检索结果-内蒙古大学图书馆

Towards Integrated process Supervision: Current Status and Future Directions

IFAC Proceedings Volumes 1994年第10期27卷 9-21页

作者： Venkat Venkatasubramanian Laboratory for Intelligent Process Systems School of Chemical Engineering Purdue University W. Lafayette IN 47907 USA

process supervision deals with tasks that are executed to operate a process plant safely and economically. These tasks can be classified as data acquisition. regulatory control, monitoring. data reconciliation, fault diagnosis, supervisory control, scheduling and planning. Whi1c these operational tasks may be intrinsically different from each other, they are, however, c1osely related and can not be treated in isolation. Hence, there exists a clear need for an integrated framework so that the operational decision-making can be made more comprehensively and effectively. While such an integrated approach is very compelling and desirable, achieving it is no simple task as there are many challenges in realizing integration. In this paper, we review these challenges and indentify the underlying issues which need to be addressed for achieving an integrated approach to process supervision. We discuss the role of artificial intelligence in this context and how it provides a problem-solving platform for integration. We also survey the current status of automated approaches to operations and conclude with some thoughts on future directions.

关键词： Artificial intelligence Failure detection Integrated plant control Monitoring

来源：评论

学校读者我要写书评

暂无评论

WAVE-NET - A MULTIRESOLUTION, HIERARCHICAL NEURAL NETWORK WITH LOCALIZED LEARNING

引用

AICHE JOURNAL 1993年第1期39卷 57-81页

作者： BAKSHI, BR STEPHANOPOULOS, G Laboratory for Intelligent Systems in Process Engineering Dept. of Chemical Engineering Massachusetts Institute of Technology Cambridge MA 02139

A Wave-Net is an artificial neural network with one hidden layer of nodes, whose basis functions are drawn from a family of orthonormal wavelets. The good localization characteristics of the basis functions, both in the input and frequency domains, allow hierarchical, multiresolution learning of input-output maps from experimental data. Furthermore, Wave-Nets allow explicit estimation for global and local prediction error-bounds, and thus tend themselves to a rigorous and explicit design of the network. This article presents the mathematical framework for the development of Wave-Nets and discusses the various aspects of their practical implementation. Computational complexity arguments prove that the training and adaptation efficiency of Wave-Nets is at least an order of magnitude better than other networks. In addition, it presents two examples on the application of Wave-Nets, (a) the prediction of a chaotic time-series, representing population dynamics, and (b) the classification of experimental data for process fault diagnosis.

关键词： Neural networks

来源：评论

学校读者我要写书评

暂无评论

CONTINUOUS process IMPROVEMENT THROUGH INDUCTIVE AND ANALOGICAL LEARNING

引用

AICHE JOURNAL 1992年第2期38卷 161-183页

作者： SARAIVA, PM STEPHANOPOULOS, G Laboratory for Intelligent Systems in Process Engineering Dept. of Chemical Engineering Massachusetts Institute of Technology Cambridge MA 02139

This article presents a methodology for the continuous detection and definition of process performance improvement opportunities, especially as these pertain to the quality of operations (such as product quality). The problem is first reduced to an essentially pattern recognition formulation for which an integrated and adaptive methodology combining analogical reasoning and symbolic induction is developed. The resulting classification of past records of data is used to support the construction of a decision support system for the generation/selection of operating suggestions leading to performance improvements. The overall approach complements the usual set of statistical tools, commonly employed to address quality management problems. The basic methodology is also extended to handle fuzzy class definitions and function learning formulations. Case studies, covering both simulated and real industrial situations, illustrate the concepts and their practical utility.

关键词： Chemical Operations

来源：评论

学校读者我要写书评

暂无评论

AUTOASSOCIATIVE NEURAL NETWORKS

引用

COMPUTERS & CHEMICAL engineering 1992年第4期16卷 313-328页

作者： KRAMER, MA Laboratory for Intelligent Systems in Process Engineering Department of Chemical Engineering Massachusetts Institute of Technology Cambridge MA 02139 U.S.A.

