ML20062F749
| ML20062F749 | |
| Person / Time | |
|---|---|
| Site: | Dow Chemical Company |
| Issue date: | 11/19/1990 |
| From: | DOW CHEMICAL CO. |
| To: | |
| Shared Package | |
| ML20062F745 | List: |
| References | |
| NUDOCS 9011280148 | |
| Download: ML20062F749 (13) | |
Text
, . . _ ,_ ..- .. . . . .- - . - - - - _ - - _ - - - _ _ - _
t
, DOW TRIG A RESEARCH REACTOR '
i MICROPROCESSOR BASED INSTRUMENTATION AND CONTROL SYSTEM FOR THE:
DOW TRIGA RESEARCH REACTOR '
1 19 November 1990 l
l l
l 9011280148 901119 PDR ADOCK 05000264 '
P PDL i
-- , . . . , , -,, . ~ . - . - - . . . , . . . -
j DOW TRIGA RESEACCH REACTOR l
l TABLE OF CONTENTS l
y INTRODUCTION The Dow TRIGA Research Reactor . .
I :
Comparison of Dow, GA, and AFFRI facilities 2-Safety channels in the DTRR 3 J LOCATION AND ENVIRONMENT 4- L POWER SUPPLIES 5-NM 1000 Independence 5 H Indicators . 5' <
t Isolators 6 NPP 1000 Hard wired connections 6 Isolation -6
-Indicators 6' WATCHDOG 7..
1 a S AFETY/NON-S AFETY FUNCTION ISOLATION 7 SCRAM SYSTEM 7 l
OTHER CHARACTERISTICS
! Failure modes 8 Self-diagnostics .8, 1 Software verification 8 l VALIDATION / TESTING PROGRAM 8 l
MAINTENANCE ANDTESTING 10 TRAINING OF OPERATORS 10L i';
F i
, . . _ . . - - - - ,, e a f--
~
. 1
, DOW TRIGA RESEARCH REACTOR '
l Ll .
INTRODUCTION i
The Dow TRIGA* Research Reactor (DTRR) is a 300 kilowatt pool type nuclear. 'l reactor designed as a Training, Research, and Isotope production facility by General Atomics. The reactor fuel, nominally 20% enriched in 235U,is fabricated in a U - i l form which provides a unique safety factor in terms a of prompt negative-Zrli 1.7 temperature coefficient of reactivity. The reactor at Dow is a research and analytical tool within the Michigan Division Analytical Laboratory where it has been used as a '
source of neutrons for activation analysis and other research functions since 1%7.
The reactor is found in a below ground pool with no horizontal beam ports or cost .
level access. The poolis located in a 20' x 25' Reactor Room which contains sample i storage facilities, experimental facilities, spare parts, and tools. The Reactor Room is .
monitored by a Continuous Air Monitor (a Geiger-Mueller detector equipped with an air pump and paper filter), and by an Area Monitor. The Continuous Air Monit x is provided with an alarm serpoint; radiation levels above the setpoint trigger aud ole and .;
visible signals and also transmit a signal to the Dow Security Dispatcher. Loss of.
power at the Continuous Air Monitor also triggers a signal to the Dow Security Dispatcher. The Area Monitor, a Geiger-Mueller detector, has a readout and alarm at the reactor console. The center of the reactor is at least 100' from the Dow security ;
fence. The reactor console and the associated laboratories are located in adjacent rooms.
The reactor is typically operated for periods of 10 minutes to several hours each day at power levels up to 300 kilowatt. The average energy output is approximately 1.5 megawatt days per year.
The original (1967) control system was upgraded in 1973 and 1974, eliminating all of the tube-type electronics and greatly improving the reliability of the entire system.
