Tennessee Eastman Process
收藏arXiv2025-09-30 收录
下载链接:
https://data.dtu.dk/articles/dataset/tennessee_eastman_reference_data_for_fault-detection_and_decision_support_systems/13385936/1
下载链接
链接失效反馈官方服务:
资源简介:
该数据集包含了20种不同的注入故障(过程干扰),用于验证AITwin概念的有效性。在模拟过程中,由于仪器限制,可能无法准确识别所有故障。此外,由于描述不足,实验并未涵盖所有注入的故障。该数据集在多个实验中评估了20种不同的故障,其任务是对模拟过程中的异常进行检测以及故障的诊断。
This dataset contains 20 distinct injected faults (process disturbances) for validating the effectiveness of the AItwin concept. During the simulation process, due to instrument limitations, not all faults may be accurately identified. Furthermore, owing to inadequate descriptions, the experiments did not cover all the injected faults. This dataset is employed to evaluate the 20 distinct faults across multiple experiments, with the core tasks being anomaly detection and fault diagnosis in simulated processes.
搜集汇总
数据集介绍

背景与挑战
背景概述
该数据集是Tennessee Eastman Process的扩展参考数据,主要用于故障检测和决策支持系统研究。数据包含28种过程故障的500次模拟(每次使用不同随机种子),以及设定点变化和模式转换模拟,所有模拟持续100小时并以3分钟为采样间隔。
以上内容由遇见数据集搜集并总结生成



