基于双BP神经网络数据融合的水声定位研究

针对多传感器水声定位的数据融合问题,对目前常用的基于BP神经网络的数据融合算法进行改进,提出了基于双BP神经网络数据融合的水声定位算法.该算法首先将传感器采集回来的数据依次输入到异常数据识别网络,进行训练,剔除异常数据点,再将异常数据识别网络所训练出来的有效数据输入到有效点融合网络,得出最终的定位目标.仿真实验表明,与常规的基于多传感器的水声定位算法相比,该方法在速度和精度上都存在一定优势.     

多源数据融合方法研究

选取工程应用中易于实现的数据融合方法进行分析研究,并以融合数据与数据真实值的偏差作为数据融合方法的稳定性判定依据,进行了三余度多源数据的数据融合仿真试验。试验结果表明,加权最小二乘法在数据融合时表现稳定,可替代常用的矩估计融合和中值融合,更有效地提高融合数据的精度。      

基于数据融合的核动力装置故障诊断方法

数据融合作为一种处理多源信息的方法适合于核动力装置的故障诊断。利用数据融合信息分级处理的思想,将核动力装置故障诊断分为3级进行,数据级采用了数据挖掘的方法对数据进行处理,对属性进行约简;特征级采用并行的3个神经网络处理数据级的约简属性,并将其输出作为决策级 Dempster-Shafer(D-S)证据理论的基本概率赋值;决策级采用了改进的D-S证据理论对神经网络的输出进行合成,克服了传统D-S证据理论无法处理冲突信息的缺陷。运用文献中的相关数据对该方法进行了测试验证,测试结果证实了该方法可正确诊断训练过的核动力装置相关故障,具有一定的应用价值。     

常规弹头爆炸产生声波机理研究

在一次靶场试验中,根据测到的弹头爆炸产生爆光和爆声的时间差进行定位计算,计算爆点到测点的距离和实测结果不一致,根据这一实际情况对常规弹头爆炸产生声波的机理进行研究,分析了常规弹头爆炸瞬间产生声波的过程,从而,得出所测到的爆声的声源不是点源,而是球面源。     

数据融合理论及其在禁核试核查中的应用研究

数据融合旨在将来自多传感器或多源的信息进行协同分析,通过一系列的融合处理,从而得出更为准确可信的结论。概述了数据融合的研究近况,从系统的角度提出了数据融合的定义,初步探讨了其理论模型。在介绍禁核试核查背景的基础上,给出了禁核试核查数据融合的基本概念,理论框架及研究内容。     

分布式三测角定位系统的数据融合

本文对集中式多探测器定位同一目标的数据融合进行了一般性的论述，进而对分布式三测角定位系统推导了一个数据融合方法，这种方法改变系统的硬件配置不可以大大提高系统的定位精度。     

基于指纹定位技术的放射源定位方法研究

放射源定位是放射性物质安全监管的重要内容之一.目前基于移动式探测器的定位方案存在应用条件、部署成本、监控时效等诸多方面的挑战.本文提出一种基于指纹定位技术的放射源定位方法,讨论了基于多探测器的放射源指纹定位系统设计方案,建立了基于环境辐射监测系统的仿真模型,设计了基于WKNN和XGBoost的放射源指纹定位算法并进行了...     

基于LSTM的核电传感器多特征融合多步状态预测

针对核电工况参数预测的问题,利用核电站传感检测系统采集的大量时间序列,提出了基于长短时记忆网络（LSTM）的多特征融合多步状态预测模型。以某核电厂实时参数系统采集到的SG1蒸汽压力传感数据为研究对象,首先针对数据缺失、采样时标不一致问题进行数据预处理,然后完成基于LSTM的多特征融合多步状态预测模型的结构设计与建模,最后将本文提出的预测模型与循环神经网络（RNN）、门控循环单元（GRU）、本文模型-全连接层1以及单变量LSTM等多步预测模型进行比较。实验结果表明,本文提出的预测模型的拟合性能和预测性能整体最优,同时也验证了基于LSTM模型的深度学习方法在核电站运行安全保障领域的适用性。      

