Estimate of the error in monitoring the energy release distribution on the basis of a conditional probability distribution density

Abstract

It is shown by means of calculations and experiments that the generally accepted practice of using an unconditional instead of a conditional probability density distribution when estimating the error in monitoring the power in the fuel channels with the highest power density in the cores of a wide class of reactors with standard power monitoring error greater than 1.5% results in a systematic underestimation of the margin up to the limits for safe operation by up to approximately 5 and 10% for regimes with and without real-time optimization of the energy release distributions, respectively. This underestimation decreases to zero monotonically when the average power is used instead of the maximum power.Two types of errors are unavoidably encountered in designing and operating nuclear reactors with high power density -a measurement error and a regulation error, most often smoothing and distribution of the energy release and other distributed parameters of the core [1]. Ordinarily, the margin up to the safe operating limit is calculated assuming that the errors are independent of the measured value and the regulation algorithm, though, as shown in [2], this is not so for large measurement errors when corrections due to modeling of the smoothing and measurement errors must be introduced. It is legitimate to examine separately to problems of determining the distribution of the measurement error: for the fuel channel with the highest power density and no concrete coordinate and for each channel of the reactor with different parameter regulation methods, for example, without smoothing or with minimization of the parameter nonuniformity over the core. A detailed examination of the first problem is presented in [2]. The present article is devoted to solving the second problem and generalizing the results obtained.First, we shall examine the results of the solution of the first problem [2], where the errors in monitoring the coefficient of nonuniformity K r of the radial-azimuthal distribution of energy release are determined in regimes without and with smoothing of the energy release distribution in RBMK-1000. However, they can also be used for other reactors with high power density, which are equipped or not equipped with in-reactor monitoring systems. In the latter case, design neutronphysical calculations are used as a means for monitoring. The regime without smoothing was studied analytically and by the method of statistical modeling. The regime with smoothing was studied only using the latter method. The two-dimensional

Stochastic effects when determining the maximum RBMK channel power

“…Taking account of the sensor error the ratio of the maximum to the average power increases to 1.506-1.511. To take account of the error, the sensor indications were varied assuming they follow the normal distribution with standard deviation 4% [2]. Table 2 shows the standard deviations of the reconstructed field from the true field.…”

“…In recent years, a nontrivial assertion has been formulated that the system monitoring the RBMK energy release systematically overstates the maximum channel power by 5% [2]. This assertion pertains not only to the operating but also any other similar operating systems.…”