Abstract
<jats:p>This study examines the application of artificial intelligence (AI) and neural network methods to support engineering calculations in the development of high-tech products. Digital twin models of polymer composite materials are analyzed, where the computational core is implemented as a regression machine learning model that predicts tensile strength based on technological and physicochemical parameters. The paper describes the composition of initial data, feature engineering, model selection, and validation using metrics such as Mean Absolute Error (MAE) and the Coefficient of Determination ($R^2$). Conditions for achieving practical results are discussed, alongside limitations related to sample representativeness, interpretability, and computational resources. The integration of statistical models with CAD and CAE methods reduces the volume of physical field testing and accelerates the design cycle</jats:p> <jats:p>考虑了人工智能方法和神经网络方法在高科技产品开发中支持工程计算的应用。 分析了聚合物复合材料的数字孪生模型 其中计算核心作为回归机器学习模型实现 该模型根据技术和物理化学参数预测拉伸强度。 描述了初始数据的组成 特征的准备 模型选择和使用平均绝对误差和确定系数的度量进行验证。 讨论了获得实际效果的条件以及与样本代表性 可解释性和计算资源相关的限制。 统计模型与CAD和CAE方法的集成减少了现场测试的体积 加快了设计周期。 关键词:人工智能,机器学习,神经网络,数字孪生,复合材料,回归,抗拉强度</jats:p>