XA9949961
IAEA-SM-353/26P
Development of the advanced I&C : ASICS and ADIOS
K. C. Kwon, C. H. Jung, J. T. Kim, D. Y. Lee, C. S. Ham
Korea Atomic Energy Research Institute
P. O. Box 105, Yusongj Taejon, 305-600, Rep. of Korea
Recently, digital technology has been introduced to instrumentation and control
(I&C) systems in nuclear power plants. It then actively proceeded, in the nuclear industry,
that intelligent technologies be applied to operation and operator support system. In this
paper, the Automatic Startup Intelligent Control System (ASICS) and Alarm and
Diagnosis-Integrated Operator Support System (ADIOS) which are connected to the
functional test facility (FTF) are described.
ASICS automatically controls the PWR plant from cold shutdown to 5% of reactor
power. The ASICS has a supervisor system and a distributed control system. The
supervisor system has a supervisor program to control the distributed control system, and
the knowledge base which has been designed from general operating procedures and the
operator's experiences. The supervisor system is implemented using an intelligent real-
time expert system shell G2. The distributed control system has four automation modes
such as Heating I mode, Heating II mode, Critical mode and Secondary mode, and each
mode has controllers and keep-up bands. The keep-up bands have a check-up function to
start each automation mode or the function to hold on some process variables to allow for
an operator's action. The ASICS function is verified to receive control input signals from
the FTF and send control results to FTF. Through this method, the developed control
algorithm is evaluated in the real-time environment. For the performance test of Mode I,
the heat-up rate was set to 27°C/hr, the start temperature and the target temperature were
60°C and 176.6°C, respectively. The pressure set point was 24kg/cm2. From the Mode I
test the heat-up time required to reach the target temperature was 7 hours. The automatic
temperature controller saved about 2 hours in heat-up time as compared to the manual
operation for heat-up. The measured heat-up slope was uniform and the pressure was
controlled constantly. For the operation test of Mode II the heat-up rate was the same as
in Mode I, the target temperature was 292°C. The pressure set point was 157kg/cm2.
From the Mode II test the heat-up time required to reach the target temperature was 8
hours. The automatic temperature controller saved about 3 hours in heat-up time when
compared to the manual operation. The measured P-T curve was located in the required
operation boundary. From these tests we could confirm the supervisory control rules and
the controllers of the distributed control system were designed well, as expected. ASICS
could save about 4 hours in heat-up time and could reduce the chance of repeated
operations and trips. ASICS could also reduce the operator's burden.
The overall objective of ADIOS is to improve the operation performance of the
man-machine interface system by integrating alarms, process values, and diagnostic
information to an expert system focused on alarm processing. The ADIOS was
171
implemented using the G2 real-time expert system shell. Its knowledge base is
constructed on the basis of the process knowledge of a Korean nuclear power plant and
using some advanced alarm processing concepts. Every alarm is treated as an object of
Alarm class. Objects consist of various attributes including tile message, process value,
alarm firing set-point, activation status, priority, acknowledgement or reset status, causal
alarm, level precursor, and so on. Therefore each alarm object with those attributes
contains all necessary information for processing and displaying itself in the system.
Some of attributes of an alarm object change their values dynamically during a run of
ADIOS. The values of process variable and status are fed into the attribute, 'process
value,' of the corresponding alarm objects. 'Acknowledgement or reset status' is used to
control the flashing display depending on the acknowledgement status of the alarm when
it is activated or deactivated. The attributes, 'casual alarm' and 'level precursor' are used
for prioritizing the alarm on the basis of the relationship with other alarms. The alarms in
ADIOS initially had their own default priorities which are then changed according to
plant-mode dependency, equipment-status dependency, multi-setpoint relationship, or
some other method. The dynamic prioritization is determined by using 'Rules' and
'Procedures' of the expert system shell. The plant-mode dependency is a method that the
alarms activated as a consequence of the plant mode change are de-prioritized from the
default priority. Multiple set-point relationship uses the relationships among several
alarms on the same process parameter. For example, when both the low and low-low
level alarms of a steam generator are activated, the priority of the low alarm should be
lowered. The activated alarms are displayed as rectangles around the process value on the
process overview mimic and chronological alarm lists with different colors in accordance
with their priorities. The diagnostic function module will be incorporated into ADIOS to
estimate and inform the root causes of some complicated failure behaviors for operators.
