典型制造过程异构动态图数据集

钢铁制造过程异构动态图模型数据集包括数值仿真案例数据、TEP仿真案例数据以及热连轧数据三部分。每部分均提供构建异构动态图所需的训练数据集，并提供用于测试基于异构动态图的面向KPI（关键性能指标）的过程监控模型监控效果的测试集。数据集涵盖了不同时间段的生产数据，时间精度达到分钟级别，空间范围涉及整个钢铁制造过程中的各个生产环节。空间精度涵盖从原材料处理到最终成品的各个步骤，数据通过高精度传感器和实时监控系统采集，保证了较高的空间精度与时间同步性。计算方式采用基于动态仿真模型和数据驱动的方法，通过数值仿真与实际生产数据对比验证数据的合理性和准确性。数据集的质量控制包括定期的数据清洗、异常值检测与剔除以及对传感器精度的校验，确保数据的完整性和可靠性。该数据集的潜在利用价值体现在为钢铁制造过程中提供基于KPI的智能监控和优化方法，能够显著提高生产效率、降低能耗、提高产品质量，并为后续的智能制造技术研发和工业应用提供基础数据支持。

This dataset, the Heterogeneous Dynamic Graph Model Dataset for Steel Manufacturing Processes, comprises three components: numerical simulation case data, TEP (Tennessee Eastman Process) simulation case data, and hot rolling mill data. For each component, both the training dataset necessary for constructing heterogeneous dynamic graphs and the test dataset for evaluating the performance of KPI (Key Performance Indicator)-oriented process monitoring models built upon heterogeneous dynamic graphs are provided. The dataset covers production data across diverse time periods, with a minute-level temporal resolution, and its spatial scope covers all production stages throughout the entire steel manufacturing workflow. The spatial coverage spans all steps from raw material handling to final finished product manufacturing. The data is collected using high-precision sensors and real-time monitoring systems, which ensures high spatial accuracy and strict temporal synchronization. The data validation and calculation adopt a hybrid framework integrating dynamic simulation models and data-driven methodologies, and the rationality and accuracy of the dataset are verified by comparing numerical simulation outputs with actual production data. Quality control measures for the dataset include regular data cleaning, outlier detection and elimination, and sensor accuracy calibration, to guarantee data integrity and reliability. The potential applications of this dataset lie in providing KPI-based intelligent monitoring and optimization solutions for steel manufacturing processes, which can notably enhance production efficiency, reduce energy consumption, improve product quality, and offer fundamental data support for subsequent R&D of intelligent manufacturing technologies and their industrial implementations.

搜集汇总

背景与挑战

背景概述

该数据集为典型制造过程异构动态图数据集，专注于钢铁制造领域，包含数值仿真、TEP仿真和热连轧数据，用于构建异构动态图以支持基于关键性能指标（KPI）的过程监控模型。数据集通过高精度传感器采集，具有分钟级时间精度和覆盖全生产环节的空间范围，旨在提高生产效率、降低能耗并为智能制造研发提供基础数据支持。