Autoassociative neural networks are feedforward nets trained to produce an approximation of the identity mapping between network inputs and outputs using backpropagation or similar learning procedures. The key feature of an autoassociative network is a dimensional bottleneck between input and output. Compression of information by the bottleneck results in the acquisition of a correlation model of the input data, useful for performing a variety of data screening tasks. The network reduces measurement noise by mapping inputs into the space of the correlation model, and the residuals of this mapping can be used to detect sensor failures. Values for missing and faulty sensors can be estimated using the network. A related approach, "robust autoassociative networks," filter both random noise and gross errors in data resulting from faulty sensors. These networks replace with a single forward pass the conventional multiple-step procedure of data rectification, gross error detection, failure identification and sensor value replacement estimation. Autoassociative networks can be used to preprocess data so that sensor-based calculations can be performed correctly even in the presence of large sensor biases and failures. These techniques are demonstrated on an example involving inferential measurement of concentrations via tray temperatures in a distillation column. Results show that the network approach is more effective than competitive linear techniques in this application.

关键词： Neural networks

来源：评论

学校读者我要写书评

暂无评论

Solving the Hidden Node Problem in Networks with Ellipsoidal Units and Related Issues

Solving the Hidden Node Problem in Networks with Ellipsoidal...

引用

1992 International Joint Conference on Neural Networks, IJCNN 1992

作者： Kavuri, Surya N. Venkatasubramanian, Venkat Laboratory for Intelligent Process Systems School of Chemical Engineering Purdue University West LafayetteIN47907 United States

ISBN: (纸本)0780305590

A feedforward network with a single hidden layer of ellipsoidal units is considered. Network training using backpropagation algorithm runs into local minima problems, attributable mainly to poor choices of initial network weights. A fuzzy-clustering algorithm based on a modified version of Kohonen's self-organizing feature maps is used to determine the initial number of hidden nodes and the initial estimates for the hidden layer weights. The algorithm is demonstrated to determine a minimal number of hidden nodes. Supervised learning is used to fine-tune the ellipsoids initialized by the cluster information. Generalization of the network can suffer when ellipsoidal units grow unnecessarily large during the network training. Unnecessarily large ellipsoids can result in an arbitrary classification of regions in the input space far from the training patterns. The ellipsoidal fan-in function is modified so that the size of the ellipsoid generated can be controlled. An example is shown to demonstrate the utility of the clustering algorithm and the classification by networks with ellipsoidal fan-in functions. © 1992 IEEE.

关键词： Self organizing maps

来源：评论

学校读者我要写书评

暂无评论

WAVELETS AS BASIS FUNCTIONS FOR LOCALIZED LEARNING IN A MULTI-RESOLUTION HIERARCHY

WAVELETS AS BASIS FUNCTIONS FOR LOCALIZED LEARNING IN A MULT...

引用

1992 International Joint Conference on Neural Networks, IJCNN 1992

作者： Bakshi, Bhavik R. Stephanopoulos, George Laboratory for Intelligent Systems in Process Engineering Department of Chemical Engineering Massachusetts Institute of Technology CambridgeMA02139 United States

ISBN: (纸本)0780305590

A novel artificial neural network with one hidden layer of nodes, whose basis functions are drawn from a family of orthonormal wavelets, is developed in this paper. Wavelet Networks or Wave-Nets are based on firm theoretical foundations of functional analysis. The good localization characteristics of the basis functions, both in the input and frequency domains, allow hierarchical, multi-resolution learning of input-output maps from experimental data. Furthermore, Wave-Nets allow explicit estimation of global and local prediction error-bounds, and thus lend themselves to a rigorous and transparent design of the network. Computational complexity arguments prove that the training and adaptation efficiency of Wave-Nets is at least an order of magnitude better than other networks. This paper presents the mathematical framework for the development of Wave-Nets and discusses various aspects of their practical implementation. The problem of predicting a chaotic time-series is solved as an illustrative example. © 1992 IEEE

关键词： Functions

来源：评论

学校读者我要写书评

暂无评论

On the Classification Characteristics of Networks with Ellipsoidal Activation Functions: A Comparative Study

引用

IFAC Proceedings Volumes 1992年第4期25卷 213-218页

作者： S.N. Kavuri V. Venkatasubramanian Laboratory for Intelligent Process Systems School of Chemical Engineering Purdue University West Lafayette IN 47906 USA

