Cross-correlation detection of nonstationary signals due to a leak in the first loop of a nuclear power plant

A leak-detection system employing high-temperature microphones has been developed for RBMK and ATR (Japan) reactors. Further improvement of the system focused on using cross-correlation analysis of the spectral components of the signal to detect a small leak at an early stage of development. Frequency-time cross analysis was used to analyze nonstationary signals due to a leak, and the wavelet transformation was used to single out the envelopes corresponding to different spectral components of the signal. Since envelope processes are less affected by distortions than are wave processes, they give a higher-degree of correlation and can be used to detect leaks with lower signal/noise ratios. Many simulation tests performed at the Fugen nuclear power plant with an ATR reactor have shown that the proposed methods can be used to detect and find the location of a small leak. __________ Translated from Atomnaya énergiya, Vol. 102, No. 3, pp. 183–188, March, 2007.     

Development of microphone leak detection technology on Fugen NPP

A method of leak detection, based on high-temperature resistant microphones, was originally developed in JNC to detect leakages with flow rates from 1 m³/h to 500m³/h. The development performed on Fugen is focused on detection of a small leakage at an early stage. Specifically, for the inlet feeder pipes the leak rate of 0.2gpm (0.046m³/h) has been chosen as a target detection capability. Evaluation of detection sensitivity was carried out in order to check the capability of the method to satisfy this requirement. The possibility of detecting and locating a small leakage has been demonstrated through the research.     

毛细管中气溶胶漏率与气体漏率相关性研究

在气溶胶泄漏评估中,需深入分析气溶胶漏率与气体漏率的相关性。本文归纳总结了气体漏率和气溶胶漏率的计算方法,对不同条件下毛细管中的亚微米级气溶胶漏率进行了实验研究,分析了上游数浓度、气体流速和上游压力对气溶胶泄漏的影响。结果表明：当气体漏率高于10^-4Pa·m³·s^-1时,气溶胶漏率和气体漏率成线性关系;当气体漏率低于10^-4Pa·m³·s^-1时,气溶胶漏率的降低速率较气体漏率的更迅速,对于亚微米气溶胶,扩散沉积是造成气溶胶损失的主要机制。为方便使用,本文提出一个在一定假设条件下利用气体漏率和漏孔长度对漏孔中的气溶胶漏率进行估算的方法。     

Techniques for processing remote field eddy current signals from bend regions of steam generator tubes of prototype fast breeder reactor

Steam generator (SG) is one of the most critical components of sodium cooled fast breeder reactor. Remote field eddy current (RFEC) technique has been chosen for in-service inspection (ISI) of these ferromagnetic SG tubes made of modified 9Cr–1Mo steel (Grade 91). Expansion bends are provided in the SGs to accommodate differential thermal expansion. During ISI using RFEC technique, in expansion bend regions, exciter–receiver coil misalignment, bending stresses, probe wobble and magnetic permeability variations produce disturbing noise hindering detection of defects. Fourier filtering, cross-correlation and wavelet transform techniques have been studied for noise reduction as well as enhancement of RFEC signals of defects in bend regions, having machined grooves and localized defects. Performance of these three techniques has been compared using signal-to-noise ratio (SNR). Fourier filtering technique has shown better performance for noise reduction while cross-correlation technique has resulted in significant enhancement of signals. Wavelet transform technique has shown the combined capability of noise reduction and signal enhancement and resulted in unambiguous detection of 10% of wall loss grooves and localized defects in the bend regions with SNR better than 7 dB.     

管道贯穿裂纹泄漏率预测

充分考虑入口为过冷水、两相和过热蒸汽等情况下的临界和非临界流动,对摩擦压降、拐角压降以及临界流模型Henry-Fauske中的相变起始点进行了修正,开发出破前泄漏(LBB)的管道贯穿裂纹泄漏率计算程序。与目前通用的裂纹泄漏率程序(PICEP和SQUIRT)分别就常规直管道、人工裂纹以及自然形成裂纹的泄漏率实验数据进行了对比。计算分析表明,该程序计算结果与实验结果吻合较好且精度高于PICEP和SQUIRT。此外,通过计算分析研究,获得了背压和滞止焓以及裂纹形貌(包括局部粗糙度、全局粗糙度及裂纹开口位移、裂纹拐角数目等)对泄漏率的影响规律。     

