Tennessee-Eastman Process Data

Tennessee Eastman Process (TEP) is a chemical process test simulation platform. The simulation model is widely used in the testing of process monitoring methods. The process flow of TEP includes five main modules: reactor, condenser, compressor, separator and stripper; There are 8 components, the products G, H and by-product F which are formed by the reaction of four reactants A, C, D, E and inert substance B. 53 variables are collected in TEP, including 41 measurement variables and 12 manipulation variables (XMV 1-12). The measured variables include 22 process variables (XMEAS 1-22) and 19 component variables (XMEAS 23-41). There are 21 fault models in TEP to simulate actual industrial process faults. The training data set consists of 960 normal samples, and the test data set consists of 960 samples, of which the first 160 are normal samples and the last 800 are fault samples. This data model can inject faults to generate data of different fault degrees.

田纳西伊斯曼过程（Tennessee Eastman Process, TEP）是一款化工过程测试仿真平台，其仿真模型被广泛应用于过程监测方法的测试验证工作中。TEP的工艺流程包含五大核心模块：反应器、冷凝器、压缩机、分离器与汽提塔；体系内共包含8种组分：由反应物A、C、D、E与惰性物质B反应生成的产物G、H以及副产物F。该平台共采集53个变量，其中包括41个测量变量与12个操纵变量（XMV 1-12）；测量变量涵盖22个过程变量（XMEAS 1-22）与19个组分变量（XMEAS 23-41）。TEP内置21种故障模型，用于模拟真实工业场景中的各类过程故障。其训练数据集包含960个正常样本，测试数据集则包含960个样本，其中前160个为正常样本，剩余800个为故障样本。该数据模型支持注入故障以生成不同故障程度的数据集。