Taming the Chernobyl Avalanche
Frigyes Reisch
Nuclear Power Safety, KTH, Royal Institute of Technology
S-10691Stockholm – Sweden
ABSTRACT
Nneutron kinetics equations are one of the
cornerstones of the theory of nuclear reactors. The awareness
of nuclear
engineering students of its importance is a precondition that
the new generation will handle the currently operating and future
nuclear power plants safely. They have to learn how to design
control systems for reactors sensitive to an avalanche situation
like power increase.
The classical neutron kinetic equations with
six delayed groups are not solved analytically. Here they are
solved both numerically
and with a corresponding block diagram and applied to a Chernobyl
type reactor. The results are displayed graphically.
The Chernobyl type reactors have positive void
coefficients. When water is replaced with steam the power is
increasing. A
sudden increase of the steam content causes a rapid power surge.
The importance of choosing the magnitude of the void coefficient
and the parameters for the automatic control system is demonstrated.
1. Introduction
Now, 22 years after the Chernobyl
accident it is important for today’s and tomorrow’s
generations of nuclear engineers to learn to design control systems
for reactors with "runaway"
characteristics.
The Chernobyl type of reactor
(RBKM) core is a huge graphite cylinder (7 m high, 12 m diameter)
and within some 1600 channels
with water and steam cooled fuel rods inside. The fission neutrons
are slowed down (thermalised) mainly in the graphite and a
portion of them is absorbed in the water. When a part of
the water is
replaced by steam (void) the absorption is reduced, causing
a positive reactivity contribution. This is the positive void
coefficient. After the accident the enrichment of the fuel
was increased and the neutron spectrum became harder resulting
in a lower
positive void coefficient
Figure 1, the Chernobyl Reactor
2. The experiment
With the Chernobyl experiment
due to the abrupt decrease of the speed of the main circulation
pumps
and the sudden drop of
the reactor pressure at low reactor power and heavy Xenon
poisoning, the steam (void) content in the coolant channels
increased
suddenly from a few percent to about 50%. Thus the positive
void coefficient - about 30 pcm/% - caused a large reactivity
insertion.
The neutron flux and thereby the reactor power increased
very fast. Due to the thermal inertia of the fuel and the
small value
of the fuel temperature coefficient the Doppler effect could
not break the power excursion. Therefore, to characterize
the process at the initial phase, to use only the reactor
kinetics
equations is sufficient.
3. Simplified neutron kinetics equations
Here
|
time (sec) |
N |
neutron flux (proportional to the reactor power) |
k |
change of the effective neutron multiplication factor
(keff) |
|
sum of the delayed neutron fractions (here 0.006502) |
ß i |
the i:th delayed neutron fraction |
l |
neutron mean lifetime (here 0.001 sec) |
I |
i:th decay constant (sec-1) |
ci |
concentration of the i:th fraction of the delayed neutrons’ precursors, |
At |
steady state, when time is zero t=0 all time derivatives
are equal to zero, |
all |
d/dt=0 and the initial value of the relative power equals
unity N(0)=1, and also no reactivity perturbation is present k=0 |
Delayed neutron data for thermal fission in
U235 is used
Group |
1 |
2 |
3 |
4 |
5 |
6 |
Fraction ßi |
0.000215 |
0.001424 |
0.001274 |
0.002568 |
0.000748 |
0.000273 |
Decay constant i |
0.0124 |
0.0305 |
0.111 |
0.301 |
1.14 |
3.01 |
The initial values of the delayed neutrons’ precursors
are;
i |
1 |
2 |
3 |
4 |
5 |
6 |
ci(0) |
17.3387 |
46.6885 |
11.4775 |
8.5316 |
0.6561 |
0.0907 |
4. Using the MATLAB
notations
x(1)=N x(2)=c1………… x(7)=c6
the code is
%Save as xprim7A.m
function xprim =xprim7A(t,x,i)
DeltaK=i*0.010*0.50; %voidcoef=i*0.010pcm/percent
void change, void increase 50percent
xprim=[(DeltaK/0.001-6.502)*x(1)+0.0124*x(2)+0.0305*x(3)+0.111*x(4)+0.301*x(5)+1.14*x(6)+3.01*x(7);
0.2150*x(1)-0.0124*x(2);
1.4240*x(1)-0.0305*x(3);
1.2740*x(1)-0.1110*x(4);
2.5680*x(1)-0.3010*x(5);
0.7480*x(1)-1.1400*x(6);
0.2730*x(1)-3.0100*x(7)];
To study the importance of the magnitude of the void coefficient,
it is enough to plot the first colon of the x matrix. The rows
of the x matrix are the time steps.
%Save as ReaktorKinA.m
figure
hold on
for i=0:1:3
[t,x]=ode45(@xprim7A,[0 0.2],[1; 17.3387; 46.6885; 11.4775; 8.5316;
0.6561; 0.0907],[] ,i);
plot(t,x(:,1))
end
hold off
5. The result
Is given in the following plot;
Figure2. Power Increase at the Insertion of 50% Void at Different
Void Coefficients
6. Block diagram
Using the same parameters a block diagram is created
here with SIMULINK
Delayed group 1
Gain1A = 0.2150 Gain1B = 0.0124 Gain1C = 0 0.0124 Integrator1
= 17.3387 (is the initial value of the first delayed group)
Delayed group 2
Gain 2A = …………………….
GainN = -1.502 [= 6.502 – 5 (= the void reactivity perturbation)]
IntegratorN = 1 (is the initial value of N)
The controller is represented with a zero pole block; (s – 1)/s(s – 1)
Here are the; zeros: [1], poles: [0 1], gain: [1]
Here the absorber rods are represented with an amplifier, the
gain is 50
Figure 3. Block diagram of the neutron kinetics
(with six delayed groups) and the automatic control system with
a PID (Proportional and Integrating “1/s” and Differentiating “s”)
controller
In
this case study, a 10 % pcm/% void coefficient is used and the
perturbation is as earlier, a50% void increase. The system response
without an automatic control system is like an avalanche
Figure 4. System response without an effective automatic control
system.
Relative neutron flux (Power) vs. time (sec)
There is of course no construction which can
take a 500 times power increase in 10 seconds. During the Chernobyl
accident the result was a disaster with an elapsed time much
less than 10 sec.
7. Automatic control
In theory one can specify a control system and
a connected absorber rod actuator to team this transient. To
achieve this, the control
action must be extremely fast and must start effectively within
a fraction of a second. However, to realize a mechanical absorber
rod operating device with the required speed is very difficult.
For this example a PID controller is chosen and the absorber
rods are represented with an amplifier which follows the output
of the controller. The result is quit reasonable.
Figure 5. System response with an automatic control system (PID).
Relative neutron flux (Power) vs. time (sec)
The Chernobyl disaster demands many analyses to
really understand what happened there and how to avoid anything
similar in the future. This article is one such analysis.
8. References
University textbooks on nuclear engineering contain the applied
equations. Textbooks on information technology and numerical
analyses contain the applied methods.
To be published, ENS 2008
Computing the Chernobyl Avalanche
Calculation of the neutron flux, fuel and moderator temperature
transients for Research Reactors
Proceeding of the NESTet, Nuclear Engineering Science and Technology,
energy technology Budapest, Hungary 4 -8 May 2008
ENS NEWS, Issue: 2006/13, Neutron Kinetics of the Chernobyl
Accident
www.euronuclear.org/e-news/e-news-13/neutron-kinetics.htm
|