管道系统泄漏检测神经网络与模式识别方法

提出了以管道系统泄漏后形成多相湍射流所引发的应力波信号时域和频域特征指标构造神经网络输入矩阵，建立对管道运行状况进行分类的神经网络模型以检测管道泄漏故障的发生。并提出以波峰、波谷、水平线、主导峰、支配强度、从属度等模式基元抽取负压波形特征，采用上下文无关文法对管道负压波进行描述，进而建立了管道负压波形结构模式分类系统，用于区别管道正常状态和泄漏状态。实验研究了这些新理论的有效性。     

Acoustic method of leak detection using high-temperature microphones

A leak-detection method based on the use of high-temperature microphones was initially developed for ATR. The next-generation system was developed for RBMK entry and exit piping. Even though there are structural differences between them, RBMK and ATR have the same type of loop cooling. This has made it possible to develop a general approach to coolant leak detection. The microphone system of leak detection has been installed and successfully operating at the Leningrad nuclear power plant. The system detects leaks with sensitivity 0.23 m³/h, and it determines the site of a leak and estimates its size. The experience gained at the Leningrad nuclear power plant has subsequently been used at the Fugen nuclear power plant, where 0.046 m³/h was chosen as the target sensitivity for detecting leaks in the ATR entry piping. The possibility of detecting and determining the site of a small leak by the method proposed has been demonstrated in the present work.__________Translated from Atomnaya Énergiya, Vol. 98, No. 2, pp. 98–105, February, 2005.     

Evaluation of methods for leak detection in reactor primary systems

No currently available, single leak-detection method combines optimal leakage detection sensitivity, leak-locating ability, and leakage measurement accuracy. Technology is available to improve leak detection capability at specific sites by use of acoustic monitoring. However, current acoustic monitoring techniques provide no source discrimination (e.g., to distinguish between leaks from pipe cracks and valves) and no leak-rate information (a small leak may saturate the system).Seven cracks, including three field-induced IGSCC specimens and two thermal-fatique cracks, have been installed in a laboratory acoustic leak detection facility. The IGSCC specimens produce stronger acoustic signals than the thermal-fatigue cracks at equivalent leak rates. Despite significant differences in crack geometry, the acoustic signals from the three IGSCC specimens, tested at the same leak rate, are virtually identical in the frequency range from 300 to 400 kHz. Thus, the quantitative correlations between the acoustic signals and leak rate in the 300–400 kHz band are very similar for the IGSCC specimens. Also, acoustic background data have been acquired during a hot functional test at the Watts Bar PWR. With these data, it is now possible to estimate the sensitivity of acoustic leak detection techniques.     

Evaluation Method of Statistical Error in. Transit Time Measured with Cross-Correlation of Two Detector Signals

In the transit time method for measuring flow velocity, the cross-correlation function of two detector signals is computed; the delay time which maximizes the function gives the transit time of the fluid between the detectors. The cross-correlation function is mathematically defined for an infinite observation time, however actually it is computed based on a finite length of time. Due to this approximation, the measured cross-correlation function has a statistical uncertainty, which induces a scattering on the measured transit time.

In this paper, it is shown that the statistical error of the transit time may be expressed by using either the derivatives of correlation functions or the power spectral densities. The spectral expression has a merit of giving an intuitive understanding of the nature of the error and it can be seen that the error becomes smaller as the higher frequency components of the spectra increase. The error also becomes smaller when the coherence function has a larger value and/or the observation time extends for a longer period.

A simplified evaluation method of the error is proposed using a trapezoidal approximation for the power spectra and this method is applied to the transit time measurement by electro- magnetic-flowmeters. The evaluated statistical error closely agreed with the measured scattering.     

