Determination of the MTC by noise analysis methods

Paper i proceeding, 2003

Determination of the moderator temperature coefficient (MTC) in pressurised water reactors has been the matter of interest for a long time. The MTC should be within certain limits at all times; it should be negative, but its absolute value should not exceed a certain limit either. Since the MTC changes continuously during the fuel cycle with burnup and the corresponding change in the boron concentration in the water, a monitoring of its value is necessary. The traditional methods that are used to measure the MTC are time consuming and expensive. The noise based methods are more flexible but so far they were rather inaccurate with a systematic underestimation of the MTC. In recent work we have pointed out that the main weakness of the noise based method is the poor knowledge of the driving source, by measuring the temperature fluctuations only in one radial point whereas those fluctuations are highly spacedependent. The problems of the noise method can be eliminated or largely improved if the core average temperature is used instead of the local temperature, leading to an improved noise estimator. This fact was confirmed by us in detailed numerical simulations. The new estimator was also tested in measurements at an operating plant. In the Swedish Ringhals-2 PWR, 12 strings of gamma thermometers are permanently installed. In the frequency range of interest to the MTC, i.e. within 0.1 - 1 Hz, these sensors act as pure noise thermocouples, measuring the temperature fluctuations of the cold junction which is in the coolant. With the signals of the gamma thermometers the fluctuations of the core averaged temperature can be estimated accurately. A full measurement was performed in 2002 with gamma thermometers taken in all 12 radial points at one axial elevation, 2 in-core neutron detectors, and one core exit thermocouple. The purpose of the present paper is to report on these measurements. It was shown that with the suggested noise estimator, based on the core average temperature fluctuations, the correct value of the MTC was obtained, by comparison to SIMULATE calculations. It was seen that using only one in-core gamma thermometer the estimated value of the MTC was in significant error, and using one single core exit thermocouple yielded a yet larger error. The performance of the new MTC estimator was also investigated in view of some parameters of the evaluation such as the FFT procedure used and the neutronic weight function used in the core averaging of the temperature. Some further improvement of the method is also touched upon.