The operation of the reactor and all of the components was critically scrutinized in 1985 1986 as part of a determination of the value of the neutron-related programs to the Dow Chemical Company, in preparation for applying for a renewal of the operating license which was to expire in December 1986. Senior Dow managers decided that the ,
program was worth the cost of relicensing, and in May 1989 an amendment to the l operating license extended it through 2009 with an increase of authorized maximum j power level from 100 kilowatts to 300 kilowatts. Given this long term commitment to operation of the reactor the concern for dependability and reliability dictated a review :
- l of the control system. The age of the. components and the lack of replacement parts led to a decision to replace the entire control system and to rebuild the control rod drives, l
- Tradenwk of General Atomics 1
DOW TRIGA RESEACCH CEACTOR and the General Atomics microprocessor-based instrumentation and control system I was chosen. A commitment was made in mid 1990 and delivery is expected in late .
I November carly December 1990, with installation shortly thereafter.
The analysis of the operating characteristics of the General Atomics Microprocessor Based Instrumentation and Control System with respect to the Dow TRIGA Mark I J
l ..
Research Reactor is based upon similar analyses of the General Atomics 250-kilowatt TRIGA Mark I teactor (GA) and the Armed Forces Radiobiological Research Institute TRIGA Mark F reactor (AFFRI).
Unlike the GA and AFFRI facilities, the DTRR is neither licensed for nor configured
{
for pulsing. The DTRR therefore has three, rather than four, control rods, and has no r need of a separate neutron channel for pulsing operation. The DTRR does, however,L l have a Technical Specification for minimum period, with an associated scram l l capability. Also unlike the GA and AFRRI facilities the DTRR has ao instrumented . I fuel elements and thus no fuel element temperature scram capability. This is a consequence of the lack of pulsing capability - the DTRR is restricted to steady state' l operation, and experience with TRIGA' reactors which do have fuel element - 3 i temperature measurement capability shows that operation at power levels several times.-
higher than the 300 kW maximum licensed power level of the DTRR will not produce fuel element temperrares sufficient to damage the fuel. ;
In other itspects, however, the facilities are similar. The most imponant shared featuit !
is the use of standard TRIGA fuel, in which the prompt negative temperature ;
coefficient pmvides an effective safety factor for control of the nuclear fission reaction.
A comparison of some characteristics of these three facilities, including the safety channels,is given in Table 1. ,
i Table 1. Comparison of Dow, GA, and AFFRI Reactors Dow GA - AFFRI Non-pulsing Pulsing . Pulsmg 300 kilowatt 250 kilowatt. 1000 kilowatt t TRIGA MARK I TRIGA MARK I TRIGA MARK F ,
NM-1000 digital -NM-1000 digital NM-1000 digital :
NPP-1000 analog NPP-1000 analog ' _
NPP-1000 analog NP-1000 analog NP-1000 analog (configured as NP-1000) '
i
.i i
i
)q i
e DOW TRIG A RESEARCH REACTOR D
I l
There are at present three safety channels at the DTRR capable ofinitiating scrams .
based upon information received from neutron detectors. Two of these provide !
independent scram capability according to power level set points, while the third I provides scram capability according to the period serpoint. A similar configuration is v i proposed for the replacement instrumentation and control system; the comparison is : 1 shown in Table 2. All other scram capabilities, which include high voltage setpoints : -
)
on two neutron detectors as well as manual scram, keyswhch scram, and loss of power : l scram, will be identical to the present system.
Table 2. Safety Channels in the Dow TRIGA Research Reactor Present system Proposed system
% power channel % power channel analog analog - NPP-1000 7 2 decades (1 kw to 100 kw) 2 decades (3 kw to 300 kw) uncompensated ion chamber uncompensated ion chamber scram capability- power level scram capability - power level wide range linear channel wide range linear channel analog digital- NM 1000 .