基于数据融合的航空γ能谱特征峰信息提取技术研究

针对航空γ能谱测量仪中多路NaI(Tl)探测器中的不同噪声水平,采用加权最小二融合估算法对能谱信息进行提取。该算法通过对NaI(Tl)探测器测量的数据进行在线方差学习,及时调整各探测器的权系数。将该方法用于航空能谱仪在标准模型上采集的数据来进行分析,可以看出,其单秒测量得到的误差范围有所改善,测量结果的准确度有所提高。     

Impact of photoneutrons on reactivity measurements for TMSR-SF1

The solid-fueled thorium molten salt reactor(TMSR-SF1) is a 10 MW_(th) test reactor design to be deployed in 5-10 years by the TMSR group.Its design combines coated particle fuel and molten FLiBe coolant for great intrinsic safety features and economic advantages.Due to a large amount of beryllium in the coolant salt,photoneutrons are produced by(y,n) reaction,hence the increasing fraction of effective delayed neutrons in the core by the photoneutrons originating from the long-lived fission products.Some of the delayed photoneutron groups are of long lifetime,so a direct effect is resulted in the transient process and reactivity measurement.To study the impact of photoneutrons for TMSR-SF1,the effective photoneutron fraction is estimated using k-ratio method and performed by the Monte Carlo code(MCNP5) with ENDF/B-Ⅶ cross sections.Based on the coupled neutronphoton point kinetics equations,influence of the photoneutrons is analyzed.The results show that the impact of photoneutrons is not negligible in reactivity measurement.Without considering photoneutrons in on-line reactivity measurement based on inverse point kinetics can result in overestimation of the positive reactivity and underestimation of the negative reactivity.The photoneutrons also lead to more waiting time for the doubling time measurement.Since the photoneutron precursors take extremely long time to achieve equilibrium,a "steady" power operation may not directly imply a "real" criticality.     

求解点堆动态方程的IGear方法

求解点堆中子动力学方程的改进的Gear方法——IGear,以其高阶和大稳定域的性质保持了Gear方法适应性强,计算精度较高的优点,并改进了它的缺点。因此能够灵活应用于各种反应性的输入。计算结果表明:从计算时间和计算精度上看,该方法都可以满足实际工程的需要,是求解点堆动态方程的一种较好的方法。     

Estimation of time-dependent neutron transport from point source based on Monte Carlo power iteration

This study proposes a method for calculating time-dependent neutron transport from a point source with a continuous-energy Monte Carlo code. To deal with neutron multiplication and attenuation in orders of magnitude, the power iteration method conventionally used to estimate the effective multiplication factor k_eff was utilized. The time of a neutron flying in a cycle from emission of its ancestor at the point source was estimated. In the estimation, the decay time of the delayed neutron precursor was considered. The neutron flux was tallied in time bins in each cycle. The source strength in the cycle was considered as the product of k_eff estimators from the first to the previous cycle. By summing up the tallied flux multiplied by the strength, the neutron flux variation with time was obtained. This method was verified against a UO₂ fuel lattice moderated and reflected by light water.     

点通量积分法在蒙特卡罗方法中的应用

介绍了指向概率法和点通量积分法两种方差减小技术在计算小概率的体通量、面流量中的应用,进而利用点通量积分法计算了在MCLLS方法的典型几何结构中的面流量计算结果并且与通用蒙卡程序MCNP的计算结果进行了比较。     

Wide-range characteristics of beam perveance and saddle point potential of LIPS-200 ion thruster

Both the long-life and multi-mode versions of LIPS-200 ion thruster are under investigation in LIP (Lanzhou Institute of Physics). To confirm the feasible ranges of the beam current and accel (abbreviation for accelaration) grid potential to apply to the thruster, the wide-range beam perveance (the state of beam focus) and saddle point potential (the lowest potential along beamlet centerline) characteristics of LIPS-200 are studied with a test-verified PIC-MCC (Particle in Cell-Monte Carlo Collisions) model. These characteristics are investigated with both the initial and the eroded states of the accel grid aperture diameter. The results show that the feasible ranges of these parameters with respect to perveance/crossover (overfocused) limit extend as the operating time accumulates, while the feasible range of accel grid potential narrows due to a reduced EBSF (electron backstreaming failure) margin. The feasible ranges determined by the initial condition are: (i) the beam current up to 0.981 A, and (ii) the accel grid potential up to −85 V. A 23% enlargement of the aperture diameter would bring up to 48 V of EBSF margin loss.     

