检索结果-内蒙古大学图书馆

International Joint Conference on Neural Networks (IJCNN)

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

An artificial neural network with one hidden layer of nodes, whose basis functions are drawn from a family of orthonormal wavelets, is developed. 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. 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. The mathematical framework for the development of wave-nets is presented and various aspects of their practical implementation are discussed. The problem of predicting a chaotic time-series is solved as an illustrative example.< >

关键词： Neural networks Artificial neural networks intelligent networks Functional analysis Input variables Laboratories intelligent systems Artificial intelligence systems engineering and theory Chemical engineering

来源：评论

学校读者我要写书评

暂无评论

REPLY TO SACKS AND DOYLE, “PROLEGOMENA TO ANY FUTURE QUALITATIVE PHYSICS”

引用

Computational Intelligence 1992年第2期8卷

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

来源：评论

学校读者我要写书评

暂无评论

NONLINEAR PRINCIPAL COMPONENT ANALYSIS USING AUTOASSOCIATIVE NEURAL NETWORKS

引用

AICHE JOURNAL 1991年第2期37卷 233-243页

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

Nonlinear principal component analysis is a novel technique for multivariate data analysis, similar to the well-known method of principal component analysis. NLPCA, like PCA, is used to identify and remove correlations among problem variables as an aid to dimensionality reduction, visualization, and exploratory data analysis. While PCA identifies only linear correlations between variables, NLPCA uncovers both linear and nonlinear correlations, without restriction on the character of the nonlinearities present in the data. NLPCA operates by training a feedforward neural network to perform the identity mapping, where the network inputs are reproduced at the output layer. The network contains an internal "bottleneck" layer (containing fewer nodes than input or output layers), which forces the network to develop a compact representation of the input data, and two additional hidden layers. The NLPCA method is demonstrated using time-dependent, simulated batch reaction data. Results show that NLPCA successfully reduces dimensionality and produces a feature space map resembling the actual distribution of the underlying system parameters.

关键词： .intrinsic data set Component Analysis HIDDEN Feature Space significant Reducing batch Nonlinear Principal Autoassociative identity mapping NLPCA method

来源：评论

学校读者我要写书评

暂无评论

Radial Basis Function Networks for Classifying process Faults

引用

IEEE Control systems 1991年第3期11卷 31-38页

作者： Leonard, James A. Kramer, Mark A. Laboratory for Intelligent Systems in Process Engineering Department of Chemical Engineering Massachusetts Institute of Technology Cambridge MA 02139 United States

Feedforward neural networks trained by backpropagation have recently been applied to fault diagnosis problems. Backpropagation produces decision surfaces that effectively separate training examples of different classes. However, separation of the training examples does not necessarily result in the most plausible or robust classifier. The decision surfaces of backpropagation networks may take on nonintuitive shapes because regions of the input space not occupied by training data are classified arbitrarily, not according to proximity to training data. In addition, backpropagation networks having no mechanism to detect when a case to be classified falls in a region with no training data and therefore lack data for a reliable classification, a serious deficiency in many practical applications. To remedy these difficulties a network is proposed that uses radial basis functions instead of sigmoid threshold units. This network is compared to the standard backpropagation network on an prototypical fault diagnosis example problem. © 1991 IEEE

关键词： Radial basis function networks Backpropagation Error analysis Neural networks Training data Chemical processes Feedforward neural networks Fault diagnosis Pattern recognition Chemical industry

来源：评论

学校读者我要写书评

暂无评论

Qualitative Analysis of Causal Feedback 9

引用

9th National Conference on Artificial Intelligence, AAAI 1991

作者： Rose, Philippe Kramer, Mark A. Laboratory for Intelligent Systems in Process Engineering Department of Chemical Engineering Massachusetts Institute of Technology CambridgeMA02139 United States

ISBN: (纸本)0262510596

An analysis of the properties of qualitative differential equations involving feedback structures is presented. The topological interpretation of this theory serves as the basis for a simulator of qualitative differential equations, QUAF. QUAF predicts the initial trend and final state of each variable, thereby elucidating the general character of the response. The approach requires that causal differential equations replace algebraic forms derived from pseudo-steady state (moving equilibrium) assumptions. QUAF is compared to the qualitative simulator QSIM on an example involving interconnected tanks, and a significant narrowing of the number of interpretations of system behavior is observed. © 1991, AAAI (***). All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Diagnosing Noisy process Data using Neural Networks

引用

IFAC Proceedings Volumes 1991年第6期24卷 547-552页

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

An account of the application of neural networks for detecting and diagnosing faults during process transients in the presence of random measurement noise is presented The approach employs a feedforward neural network with backpropagation as the learning algorithm. Two representation techniques for capturing dynamic process trend data in the form of a timeseries as well as in the form of a moving average have been developed and used for training the network. Networks with various number of hidden units have been tested and compared with respect to their performance in recall and generalization. The time required for training increases substantially when the level of noise in the training data increases. Trained networks have excellent recall performance even in the presence of noise, in both steady state and dynamic processes. The networks are successful in diagnosing untrained, noisy single fault patterns in a large majority of the steady state and dynamic cases tested.

关键词： Failure detection fault diagnosis dynamic processes neural networks artificial intelligence

来源：评论

学校读者我要写书评

暂无评论

Multi-Scale Descriptions of Real-Time Trends and their Impact in Pattern Recognition

Multi-Scale Descriptions of Real-Time Trends and their Impac...