10 decades (to 100 kw) 10 decades (to 300 kw) 1 compensated ion chamber fission chamber scram capability - power level . scram capability - power level log channel log channel analog digital- NM 1000 L
10 decades (to 100 kw) 10 decades (to 300 kw)- 4 fission chamber fission chamber scram capability - period scram capability- period There are at present two interlocks intended to prevent a) withdrawal of control rods after the count rate in the log power channel has dropped below a setpoint and b)-
withdrawal of any two control rods simultaneously by manual operation of the rod drive switches. . Both interlocks, required by the Technical Specifications, will be preserved in the proposed replacement system.
l 3-l L, .
i
)
DOW TRIGA RESEARCH REACTOR l
l l
LOCATION AND ENVIRONMENT l
The NM 1000 and the rack containing the NPP 1000, the control rtxi drive translators, and the DAC system, including the computer, will be placed in the reactor room, a 20' 1 X 25' structure built of concrete blocks with a steel framework. This room is served l by the same air conditioner / heater / alt movement system that provides conditioned air l for the console room, two associated laboratories, and four offices,. This air is not recirculated; the reactor room air is vented directly to the outside of the building. This system is independent of the rest of the building. The air in the reactor room is chilled or heated, as required, and is subject to humidity control which climinates the usual wintertime static electricity problems experienced in Michigan. The air tumover rate I in this room is approximately 17 room volumes per hour, l The reactor console, with the display units, the CSC computer, and the control switches will be placed in the control room adjacent to the reactor room, replacing the present control system. 'Ihe control room is a 13' X 20' portion ot'the facility which is configured as an office no chemicals or radioactive materials are allowed in this room. The temperature, humidity, and air flow are controlled by the same system that serves the reactor room. The air turnover rate in the control room is similar to that for other office areas; this room is kept at a positive pressure with respect to the reactor -
room and the associated laboratories to prevent possible transfer of unwanted materials -
into the control room.
In these two rooms the normal temperature range is 68 to 72 F and the humidity is controlled in the summer by the air conditioner and in the winter by steam injection
, units; humidity levels of 35% to 50% are the norm.
The NM 1000 is an industrial neutron monitoring system used in research reactors and in nuclear power plants. It is contained in two 24" X W" X 10" NEMA enclosures, I
which are to be mounted on the concrete block wall of the reactor room using steel framework. The components of the NM 1000 are mounted on sturdy platforms within the NEMA enclosures, which provide massive thermal pathways for dissipation of heat and extremely rugged constructi9n to prevent significant movement.
The NPP 1000 and other equipment is mounted in a multi shelf cabinet in the reacter l
room and is not otherwise attached to the building. This cabinet contains power '
supplies and power conditioning equipment; relay boards, optical isolator boards, and input scanner boards; devices e monitor and condition signals for temperature, conductivity, and other parameters; the NPP-1000 power safety channel; various analog termination and serial communications boards; and an industrially-hardened ,
inicroprocessor based computer. The cabinet and the components are constructed to l industrial standards and are expected to withstand the stresses of this environment.
. 4 w-
- y w*qw w - s-- - ,.--.g w- --- --eyev-w w --
m'Ww
i 4 ,
,. DOW TRIGA RESEARCH REACTOR 1
This equipment has not been subjected to seismic qualification testing, nor has the equipment which it replaces. A major characteristic of this and similar TRIGA l reactors is that an inadvertent scram has no significant safety consequences, primarily I due to the quite low amount of residual heat in the fuel elements, which precludes fuel damage following an emergency shut-down. Scrams are initiated by interruption of current to the control rod magnets followed by Sravity assisted insertion of the control rods; no other mechanisms are required to shut down the reactor.
POWER SUPPLIES The power supplies are passively buffered to protect against normal fluctuations of ,
applied power and against line surges. The power lines at the present installation have been quh stable; no instances of loss of control of the reactor within the 20-year l experience of the most senior SRO have been recorded.
The computers have the normal lithium battery system which keep the clocks operating during shutdown.
The NM 1000 has a battery powered non volatile random access memory which serves to store data during loss of power, i Loss of power rest.as in a scram situation; the reactor is shutdown in a very short -
period of time by the gravity assisted insertion of the control rods. There are no safety conse quences of such a shutdown since there is very little residual heat in the fuel elements. For this reason there are no plans to install uninteruptable power supplies for i
this system.