示踪技术法检测反应堆堆厅气体渗漏率

采用SF₆示踪技术检测某反应堆3个堆厅在常温常压下的气体渗漏率,考察堆厅渗漏气体在主要工作场所的累积效应和到达建筑物内人员疏散通道关键部位的时间和累积强度。结果表明,3个反应堆堆厅在检测条件下的气体渗漏率分别为（7.30±0.16）×10^-4、（1.88±0.12）×10^-4和（2.07±0.07）×10^-4h^-1。堆厅渗漏气体在各工作间以较快速度积累,5h左右在工作间内达到极值,在一楼更衣室内的累积效应明显;堆厅渗漏气体在人员通道的累积幅度较小,约10h达到极值浓度。     

强背景噪声下反应堆内冲击信号的特征提取

压水反应堆主冷却剂系统中松动部件的在线监测对核电厂的安全运行至关重要,但监测信号往往会受到流致振动和其他设备运行而产生的强背景噪声的干扰。为增强信号的冲击特征,本文提出了一种基于变分模态分解（VMD）和小波包变换（WPT）相结合的信噪分离和特征提取方法。首先,采用VMD算法将含噪声的冲击信号分解成不同频率成分的本征模态函数（IMF）,并基于各模态函数间的相关系数确定分解过程的模态数量;然后,利用峭度和相关系数构建加权峭度指标,并依据加权峭度指标选取IMF,重构冲击分量较强的新信号;最后,利用WPT算法进一步对新信号进行去噪处理。采用所提出的算法对仿真模拟和冲击实验获取的信号进行特征提取,均成功分离出冲击分量,验证了该方法的有效性。     

Development of methodology to optimize management of failed fuels in light water reactors

Fuel cladding is one of the key components in a fission reactor that confines radioactive materials inside a fuel tube. During reactor operation, however, cladding is sometimes breached, and radioactive materials leak from the fuel pellet into the coolant water through the breach. The primary coolant water is therefore monitored so that any leak is quickly detected; coolant water is periodically sampled, and the concentration of radioactive iodine 131 (I-131), for example, is measured. Depending on the measured leakage concentration, the faulty fuel assembly with leaking rod is removed from the reactor and replaced immediately or at the next refueling. In the present study, an effort has been made to develop a methodology to optimize the management for replacement of faulty fuel assemblies due to cladding failures using measured leakage concentration. A model numerical equation is proposed to describe the time evolution of an increase in I-131 concentration due to cladding failures and is then solved using the Monte Carlo method as a function of sampling rate. Our results indicate that, to achieve rationalized management of failed fuels, higher resolution to detect a small amount of I-131 is not necessarily required, but more frequent sampling is favorable.     

速调管走廊抗干扰技术研究

合肥国家同步辐射实验室速调管走廓内存在着高频电磁泄漏，为确保走廓内的通讯信号不被干扰，采用抗干扰技术对泄漏噪声进行了抑制，并对抑制前后的泄漏噪声作了对比测试，结果显示了所采用的抗干扰技术的有效性。     

蒸汽发生器二次侧^16N迁移时间的计算模型

用测量蒸汽中＾１６Ｎ的放射性活度来监测蒸汽发生器传热管的泄漏，是一项新的技术，该监测系统确定泄漏率的一个重要参数是＾１６Ｎ在蒸汽发生器二次侧的迁移时间，本文认为＾１６Ｎ泄漏工质是以汽相形成随二次测工质运动，根据蒸汽发生器二次侧工质的流动特点，将迁移的时间分成四段计算，并重点提出了管束区汽相运动的速度分布计算模型，用本文模型对秦山核电厂蒸汽发生器的＾１６Ｎ迁移时间进行了计算，并与法国电力公司的计算结     

On-Line Leak Detection Method for OWL-1 Loop by ARX Modeling Using Dewpoint Signals

Model identification technique based on ARX (autoregressive model with exogenous variable) process was applied to dewpoint data recorded at OWL-1 (Oarai Water Loop No. 1) loop cubicle in JMTR (Japan Materials Testing Reactor) and the dynamical interrelationship between the supply and exhaust dewpoints in the ventilation system of the cubicle was empirically determined.

It was shown that the information so derived on the dewpoint dynamics can assist to enhance the sensitivity of leak detection, if it was incorporated into a leak monitoring system for the OWL-1 loop.

A simple digital filter incorporating the dewpoint dynamics was designed in an attempt to develop an efficient leak monitor for the OWL-1 loop. This filter was applied to the dewpoint data recordings during an abnormal leak that had occurred at the OWL-1 loop in the 43 rd cycle of JMTR operation, which demonstrated the effectiveness of the present method for leak detection at its early stage.     

