Skip directly to search Skip directly to A to Z list Skip directly to page options Skip directly to site content

NIOSHTIC-2 Publications Search

Search Results

Artificial Intelligence in Process Plant Safety.

Authors
Venkatasubramanian-V
Source
Laboratory for Intelligent Process Systems, School of Chemical Engineering, Purdue University, West Lafayette, Indiana 1990 Nov:72 pages
Link
NIOSHTIC No.
00201456
Abstract
A process general inference procedure was presented that derived the malfunction hypotheses from the structure and behavior of the process facility. A framework for developing causal model based approaches to fault diagnosis of chemical process facilities was proposed. This framework involved the development of unit centered models that incorporate first principles knowledge of the process such as conservation laws. This approach also treated the structural connectivity of the process explicitly. A framework was also developed for the integration of deep level and compiled knowledge for process diagnosis. An object oriented architecture was developed for facilitating knowledge representation and reasoning. The feasibility of this novel approach was demonstrated in the form of a prototypical expert system, called MODEX2, for Model Oriented Diagnostic EXpert 2. Also presented was a mechanism called causality based failure driven learning for the automatic acquisition of diagnostic knowledge. This technique refined a heuristic in the event that the hypothesized causal origin offered by the heuristic was incorrect. The heuristic was thus strengthened, preventing it from firing if a similar situation were to arise. A technique was also developed for the automatic generation of process trend descriptions from sensor data. This trend description can be used to perform diagnoses under transient conditions of the process.
Keywords
NIOSH-Grant; Control-technology; Chemical-manufacturing-industry; Chemical-industry-workers; Occupational-safety-programs; Automation;
Contact
Chemical Engineering Columbia University Chme Building New York, New York 10027
Publication Date
19901120
Document Type
Final Grant Report;
Funding Amount
141129.00
Funding Type
Grant;
Fiscal Year
1991
NTIS Accession No.
PB92-115724
NTIS Price
A05
Identifying No.
Grant-Number-R01-OH-02352
NIOSH Division
OEP
Priority Area
Control Technology and Personal Protective Equipment; Research Tools and Approaches; Control-technology;
Source Name
Laboratory for Intelligent Process Systems, School of Chemical Engineering, Purdue University, West Lafayette, Indiana
State
IN; NY;
Performing Organization
Columbia University New York, New York, New York
TOP