摘要
由一名新闻记者-机器人与机器学习的工作人员新闻编辑-每日新闻-调查人员发布了关于人工智能的新报告。根据NewsRx编辑从泰国曼谷发来的新闻报道，这项研究称，“热裂解炉是石油化工行业的一种重要设备，通常用于将重烃裂解成短链，并作为副产品生产焦炭。生成的焦炭的沉积会提高外盘管壁的温度，需要定期维修炉以防止盘管失效。”这项研究的资助者包括石化和Mate Rials技术卓越中心；Kasetsart大学工程学院。为此，本文提出了一种基于后验特征的机器学习方法，该方法包括一个两级机器学习模型，该模型旨在提高预测精度，降低特征敏感性，并将标签分类为一个星期范围的标签。该模型将传感器特征提取到后验概率类Labe L分数中,然后对这些分数进行处理和分类,为第二级模型生成特征,结果表明,该模型能够提取过程变化,识别服务需求.与传统分类模型相比,该模型对洁净数据集和焦炭污染数据集的分类精度分别提高了23.94%和17.67%。
Abstract
By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News-Investigators publish new report on ar tificial intelligence. According to news originating from Bangkok, Thailand, by NewsRx editors, the research stated, "A thermal cracking furnace is an important equipment in the petrochemical industry that is typically used for breaking lon g hydrocarbons into short chains and producing coke as a byproduct. Deposition o f the generated coke increases the temperature at the outside coil wall, necessi tating regular furnace maintenance to prevent coil failure." Funders for this research include Center of Excellence on Petrochemical And Mate rials Technology; Faculty of Engineering, Kasetsart University. Our news reporters obtained a quote from the research from Kasetsart University: "Therefore, this study proposed a machine learning approach with a posteriori-b ased feature to predict the service life of the furnace to runtime failure. The proposed approach consists of a two-level machine learning model, which aims to improve prediction accuracy and reduce feature sensitivity. The label is classif ied as a week-range label, which can be categorized by classification criteria i nto three classes: weekly, bi-weekly, and quarter-weekly. The first-level model is utilized to extract sensor features into the posterior probability class labe l score. These scores are then processed and sorted into moving windows to gener ate features for the second-level model. The results showed that the proposed mo del could extract process variation and identify service needs, which improved c lassification accuracy by 23.94 % and 17.67 % for th e clean and coke-contaminated datasets compared to the conventional classificati on model, respectively."
NETL
NSTL