Small Leak Detection by Measuring Surface Oscillation during Sodium-Water Reaction in Steam Generator

Small leak sodium-water reaction tests were conducted to develop various kinds of leak detectors for the sodium-heated steam generator in FBR. The super-heated steam was injected into sodium in a reaction vessel having a sodium free surface, simulating the steam generator. The level gauge in the reaction vessel generated the most reliable signal among detectors, as long as the leak rates were relatively high. The level gauge signal was estimated to be the sodium surface oscillation caused by hydrogen bubbles produced in sodium-water reaction.

Experimental correlation was derived, predicting the amplitude as a function of leak rate, hydrogen dissolution ratio, bubble rise velocity and other parameters concerned, assuming that the surface oscillation is in proportion to the gas hold-up. The noise amplitude under normal operation without water leak was increased with sodium flow rate and found to be well correlated with Froud number. These two correlations predict that a water leak in a “MONJU” class (300 MWe) steam generator could possibly be detected by level gauges at a leak rate above 2g/sec.     

Application of an eddy-current type flowmeter to void detection at the LMFBR core exit

To apply an eddy-current type flowmeter to void (bubble) detection, out-of-pile test has been carried out in a sodium loop, and the void detection technique is applied to the subassembly outlet flowmeters of JOYO. The experimental apparatus used 7 simulated fuel subassemblies under 500°C sodium conditions with a volumetric void fraction of up to 2%. It was found that the phase of the void signal (vector) is different from that of the flow signal (vector). Therefore, the signal-to-noise ratio of void detection was highly improved by choosing the optimum phase angle for synchronous detection. At the optimum phase, the flow-induced fluctuation signal (background) is minimum and the void signal nearly maximum. In addition to the improved void-detection technique, basic information about amplitude probability density functions and power spectral densities of void signals are presented. Using the flowmeter installed in the core of JOYO, void-detecting characteristics have been studied.     

A method for weak impact signal discrimination based on para-approximate entropy

Impact signal of the loose part in a nuclear power plant provides vital information about its mass and location. Hence detection of the impact signal is of great significance. In a practical NPP working environment, some weak impact signals may submerge in background noises. A signal detection and impact signal discrimination method is proposed in this paper. Teager energy operator is used to process a whitened signal using auto-regression model. It increases the signal to noise ratio (SNR) and makes the impact signal detection easy. Para-approximate entropy method is proposed to process the energy envelop for discriminating the loose part impact signal from other interferences. Simulated experiment has been carried out to verify the proposed method. Data analysis proves that it is capable to detect and discriminate a weak impact signal with low SNR.     

Acoustic Detection for Small-Leak Sodium-Water Reaction

Characteristics of acoustic signal produced by sodium-water reaction due to steam injection and by Ar gas injection into sodium were experimentally investigated. Acoustic signal was measured by using Kistler 808A and 815A5 accelerometers. Root mean square (RMS) measurements and frequency analysis of the signal were conducted.

The RMS measurements could detect a small water leakage into sodium, as small as 0.07 g/sec, in the present loop. The peaks in a frequency spectrum were caused by the natural vibration of a rod on which the acoustic transducer was mounted.

The RMS was approximately proportional to the one-third power of the steam leak rate and increased to some extent with the ambient sodium temperature. RMS values, both for sodium-water reaction and Ar gas injection, were about the same order of magnitude, when the data were plotted against the volumetric flow rates of steam and Ar gas.     

Dynamic analysis of the NPP pressure sensor system

The aim of the work was to develop on-line methods of control and diagnostics of pressure sensors at a NPP. The analysis was carried out along two lines:

1. i) The detection system including the sensor itself was modeled theoretically to obtain and study its transfer function, and to establish correspondence between the spectral characteristics of the pressure signal.

2. ii) The numerical processing of the signal using the AR technique to reconstruct the transfer function and evaluate the system's response, to a step impulse, as well as the spectral analysis. The hydraulic model presented indicate that the spectral lines observed at 10 Hz are due to longitu-dinal oscillations of water in the system branches, while the response time of the detection system is effectively the time of signal retardation in the porous ceramic diaphragms of the δ-cell of the sensor itself.