引用

American Control Conference (ACC)

作者： George Stephanopoulos Laboratory for Intelligent Systems in Process Engineering Department of Chemical Engineering Massachusetts Institute of Technology Cambridge MA USA

Recognition and interpretation of real-time process trends is the center-piece of the socalled intelligent controller, which recognizes faults, performance degradation, and scopes the problems associated with efficient disturbance rejection and process model/controller adaptation. In this paper we will show how a unifying symbolic representation of process trends leads to a systematic extraction of their temporal features and precise classification of trends. Such classification may proceed in a syntactic, semi-quantitative, or fully quantitative manner, depending on the characteristics of the problem being solved. Furthermore, we will argue that a multi-scale representation of process trends is essential in pattern recognition, and we will show how the unifying representation of trends allows for the decomposition of signals at various levels of abstraction.

关键词： Pattern recognition Signal processing Adaptive filters Degradation process control Fault diagnosis Real time systems Laboratories intelligent systems systems engineering and theory

来源：评论

学校读者我要写书评

暂无评论

QUALITATIVE MODELING AND FAULT-DIAGNOSIS OF DYNAMIC processES BY MIDAS

引用

CHEMICAL engineering COMMUNICATIONS 1990年第1期96卷 205-228页

作者： OYELEYE, OO FINCH, FE KRAMER, MA Laboratory for Intelligent Systems in Process Engineering Department of Chemical Engineering Massachusetts Institute of Technology Cambridge MA 02139

The Model Integrated Diagnostic Analysis System (MIDAS) is a program for diagnosing abnormal transient conditions in chemical, refinery, and utility systems. MIDAS employs causal reasoning using an event model derived from piping and instrumentation diagrams, and from quantitative process models. Root causes typically considered are equipment degradation and failure, incorrect manual actions, and external disturbances. By prompt and accurate diagnosis of these conditions, MIDAS can reduce the risk of safety hazards, material wastage, and unnecessary *** paper extends and complements a previous report on MIDAS† by detailing the qualitative modeling techniques used in MIDAS, and presenting the results of a simulation case study. The modeling methodology employs a new result on qualitative modeling to help resolve ambiguity from feedback loops and locate apparent “non-local causality”. The event model allows MIDAS to exploit the sequence of malfunction propagation in its internal reasoning. The models include control system responses and MIDAS uses this knowledge to correctly identify malfunctions even when the primary symptoms are concealed by control system compensations. The reasoning methodology used to perform the diagnosis is process *** case study reported involves diagnosis of a reactor-heat exchange process with associated control systems, with 10 sensors and over 100 potential malfunctions. The results show that MIDAS includes in its hypothesis set the correct malfunction 98.9% of the time. Of the scenarios tested, 40% exhibited control system compensation, 20% exhibited inverse response, and over 30% included out-of-order events.

关键词： Fault diagnosis Causal reasoning process safety Alarm analysis Deep knowledge expert systems

来源：评论

学校读者我要写书评

暂无评论

REPRESENTATION OF process TRENDS .1. A FORMAL REPRESENTATION FRAMEWORK

引用

COMPUTERS & CHEMICAL engineering 1990年第4-5期14卷 495-510页

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

A formal methodology capable of transforming time-records of process variables into meaningful and explicit descriptions of process trends in real-time is presented in this paper. The representation is based on the formal definition of “temporal episodes,” which provide descriptions of trends by constructing “histories of episodes” making explicitly all the important “domain landmark values” specified by the user as well as the “geometric landmark values” generic to trends. Model inaccuracies are confined by establishing specific measures of information lost in the descriptions and keeping them below those of the process models. Despite the simplicity of the representation primitives, it is shown that the representation can provide complete, correct, robust and very compact models for process trend histories. In contrast to pure qualitative descriptions, the new representation contains sufficient quantitative information to allow for powerful inferences of the trends of the unmeasure variables.

关键词： process Control

来源：评论

学校读者我要写书评

暂无评论

ARTIFICIAL-INTELLIGENCE IN process engineering - CURRENT STATE AND FUTURE-TRENDS

引用

COMPUTERS & CHEMICAL engineering 1990年第11期14卷 1259-1270页

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

Recent advances in artificial intelligence have changed the fundamental assumptions upon which the progress of computer-aided process engineering (modeling and methodologies) during the last 30 yr has been founded. Thus, in certain instances, numerical computations today constitute inferior alternatives to qualitative and/or semi-quantitative models and procedures which can capture and utilize more broadly-based sources of knowledge. In this paper it will be shown how process development and design, as well as planning, scheduling, monitoring, analysis and control of process operations can benefit from improved knowledge-representation schemes and advanced reasoning control strategies. It will also be argued that the central challenge coming from research advances in artificial intelligence is "modeling the knowledge", i.e. modeling: (a) physical phenomena and the systems in which they occur;(b) information handling and processing systems;and (c) problem-solving strategies in design, operations and control. Thus, different strategies require different forms of declarative knowledge, and the success or failure of various design, planning, diagnostic and control systems depends on the extent of actively utilizable knowledge. Furthermore, this paper will outline the theoretical scope of important contributions from AI and what their impact has been and will be on the formulation and solution of process engineering problems.

关键词： Artificial Intelligence

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：