数字化模型中,构建虚实融合的数字孪生体。然后,利用类深度支持向量数据描述方法构建多个复合参量的特征超球体,依据不同的概率密度分布划分特征空间中不同参量的对应区域,实现异常参量靶向识别和计量溯源。此外,基于计量无异常样本数据提取多维敏感特征向量,利用广义多核聚类分析方法自适应计算邻域距离,评估计量准确度等级。最后,将“算测融合、溯评一体”的数字孪生计量方法推广至室外1.2 km三维标准基线,为实现数字化精准计量提供可行的初步解决方案与应用思路。

       

      Abstract: To address the challenge of composite geometric parameter metrology for large infrastructure structures in complex environments, the limitations of traditional metrology technology—restricted to controlled environments, single parameters, and discrete points—make it difficult to meet the requirements of “comprehensive measurement, accurate measurement, and controllable traceability evaluation”. Introducing digital twin technology into the metrology field can effectively solve the problems of dynamic, continuous, and stable measurement of composite parameters and intelligent data processing, thereby enhancing geometric metrology under multi-dimensional, multi-parameter, and multi-environment coupling. This enables full-sample, multi-scenario digital metrological calibration. Firstly, measurement experiments were conducted using an indoor 80 m standard baseline to establish a multidimensional parameter database under standard scales. A thermo-mechanical-fluid-solid multi-field coupling mechanism model was constructed, revealing the virtual-physical mapping mechanism of composite parameters. Secondly, a three-dimensional multibody dynamics digital model was established. Mechanistic models, data-driven models, statistical models, and empirical knowledge were integrated to build a digital twin that fuses virtual and real aspects. Then, multiple characteristic hyperspheres of composite parameters were constructed using a deep support vector data description method. Based on different probability density distributions, the feature space was partitioned into corresponding regions for different parameters, achieving targeted identification of abnormal parameters and metrological traceability. Additionally, multidimensional sensitive feature vectors were extracted from anomaly-free metrological sample data. The generalized multi-kernel clustering analysis method was used to adaptively calculate neighborhood distances, evaluating the metrological accuracy level. Finally, the digital twin metrology method that integrates "calculation-measurement fusion and traceability-evaluation unity" is extended to a 1.2km three-dimensional outdoor standard baseline, providing a feasible preliminary solution and application approach for achieving digital precision metrology.

       

    /

    返回文章
    返回