检索结果-内蒙古大学图书馆

22nd World Congress of the International Federation of Automatic control (IFAC)

作者： Hauesser, Felix Himmel, Andreas Findeisen, Rolf Tech Univ Darmstadt Control & Cyber Phys Syst Lab Darmstadt Germany

ISBN: (纸本)9781713872344

Many industrial processes consist of a series of interconnected components. While the components are often individually controllable, the potentially complex interaction mechanisms limit the design options for a process-wide controller. Typical reasons are that parts of the component interactions are difficult to model or that the overall, detailed physical model is very complex. Due to these challenges, using model-based control approaches, such as model predictive control, often becomes infeasible. To this end, an adaptive two-degree of freedom control concept for reference tracking tasks is proposed. A data-driven surrogate model of the coupled system is trained, focusing on the reference tracking objective, and exploiting the model structure. This surrogate model is used in a two-degree of freedom control concept, splitting the tasks into reference tracking and disturbance rejection. The proposed control concept applies to a broad class of reference tracking processes subject to partially unknown system dynamics. It is tested exemplarily to control an interconnected three-tank system. Copyright (c) 2023 The Authors.

关键词： control of distributed systems Model predictive and optimization-based control machine learning and data analytics in process control Nonlinear process control

来源：评论

学校读者我要写书评

暂无评论

Design of Multi-Model Linear Inferential Sensors with SVM-based Switching Logic 22

Design of Multi-Model Linear Inferential Sensors with SVM-ba...

引用

22nd World Congress of the International Federation of Automatic control (IFAC)

作者： Mojto, Martin Lubusky, Karol Fikar, Miroslav Paulen, Radoslav Slovak Univ Technol Bratislava Fac Chem & Food Technol Bratislava Slovakia Slovnaft as Bratislava Slovakia

ISBN: (纸本)9781713872344

The data-based design of multi-model linear inferential (soft) sensors ( MIS) is studied. These promise increased prediction accuracy yet simplicity of the model structure and training. The state-of-the- art approach to the MIS design consists of three steps: 1) data labeling (establishing training subsets for individual models), 2) data classification (creating a switching logic for the models), and 3) training of individual models. There are two main issues with this concept as steps 2) & 3) are separate: (i) discontinuities can occur when switching between the models;(ii) data labeling disregards the quality of the resulting model. Our contribution aims at both the mentioned problems, where, for the problem (i), we introduce a novel support vector method (SVM)- based model training coupled with switching logic identification and, for the problem (ii), we propose a direct optimization of data labeling. The proposed methodology and its benefits are illustrated on an example from the chemical engineering domain.

关键词： machine learning and data analytics in process control Monitoring and performance assessment Monitoring of product quality and control performance

来源：评论

学校读者我要写书评

暂无评论

A Log-Likelihood-Based Evaluation Metric for the Reproducibility and Simplicity of Logistics Graphs 22

A Log-Likelihood-Based Evaluation Metric for the Reproducibi...

引用

22nd World Congress of the International Federation of Automatic control (IFAC)

作者： Takakura, Yuriko Mori, Junichi Kobayashi, Hirokazu Nippon Steel Corp Res & Dev Proc Res Labs Chiba Japan

ISBN: (纸本)9781713872344

Logistics are sometimes complex and usually entail the interrelation of different processes. It is beneficial to visualize actual process-flow logs to better understand the underlying processes. However, it is difficult to analyze logistics using graph visualization as the graphs are typically quite large and complex. In the field of process mining, several metrics have been proposed in previous studies to examine the quality of process models created using process mining algorithms. However, these metrics evaluate the efficacy of the graph in terms of reproducing the actual process-flow logs. Thus, there is often process-flow information loss in the resulting graph structure when nodes or edges are aggregated to simplify the graph. To address these limitations, in this paper, we propose a maximum log-likelihood-based metric for measuring the reproducibility of graphs and define the concept in terms of how much of the actual process-flow information the graph retains. To obtain the metrics, we developed a graph model that can generate process-flow logs according to its probability parameters. In addition, we also developed an evaluation metric, which is a weighted sum of both the log-likelihood and model dimensions. An empirical evaluation was conducted using the actual process-flow patterns of steel-making process data. The results revealed that the maximum log-likelihood-based metric effectively evaluated the process by which the graphs with node and edge aggregations reproduced the actual process-flow patterns. Copyright (c) 2023 The Authors. This is an open access article under the CC BY-NC-ND license (https://***/licenses/by-nc-nd/4.0/)

关键词： machine learning and data analytics in process control Digital twins for power and process systems data visualization Logistics in manufacturing Production & logistics over manufacturing networking

来源：评论

学校读者我要写书评

暂无评论

A Log-Likelihood-Based Evaluation Metric for the Reproducibility and Simplicity of Logistics Graphs

引用

IFAC-PapersOnLine 2023年第2期56卷 6174-6180页

作者： Yuriko Takakura Junichi Mori Hirokazu Kobayashi Nippon Steel Corporation Research & Development Process Research Laboratories Chiba Japan

来源：评论

学校读者我要写书评

暂无评论

Design of Multi-Model Linear Inferential Sensors with SVM-based Switching Logic

引用

IFAC-PapersOnLine 2023年第2期56卷 2305-2310页

作者： Martin Mojto Karol Ľubušký Miroslav Fikar Radoslav Paulen Faculty of Chemical and Food Technology Slovak University of Technology in Bratislava Bratislava Slovakia Slovnaft a.s. Bratislava Slovakia

The data-based design of multi-model linear inferential (soft) sensors (MIS) is studied. These promise increased prediction accuracy yet simplicity of the model structure and training. The state-of-the-art approach to the MIS design consists of three steps: 1) data labeling (establishing training subsets for individual models), 2) data classification (creating a switching logic for the models), and 3) training of individual models. There are two main issues with this concept as steps 2) &3) are separate: (i) discontinuities can occur when switching between the models; (ii) data labeling disregards the quality of the resulting model. Our contribution aims at both the mentioned problems, where, for the problem (i), we introduce a novel support vector method (SVM)-based model training coupled with switching logic identification and, for the problem (ii), we propose a direct optimization of data labeling. The proposed methodology and its benefits are illustrated on an example from the chemical engineering domain.

关键词： machine learning and data analytics in process control Monitoring and performance assessment Monitoring of product quality and control performance

来源：评论

学校读者我要写书评

暂无评论

data-driven adaptive two-degree of freedom control of interconnected systems for reference tracking

引用

IFAC-PapersOnLine 2023年第2期56卷 2713-2718页

作者： Felix Häusser Andreas Himmel Rolf Findeisen

关键词： control of distributed systems Model predictive and optimization-based control machine learning and data analytics in process control Nonlinear process control

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：