NM 1000 The NM 1000 digital neutmn monitoring system uses a fission chamber, mounted on !
the periphery of the reflector of the reactor core, as a source of signals proportional to the power level of the reactor. The total system is independent of any other safety l channel, and it provides four sets ofinformation to the console through the DAC l computer:
wide-range log power,10 decades to 100%, displayed on the computer-driven l display and on the hard wired LED display, both at the console
[
l l reactor period, 30 seconds thrcugh infinity to +3 seconds, displayed on the l computer-driven display and equipped with a scram setpoint, independent of the DAC and CSC computers 5- ,
___ _ ._ _ _ _ . _ _ _ _ .~ _ . _ _ _ _ _ _ . . _ _ _ _
l DOW TRIGA RESEARCH REACTOR i
wide range linear power, ten decades to 1000 kW, displayed as an auto ranging graph on the computer-driven display and on the hard-wired display, both at the console linear power, two decades to 120% of full power, displayed on the computer- .
driven display and on the hatd-wired LED display, both at the console, and I equipped with a scram setpoint independent of the computers Signals from the fis:lon chamber neutron detector are passed to the NM 1000 through a shielded cable, reducing the pickup of electrical noise signals and enabling the wide-range channel to fully utilize the sensitive neutron detector. Thl permits the use of one detector for both very low level signals (start up range) to the maximum permitted power level. Signals from the NM 1000 to the DAC system.(and then to the computer-driven displays) are passed through optical isolators to prevent conducted noise signals from passing through to the computers, displays, and control systems. Any electronic noise signals which do get into the system are expected to cause a scram, probably with the period circuit.
The scram relays are hard wired into the magnet power circuit, independent of the DAC and CSC computers.
)
NPP 1000 i The NPP-1000 analog neutron monitoring system is an independent channel using an uncompensated ion chamber attached to the penphery of the reactor reflector as the i neutron detector. The normal pulsing functions associated with the NPP 1000 have been disabled so that this channel functions only as an independent neutron monitoring channel, two decades to 120% of full power, with hardwired scram capability.
The power and signal leads from the NPP 1000 to the uncompensated ion chamber are shielded to reduce pickup of electrical signals. Outputs from the NPP 1000 to the DAC and console are fed through opticalisolators to reduce the effects of electromagnetic interference signals.
! The NPP-1000 provides signals to the computer driven display through the DAC and t
CSC computers, and to the hard wired LED display at the console.
, The scram relay is hard wired into the ma[ net power circuit, independent of the DAC and CSC computers.
6-t
DOW TRIGA RESEAOCH CEACTOR ,
1 ,
l J
WATCHDOG FUNCTION I The CSC computer is equipped with a watchdog circuit which monitors performance
- of a number of software functions. If the watchdog timer is not reset within ten seconds by one or more of these software functions then the watchdog initiates a scram !
condition. This ten second setpoint is fixed and can not be changed by the operator.
The purpose of this circuit is to prevent operation of the reactor while the computers are not performing properly. The operation of the watchdog is tested regularly under a Technical Specification surveillance requirement. '
ISOLATION BETWEEN SAFETY AND NON SAFETY FUNCTIONS The NM 1000 and NPP 1000 scram functions are hard wired to the scram circuits and are completely isolated from the DAC/CSC computers and the rest of the system.
Signals from both the NM 1000 and the NP 1000 systems are hardwired to bargraph indicators at the console, again isolated from all other components. Signals from the NM 1000 and NP 1000 neutron channels are passed, through opticalisolators, to the DAC and thence to the CSC and to the computer-driven display. The opticalisolators are of commercial grade and have been tested by the manufacturer to standard commercial criteria.
Communication between the DAC and CSC computers is through high speed serial communications ports and the computers are not isolated from each other.
The CSC computer communicates with the control console through the DIS 064 board.
Isolators are not used.
)
! SCRAM SYSTEM The proposed scram system will contain all of the features of the present system with the addition of the watchdog scram and with the substitution of the digital NM 1000 l power and period scrams for the analog power and period scrams presently available.
