[en] Some attempts have been made to investigate noise sources in a boiling water reactor (BWR) by multivariate random data analyses. Autoregression and multivariate coherency such as partial and/or multiple coherency have been introduced to the analysis of time series data gathered from a medium-sized BWR plant (BWR-3) of 460-MW electric power to evaluate linear relations among multiple inputs and outputs that are coupled with each other by sophisticated feedbacks. Through these attempts, the main local disturbance that leads to the peak in the spectrum of reactor power noise and is classified as global noise has been concluded to be caused by noise sources originated, not outside the reactor core, but inside the reactor core itself. Furthermore, the noise sources in the core have been found to be the turbulence of bubble generation and extinction in the lower region of coolant flow channel. It is found that the noise sources have different resonant frequencies that depend on the running speeds of coolant flows in fuel assemblies near the bottom local detector. It can also be shown that pressure waves induced by the local disturbances propagate into the coolant water in the lower core plenum, where they are mixed together into a single-pressure wave whose resonant frequency corresponds to the peak frequency in the spectrum of reactor power noise

[en] Multivariate autoregressive (AR) procedures are introduced as diagnostic tools to extract dynamic characteristics for detection of malfunctions of a boiling water reactor (BWR) power plant. The problem of estimating AR matrices is equivalent to identifying, from measured random signals of a BWR station, the dynamic parameters of a stationary linear discrete time system derived from an unmeasured uncorrelated white-noise process. To explain the characteristics of a derived AR spectra, a general multiple-input, single-output model is discussed. The experiments were carried out in a 460-MW(e) BWR station. The power spectral density of the averaged neutron flux is decomposed into terms corresponding to sources of noise at points of measurement where the origin of the noise in the neutron fluctuation is studied. It is shown from the analysis that a disturbance of high intensity in neutron fluctuation of the BWR is not caused by the process variables such as core flow but is possibly caused by the inherent noise, specifically defined in the paper, of the neutron flux itself

[en] A method for real-time stability monitoring is developed for supervising the steady-state operation of a boiling water reactor core. The decay ratio of the reactor power fluctuation is determined by measuring only the output neutron noise. The concept of an inverse system is introduced to identify the dynamic characteristics of the reactor core. The adoption of an adaptive digital filter is useful in real-time identification. A feasibility test that used measured output noise as an indication of reactor power suggests that this method is useful in a real-time stability monitoring system. Using this method, the tedious and difficult work for modeling reactor core dynamics can be reduced. The method employs a simple algorithm that eliminates the need for stochastic computation, thus making the method suitable for real-time computation with a simple microprocessor. In addition, there is no need to disturb the reactor core during operation. Real-time stability monitoring using the proposed algorithm may allow operation under less stable margins

[en] An on-line method of predicting reactor water level and pressure under small loss-of-coolant-accident (LOCA) conditions has been proposed. The method features initialization of a simplified reactor model in transient conditions. This is done by using plant data in time series and estimating unknown parameters, such as break area, by a nonlinear optimization method. Off-line simulations were performed for small LOCAs (break area: about100 cm²) using the results of a reference LOCA analysis code for plant data. These showed that the difference from the reference calculation during a 10-min period was within 10%, normalized by the value at the initialization time, and the estimation error for break area was below 11%. The overall CPU time required for the prediction was below 10 s in a general purpose computer of 10 million instructions per second

[en] This paper summarizes activities and experiences concerning noise measurement and analysis in the BWR power plant SHIMANE I since its beginning of operation, of the Hitachi Atomic Energy Research Laboratory (HAERL) in collaboration with Chugoku Electric Power Company. The possibility of using noise analysis for safety monitoring, detecting abnormalities at an incipient stage and further, diagnosing the abnormal condition is discussed. The discussion on noise sources in reactor power fluctuation during normal operation is also briefly summarized. (author)

[en] A multivariable autoregressive (MAR) method is applied to the core stability estimation of a boiling water reactor-5 operation. Noise data measured during steady-state operations at startup tests are used. In this method, the closed loop transfer function from reactor pressure to reactor power is identified from reactor noise data and transformed into an impulse response function. The decay ratio representing stability characteristics is evaluated from this function. The variation range of decay ratio estimates obtained by this method is sufficiently small, if the analyzing conditions are appropriately selected. The value of the decay ratio is 0.23 during natural circulation and decreases with core flow, reaching close to zero at the rated power. A similar power dependence for the decay ratio is seen in results from a core stability analysis code. The MAR method is a useful tool for stability estimation, even if no external disturbance tests are conducted

[en] A computerized operator support system for post-trip transient control of BWRs is described which has the purposes of anomaly occurrence diagnosis, reactor state estimation and prediction in the transient, and provision of advisable operational tactics for operator's use to lead the reactor to safe shutdown. In order to fulfill these purposes, the system involves the following major functions: the event diagnosis function, the operator guidance function with an event based method followed by a symptom based method for enhancing guidance reliability, the reactor state prediction function using a BWR dynamic simulator of faster than real time phenomenon, and the state estimation function using a model adjustment method. The usefulness of the system is discussed through the simulation tests of these functions. (author)