Impact of edge localized mode on photo-elastic-modulator based motional Stark effect polarimetry

Intermittent increase in background light induced by edge localized mode (ELM) in high confinement mode (H-mode) plasma gives disturbances to polarization angle of incident light measured by motional Stark effect (MSE) polarimeter using photo-elastic modulator (PEM), even if the background light is not polarized. The error in the polarization angle γ_PEM determined by the light intensity at the frequency modulated by the PEM originates in a broadband frequency spectrum of the intermittent background light and applies to PEM based polarimeter in general. The impact of the background light induced by the ELM on the error in γ_PEM is numerically investigated in terms of the polarization angle of incident light γ, the peak intensity of background light normalized by the intensity of beam emission X_ELM, ELM frequency f_ELM, ELM decay time τ_ELM and the time constant of low-pass filter (LPF) per single stage τ_LPF. The error increases almost proportionally with γ between ±22.5° when inphase spectral amplitude is used in γ_PEM calculation. The error becomes quite large for γ approaching to the direction of PEM axis (±22.5°), when absolute spectral amplitude is used. The error increases with X_ELM but does not depend on τ_ELM. When the ELM frequency is sufficiently lower than a critical ELM frequency that is given by inverse of an effective LPF time constant (τ_LPF times the number of LPF stages), the error has its peak, the height of which is independent of f_ELM, soon after the ELM and vanishes between subsequent ELMs. However, when f_ELM is higher than the critical ELM frequency, errors induced by subsequent ELMs pile up and the error increases with f_ELM. In case of no pileup, the maximum error becomes larger for smaller τ_LPF, because impact of ELM temporally concentrates in a shorter period just after the ELM. The error induced by ELMs is roughly estimated to be ∼0.002° in a type-I ELMy H-mode JT-60U discharge, which is much smaller than other errors coming from calibrations (about 0.1-0.2°) and temporal fluctuation (less than about 0.05°), mainly since the background light induced by ELMs is weak compared to the beam emission X_ELM ∼0.07 in JT-60U. However, MSE diagnostics experiencing large background light (X_ELM) should take this kind of error into account.     

严重事故有限影响准则的环境影响系数计算方法研究

欧洲用户要求（EUR）提出的有限影响准则（CLI）常被用于针对严重事故的放射性释放验收标准,由于EUR并未公开关于CLI相关环境影响系数的确定方法,国内相关机构对其进行了尝试性研究,并给出了计算CLI相关系数的计算公式和计算结果。为了制定适合中国厂址的CLI,本文在国内相关研究成果的基础上,通过对CLI设计理念进行探讨,对CLI相关系数的确定方法进行了修正,并提出了一种新的确定CLI相关环境影响系数的方法,该方法确定的环境影响系数与EUR给出的相关系数相比具有一定的相似性,该方法可进一步推广至其他国内厂址,为制定适用于国内厂址的CLI提供参考。     

基于DSP技术的多道核脉冲幅度分析器的设计

介绍以DSP2407芯片为核心,采用专用峰值幅度采样保持电路,根据输入信号上升沿及峰值信息进行分析的多道核脉冲幅度分析器系统的原理和硬件结构.在合金分析及同位素密度成像系统中,采用本脉冲幅度分析器进行高速数据采集,取得了满意的效果.也可适用于其他核脉冲信号峰值幅度分析的场合.     

基于蒙卡-点核积分耦合的核电厂退役辐射场计算

针对核电厂退役现场中复杂和简单几何区域并存的问题,以及退役过程对辐射场计算提出的快速、准确的要求,提出了蒙卡和点核积分耦合计算方法,并开发了耦合计算程序。对于单体源模型,由于模型简单,耦合方法、点核方法与蒙卡方法的计算结果均较接近,相对偏差小于10%,但耦合方法的计算速度比蒙卡方法提高了3.5倍;采用耦合方法和蒙卡方法分别计算了多体源模型的辐射场,结果表明耦合方法的计算准确度较高,相对蒙卡方法偏差小于40%,计算时间约为蒙卡方法的1/5。以秦山一期退役辐射场为例开展了计算分析,结果显示,耦合方法与蒙卡方法的计算结果相差不超过1个数量级,但前者计算速度比后者提高了约3倍,说明耦合方法可以满足核电厂退役辐射场计算需要。