The scram mechanism involves interruption of the power flowing through the control l rod magnets, allowing the control rods to be fully inserted under the influence of -
gravity.
Any interruption of power initiates a scram condition, either loss of AC power or l turning the keyswitch.
Other scram initiating events include operatio t of the manual scram switch, a period a shorter than the set point, and two independent power level set points.
7
- I
, DOW TRIGA RESEARCH REACTOR OTHER CHARACTERISTICS I An analysis of the scram circuit was performed by the University of Texas (Austin),
using a fault tree appmach, to evaluate various failure modes of the reactor safety system, including the unique characteristics of the software of the digital NM 1000. 1 The study concluded that a failure of all safety systems, which would be required for a :
failure to scram, was extremely unlikely. '
The failure analysis for the DTRR involves an insenion of $3.00 (the maximum licensed excess teactivity)in a very shon time, considering the fuel to be aluminum-clad low hydride TRIGA fuel elements (DTRR Safety Analysis Report,1986). In !'
reality, onl" one of the 78 fuel elements is aluminum-clad, and by the Technical' Specifications t!,is one fuel element must reside in either of the two outer rings of the core, where it will not ne exposed to the maximum neutron flux or power level. <
A funher accident anal. ' sis was produced for the 250 kW GA reactor, which is a Mark I similar to the DTRR again using an aluminum-clad low hydride TRIGA fuel core (General Atomics 1T.F 252,1988), for a $5.00 reactivity insenien. This analysis can .
apply to the DTRP. because the reactors have identical cott structure, reflectors, sample handling "acilities, pool size, and water content. The OA reactor is a pulsing reactor but the D"RR is not; altheigh there is no reasonable way to add the entire excess reactivity to the DTRR reactor in a short time, the analyses show that the consequences of such an insenion of reactivity would not damage the fuel. -
The instrument features some self diagnostics programs other than the watchdog circuits. The NM 1000, in panicular, performs such a function continuously, as does the computer operating the communications actwork. The function of these self.
diagnostics systems is to dump the computers out of the operating mode into the IC-DOS operating system, with consequent activation of a scram sequence.
The software used for the Dow installation is the same as that used for the GA system.
It is expected that the verification and validation process used by GA (General Atomics E1171002 March 1989)is transferable to the DTRR.
VALIDATION / TESTING Extensive testing of the OA microprocessor based instrumentation and control system at the GA, AFFRI, and McClellan installations indicates that this system is capable of operating properly. GA also performed an extensive verification and validation testing of the software. The development of the system by GA,its customers,'and the i
Console User's Group has resolved discrepancies and operating problems, and the:,e i
three systems are now in full operation. 1 8-
DOW TRIGA RESEASCH REACTOR l
The Dow initial testing program is intended to show that this particular instrument is capable of performing the assigned tasks. In addition to the initial testing program a self-diagnostics feature allows continuous on line testing, and the routine surveillance tasks normally performed by the reactor operators help to assure continued operation and rapid discovery of problem areas. Finally, this system, unlike its predecessor, will be backed up by the availability of replacement parts. A considerable assortment of ,
spares has been purchased and tested in operation in this instrument, and will be on hand at this facility. In addition, GA and its suppliers will provide quick service and supply of other parts as necessary.
At least one of the control rod drives must be changed to a stepping motor drive in order frx the new system to be able to operate in the sen'o mode. It has been decided to have all three control rod drives retumed to GA to be rebuilt and equipped with stepping motor drives in order to enhance reliability. This will make every item above the water line, with the exception of the lazy susan drive, either new or completely rebuilt, providing a greater assurance of reliability, liowever, the present control system cannot operate the stepping motor control rod drives; and the new system cannot operate the old servo drives. A procedure has been devised to provide testing and assurance that the new system will operate properly before relinquishing the scram circuits of the old system.
