D. Wade

Argonne National Laboratory,
United States of America

Technical specifications that govern plant operations require that active safety systems be periodically validated and/or recalibrated as a means to assure that they continue to perform their required safety function. Passive safety features are subject to ageing phenomena over the multidecade life of the plant, and so a means is needed to periodically reconfirm that they also remain always capable of performing their required safety function.

The means to accomplish this reconfirmation is specific to the safety function being performed and to plant design, but the philosophy of periodic checking of passive safety features under technical specification requirements can be illustrated for the specific case of liquid metal cooled fast reactors that rely on a reactor vessel auxiliary cooling system (RVACS) for passive decay heat removal and thermo-structural reactivity feedbacks to self-regulate power output to match externally imposed heat removal rates.

First, in the case of the RVACS, performance degradation might occur due to partial clogging of ambient air circulation channels with dust, rodent nests, flooding of the lower regions of the ducting, etc.; additionally, changes of emittance properties of radiation surfaces due to oxidation or dust layers, etc., might increase heat transport impedance. Continuous heat balances on the always operating RVACS heat rejection rate can be performed in a completely straightforward manner by monitoring air flow rate and temperature rise versus reactor power level. The heat balance instrumentation will, of course, require periodic recalibration in its own right.

The thermo-structural reactivity feedbacks that govern power self-regulation are integral feedbacks which depend on temperature profiles in the reactor; they affect reactivity directly through Doppler and density coefficients of reactivity and indirectly through structural displacements which affect neutron leakage rates. Their components change versus burnup and age due to changing fuel composition and due to structural relaxations of core support structure, core clamping mechanisms, and creep of the fuel wrapper. Periodic reconfirmation to show that thermo-structural feedbacks remain in the range necessary to assure passive matching of power to external heat removal rate rests on the fact that such feedbacks are composite feedbacks with respect to externally controllable variables. These externally controllable variables are the inlet coolant temperature, the forced circulation flow rate, and the reactivity vested in control rods. Specifically, asymptotically — after transients die away — normalized power, P, depends on these external variables via a quasi reactivity balance as:

AP s 0 = (- 1)A + ^P-ljВ + 8TinC + Apext

where F is the normalized primary flow rate, and STin is change in coolant inlet temperature from its operating value.

Integral reactivity coefficients A, B, and C have the following physical interpretations:

— C is the reactivity vested in the deviation of core inlet temperature from its nominal value;

— B is the reactivity vested in the coolant average temperature rise above the coolant inlet temperature;

— A is the reactivity vested in the fuel average temperature rise above the coolant average temperature.

They are measurable in-situ on the operating power plant in a non-intrusive way by introducing step changes in flow rate, coolant inlet temperature and external (rod) reactivity and then measuring the asymptotic

• If Dpext is changed while inlet temperature and flow remain fixed, the power will asymptotically self-adjust to:

= 1 + — AP ext / B

1 + A / B

• If flow rate is changed while inlet temperature and Apext remain fixed, the power will asymptotically self­adjust to:

P

P = 0 B’

A +

ё FU

• If inlet temperature is changed, ST, while Dpext and flow rate remain fixed, the power will asymptotically self-adjust to:

This procedure would yield three equations for the three unknowns, A, B and C, which would determine their current values on the operating reactor itself. The efficacy of such measurements in determining the values of A, B, and C on an operating reactor connected to the grid was demonstrated [2] at EBR-II.

Some small and medium sized reactors rely on natural circulation in which case flow, F, is not externally controllable, but instead is a function of power F= f(P). Assuming f(P), it could be represented as a quadratic:

F = a + bP + cP2

Several additional step changes in Dpext and/or dTinlet would be sufficient to determine the values of A, B, and C.

More elegant methods have been developed based on continuous monitoring and noise analysis techniques — taking advantage of spontaneous fluctuations or small purposeful power spectral density inputs to the externally controlled state variables.

These examples for liquid metal cooled fast reactors illustrate the approach that can be taken for periodic reconfirmation of the ability of passive safety features to perform their safety function. Other reactor types with different passive features may employ alternative approaches.