The standard feedforward network considers a sigmoid squashing function and a linear activation function. These networks exhibit extrapolation problems because the linear discriminant functions create half spaces of high activation. For fault diagnosis, we are specifically interested in activation functions that generate closed decision regions. An ellipsoidal activation function provides a closed discriminant function. The standard backpropagation algorithm does not guarantee the proper determination of the ellipsoidal activation functions as it may run into local minima problems attributable mainly to bad choices of initial network weights. A clustering algorithm based on Kohonen's self-organizing feature maps has been developed to determine the initial number of hidden nodes and the initial estimates for the hidden layer weights. The algorithm is demonstrated to determine a minimal number of hidden nodes. Backpropagation is used for fine-tuning the ellipsoids initialized by the cluster information. The proposed algorithm avoids having to determine the number of hidden nodes a priori. Examples are shown to demonstrate the utility of the clustering algorithm and the classification by networks with ellipsoidal activation functions.

关键词：

来源：评论

学校读者我要写书评

暂无评论

An Integrated Framework for process Monitoring, Diagnosis, and Control using Knowledge-Based systems and Neural Networks

引用

IFAC Proceedings Volumes 1992年第4期25卷 49-54页

作者： R. Rengaswamy V. Venkatasubramanian Laboratory for Intelligent Process Systems School of Chemical Engineering Purdue University West Lafayette IN 47907 USA

Every day in process plants, operators deal with the tasks of monitoring a process and assessing its current state, detecting and diagnosing any abnormal behavior, and taking appropriate control actions. While these tasks are intimately related to each other and cannot really be treated as isolated tasks, typical approaches in the past have looked at monitoring, diagnosis, and control problems separately. In this paper, we discuss a conceptual framework for addressing these issues in an integrated manner. This paper will discuss two kinds of integration that are needed for a satisfying solution to this process management problem: (i) the integration of tasks, namely, monitoring, diagnosis, and control and (ii) the integration of solution approaches. The solution approaches we focus upon are knowledge-based systems and neural networks.

关键词： Expert systems Failure detection Integrated plant control Neural nets

来源：评论

学校读者我要写书评

暂无评论

Temporal Representation of process Trends for Diagnosis and Control

引用

IFAC Proceedings Volumes 1992年第4期25卷 109-114页

作者： B.R. Bakshi G. Stephanopoulos Laboratory for Intelligent Systems in Process Engineering Department of Chemical Engineering Massachusetts Institute of Technology Cambridge MA 02139 USA

A formal methodology is developed for the temporal representation of process trends, and the pattern-based diagnosis and control of chemical processes. Measured process data are analyzed at multiple scales to extract quantitative and qualitative temporal features. The multi-scale analysis procedure is based on decomposition of signals in the time and frequency domains, using wavelets. Relationships are then learned between extracted features in the measured data, and process conditions, using decision trees. This mapping is explicit and clear, and is easily converted into simple rules, which are often physically interpretable, and provide insight into the process. Case studies are presented to illustrate, a) the extraction of distinguished features in a trend and analysis of its components, and b) pattern-induced supervisory control of a fed-batch fermentor.

关键词： Hierarchical systems multi-scale analysis supervisory control artificial intelligence classification decision trees wavelet theory filtering

来源：评论

学校读者我要写书评

暂无评论

Solving the hidden node problem in networks with ellipsoidal units and related issues

Solving the hidden node problem in networks with ellipsoidal...

引用

International Joint Conference on Neural Networks (IJCNN)

作者： S.N. Kavuri V. Venkatasubramanian Laboratory for Intelligent process Systems School of Chemical Engineering Purdue University West Lafayette IN USA

A feedforward network with a single hidden layer of ellipsoidal units is considered. A fuzzy-clustering algorithm based on a modified version of Kohonen's self-organizing feature maps is used to determine the initial number of hidden nodes and the initial estimates for the hidden layer weights. The algorithm is demonstrated to determine a minimal number of hidden nodes. Supervised learning is used to fine-tune the ellipsoids initialized by the cluster information. Generalization of the network can suffer when ellipsoidal units grow unnecessarily large during the network training. Unnecessary large ellipsoids can result in an arbitrary classification of regions in the input space far from the training patterns. The ellipsoidal fan-in function is modified so that the size of the ellipsoid generated can be controlled. An example is shown to demonstrate the utility of the cluster algorithm and the classification by networks with ellipsoidal fan-in functions.< >

关键词： intelligent networks Ellipsoids Backpropagation algorithms Clustering algorithms Supervised learning Laboratories intelligent systems Chemical engineering Size control Classification algorithms

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：