To verify only the functional effectiveness of the developed alarm and diagnosis system
without investigation of operational performance, it was tested and operated with the test
scenarios generated in the real-time test facility by activating malfunctions to simulate
abnormal plant conditions.
In applying the digital technology to nuclear power plants, we should ensure that it
does not endanger the safety and reliability of the plant. Testing and validation of the
function and performance of a digital system should be done in a realistic environment
prior to it's installation in an NPP. The objective of instrumentation and control FTF is
to test and validate newly developed digital control and protection algorithm, alarm
reduction algorithm, and the performance of operator support system, etc. The FTF
provides a simulated testing environment as an experimental test bed. The FTF software
consists of a mathematical model which simulates a three loop, 993 MWe pressurized
water reactor, and a supervisory program that comprises all the instructions necessary to
run the FTF. The hardware equipment provides an interface between host computer and
simple test panel or developed target systems to be tested. The interface module can
provide Ethernet or VXI interface to developed prototype using shared memory and also
provide the display page for the value of simulated variables. Graphic user interface
supports an easy and friendly interface between FTF and users. It is implemented through
a Picasso-3 graphic tool developed by the Halden Reactor Project. The FTF is applied to
an ASICS and ADIOS as shown in Figure 1 to test its algorithm and performance. The
172
G2 Res-Time Expert System SS»5
ASICS Supervisory System
* Ethernet
FOXSORO I/A Series DCS
I Hea&xj Mode 1 BeaJngtttoSeti j j 2n<SS*x5e CritoSMocfe!
RCS RCS RCS RCS f>2R (' | SS Es&maies
Temp Press Temp Press levei j j Level Cribcai Point
Control Control Control Control Coniroij J Confroi Projran-.
Annunciator Window Tiies
HardwiPS
Modeling Functional
•"•?"•
J- interfece Computer Test
•*•
r ADIOS Host Computer Facility
4. Process
Modeling
X-Tenninaf
VX! interface
Snp«t {Output] inpatiOutpui:
AJamt Processing
Large Scaie Schematic
rt - ,, ~,t,. G2GSI
Alarm List ASann Processing Unit Test Panel
iv. W
J 0...0
SVV1 3W2O
CRT Displays
Figure 1. ASICS and ADIOS with Test Facility
results of the test show good operational performance of the FTF in normal and transient
conditions.
ACKNOWLEDGEMENT
This work has been carried out under the nuclear research and development program by
Ministry of Science and Technology in Korea.
REFERENCES
[1] C. H. Jung, et al, "Development of Automatic Start-up Intelligent Control System for
PWR Plant," Cognitive Systems Engineering in Process Control, Kyoto, Japan, pp.
51-56, Nov. 12-15, 1996.
[2] Y. Inazumi and M. Takashima, "Automatic System for Plant Heatup and Cooldown
Operations in Japanese PWR Plants," International Symposium of NPP I&C, Tokyo,
1992.
173
[3] D. Y. Lee, et al., "An Implementation of Alarm and Diagnosis-Integrated Operator
Support System," Cognitive Systems Engineering in Process Control, Kyoto, Japan,
pp. 63-70, Nov. 12-15,1996.
[4] I. S. Kim, "Computerized System for On-line management of Failures: A State-of-
Art Discussion of Alarm Systems and Diagnostic Systems Applied in the Nuclear
Industry," Reliability Engineering and System Safety, Vol.44, pp.279-295,1994.
[5] R. M. Edwards, K. Y. Lee, Daniel E. Hughes, "Testbed for Nuclear Plant
Instrumentation and Control Validation," Proceedings of the 1996 American Nuclear
Society International Topical Meeting on Nuclear Plant Instrumentation, Control and
Human Machine Interface Technologies, pp.287-294, University Park, PA, USA,
May 6-9, 1996.
[6] Humberto E. Garcia, Richard B. Vilim, Eric M. Dean, "Hierarchical Control of
Reactor Inlet Temperature in Pool-Type Plants - II : Implementation and Results,"
Nuclear Science and Engineering, pp.337-347, Vol.125, Mar. 1997.
[7] K. C. Kwon, et al., "Development of the Test Simulator for Advanced
Instrumentation and Control Research," IEEE 6th Conference on Human Factors and
Power Plants, pp. 7-1 ~ 7-6, Orlando, USA, June 8 -13,1997.
174