The NM 1000 will be connected to a spare fission chamber neutron detector located on the periphery of the reficctor. The stepping-motor control rod drives will be controlled by the new console. The present safety channels (percent power, wide range log with period, and the wide range linear channels) will be maintained in operation, and the <
scram loop involving all of these will be included in the scram loop for the new console. In this way the scram loop will consist of all of the present scrams plus the period end power level scram of the digital NM 1000 system. The NM-1000 calibration and scram settings will be tested in this configuration.
When the NM 1000 has been checked the NPP 1000 will replace the wide range linear channel, usin the compensated ion chamber connected as an uncompensated chamber.
l The NPP 10 ' will be calibrated and tested, while the reactor is protected by the NM-l 1000, the wib-range log (analog ) channel, and the percent power (analog) channel.
I l
' Following successful testing of the NPP-100, the NM 1000 will be transfermd to the fission chamber now used for the wide-range log power (analog) channel and calibrated and tested, with the reactor protected by the NPP-1000 and the original >
percent power (analog) channel.
Following the testing of the safety channels the com.nl rods wG be calibrated using power levels not to exceed ten kilowatts.
9
DOW TRIG A RESEARCH [EACTOR i
When the control rods are calibrated a thermal calibration will be undertaken (annual checklist procedure, task #2) which involves operation at about 80% of full power for a period of time while the rate of rise of the pool temperature is measured. The neutmn detectors will be adjusted to give the correct power level readings.
Finally the percent power (analog) channel from the old console will be disconnected and the scram circuit of the new console will control the reactor. The old console will be removed.
MAINTENANCE AND TESTING An extensive program is in place to evaluate the operation of the components of the present control system. His program includes regularly scheduled surveillance and test procedures which %clude those items required by the Technical Specifications as well as a number of other items. These procedures are intended to determine the operability of the systems and to give early waming of possible malfunctions. "
Major daily checks of all scram systems, channel tests of a number of information systems, and operation of the reactor at a specified low power are used to document the operability of the systems and to evaluate the operating characteristics of the reactor itself.
Other tasks perfonned at monthly, semi annual, and annual intervals include calibrations ofimportant systems (including the control rods, the power levels, and the radiation sensing instruments) as well as a number of non safety-related parameters.
The rigor of these procedures, which have been in place for a number of years, has been adequate to assure proper maintenance of the present control system, and is expected to provide continued assurance of proper operation and detection of problem areas with the proposed system.
l TRAINING OF OPERATORS Three licensed SROs and two trainees (the trainees each have more than five months of experience operating the DTRR) participated in training sessions at the General Atomics facility in San Diego. About 1/2 of the 40 hour4.62963e-4 days <br />0.0111 hours <br />6.613757e-5 weeks <br />1.522e-5 months <br /> training course was devoted '
to classroom training, covering details of the operation of the console, the structure and
- operation of the NM 1000 and NPP 1000 neutron charnels, the software, the safety circuits, and the computers.
10
' l
. .. l DOW TRIGA RESEARCH REACTOR j 1
The remainder of the time was used mostly in hands-on operation of the Dow console i during its final stages of testing and adjustment - as a test stand operation, not connected to a reactor or neutron detectors operating the computers, following the testing sequences, and going through the NM 1000 and NPP 1000 circuits in great ,
detail, generally reinforcing the classmom instruction. j Finally, each Dow person manipulated the controls of and performed several operations on the GA 250 kW reactor. Operations of this reactor as part of training programs for non GA personnel have been approved by the OA Safety Committee and .
were performed under the direction of a licensed SRO of the GA facility. Operations included startup, intermediate reactivity changes, and shutdown. Although the GA instrumentation and control system is not identical to the proposed Dow installation the steady state operations are quite similar. The Dow people performed such operations as they would be performed at the Dow facility, including startup, approach to power l using relatively long periods consistent with the Dow technical specifications, manual l i
operation at power, and automatic mode steady state operation. Particular attention ,
was paid to observation and control of the period, which in the proposed system is generated by the digital NM 1000 circuit, as compared to the current Dow system where the period information is derived from the analog wide range log power channel.
I i
11-
, _ - - _