Systems/Process Monitoring, Diagnostics and Control
The goal of the Nuclear Engineering Division's research on advanced monitoring, diagnostics, and control
is to develop and demonstrate software tools for optimizing the operation of nuclear power plants and
other engineering systems. The role of this software is to generate and substantiate information on-line
about the condition of plant systems; this information can then be used to maximize plant availability
and power output, optimize plant maintenance functions, and help operators cope safely and effectively
with potential upsets.
The Division's activities in this area are conducted in cooperation with electric power utilities and researchers from other organizations. An Advanced Simulation and Controls Laboratory has been established to support these collaborative development efforts.
Research has been directed toward the development and demonstration of software capabilities in the following areas:
- Process Monitoring & Signal Validation
- Diagnostic & Advisory Systems
- Advanced (AI-based) Nonlinear Controllers for Industrial Processes
In addition to research on the foregoing software capabilities, future research will address additional
areas of computer-based operations support that are relevant both to the extended time scale of the
reactor operating cycle and to the shorter time scale of plant transients. In the former category,
techniques will be investigated for optimizing plant state estimates and for using this information
to help maximize plant availability and power output within approved safety limits. Developments in
the latter category will include system identification and signal validation techniques applicable
to dynamic conditions, and software for optimizing maneuvering control sequences and strategies for
coping with plant upsets. Research will also be performed to address the challenging problems involved
in demonstrating the reliability and validity of the various digital-based operations support, control,
and safety protection software.
Artificial Intelligence / Expert Systems Expertise
Nuclear Engineering Division personnel have considerable expertise in the areas of structural mechanics, thermal-hydraulics and materials behavior. This expertise can be implemented in structured artificial intelligence (AI) knowledge bases for application either as off-line AI based design tools (AICAD) or as on-line system diagnostic and supervisory aids. Expert system rules and neural network topologies have been used to develop these applications.
Division engineers have extensive experience in the development of expert systems for process data validation, plant state verification, and process fault diagnosis (sensor and other equipment or system failures, human error, etc.). Knowledge database development uses a combination of experience-based expert system rules with rules generated from analytical models describing the processes. For the expert system to be used on-line, either as an operator aid or to provide input to the plant control system, process models and neural net simulators of the process that run in real time have been developed.
The Plant Analysis and Diagnostics Section specializes in the development, adaption and application of signal processing and pattern recognition techniques and state-of-the-art AI-based algorithms for solving issues associated with the surveillance, monitoring, diagnostics and control/management of systems. Applications have been to a wide variety of systems ranging from thermal-hydraulic systems in reactor plants to NOX control in fossil plants to bioinformatics work on the human genome.
Last Modified: Tue, September 14, 2010 6:39 PM