分布式声传感器井中背景噪声统计特性分析

分布式声传感器技术(DAS: Distributed Acoustic Sensor)是一种新近提出的高精度勘探数据采集方法,受制于复杂背景噪声的影响,DAS记录处理水平有待进一步提高.针对上述问题,采用基于替代序列的时间序列平稳性检验方法和Lilliefors检验法对DAS记录中复杂背景噪声的平稳性和高斯性进行研究.用以分析的实际地震数据是按照勘探工业要求采集的井中DAS记录.研究结果表明,DAS背景噪声是非平稳、非高斯随机过程且噪声特性随记录时长发生变化.相关研究成果实现了针对DAS背景噪声特性的可靠认知,对提高DAS记录处理水平具有重要参考价值.     

复杂地层岩石力学参数实时求取方法——以塔里木盆地巴楚隆起为例

非常规油气、干热岩、超深井等领域的钻完井工程对岩石力学服务的及时性与准确性提出了迫切需求,基于矿物及干酪根等岩石骨架组成成分具有固定的岩石物理及岩石力学特性,通过低成本、低风险、高实时性、高采样率的岩石成分录井技术开展岩石力学参数求取方法研究,以偶极子阵列声波计算岩石力学参数的模型作正演,确定通过矿物计算岩石力学参数的可行性,然后以组成矿物的元素成分直接构建岩石力学参数,并对两种方法进行对比。653个复杂岩性数据点,正演法计算的杨氏弹性模量、泊松比的平均相对误差分别为16.3%、9.8%;直接法计算的杨氏弹性模量、泊松比的平均相对误差分别为10.0%、6.8%。结果表明,通过元素直接构建的岩石力学参数具有更高的精度,在岩性相对简单的地区应用效果更好。     

基于行车声音端点检测的交通量统计

基于传统特征的行车声音端点检测法存在重叠有车段识别率低、双门限阈值较难确定的问题,针对这两个问题,探索性地将梅尔频率倒谱系数(Mel frequency cepstral coefficients,MFCC)倒谱距离特征和短时能量特征进行了融合并应用于交通量检测。首先选取了周围环境较为安静的一个双车道路段,并采集了该路段上包含重叠有车段的行车声音;其次提取了行车声音的短时能量特征和MFCC倒谱距离特征,并对它们在端点检测中的优劣进行了分析对比;再次提出了一种融合短时能量特征和MFCC倒谱距离特征的新特征,并基于新特征将传统的双门限判决思路改进成了单门限判决思路;最后利用新特征对有车段进行端点检测并统计交通量。实验结果表明:基于融合特征的端点检测方法能有效解决重叠有车段识别率低和双门限阈值较难确定的问题。     

Elucidating the diel and seasonal calling behaviour of Elachistocleis matogrosso (Anura: Microhylidae)

ABSTRACT

Acoustic monitoring provides the opportunity to study ecological processes that are difficult to assess with traditional surveys. Elachistocleis matogrosso is an anuran species, described in 2010, for which limited biological information is available. This study investigated the calling activity of the species in the north-eastern portion of the Pantanal, Brazil, a wetland area with marked seasonality between the dry and wet seasons. The calling activity of E. matogrosso was monitored using automated digital recorders in combination with automated signal recognition software over two different annual cycles. The species was vocally active only during the wet season (October – April), with a peak in November-December during the 2013–2014 annual cycle and in February-March during the 2015–2016 annual cycle. The peak calling activity occurred at dusk. This species has nocturnal habits and an explosive breeding activity. The detection of the species was intermittent, which suggests that environmental predictors or site-specific conditions might play an important role in species detection. Moreover, this intermittent occupancy indicated that surveys that employ traditional field techniques would likely fail to detect this species. We describe an effective protocol for detecting E. matogrosso with acoustic monitoring, which requires recording during 20 days in February from 17:01 to 05:00. Our procedure would be easy to adapt to other anuran species, and it could be used for investigating new localities and assessing population changes over time.     

Acoustic emission detection of 316L stainless steel welded joints during intergranular corrosion

This study analyzes acoustic emission (AE) signals during the intergranular corrosion (IGC) process of 316L stainless steel welded joints under different welding currents in boiling nitric acid. IGC generates several AE signals with high AE activity. The AE tech-nique could hardly distinguish IGC in stainless steel welded joints with different welding heat inputs. However, AE signals can effectively distinguish IGC characteristics in different corrosion stages. The IGC resistance of a heat-affected zone is lower than that of a weld zone. The initiation and rapid corrosion stages can be distinguished using AE results and microstructural analysis. Moreover, energy count rate and am-plitude are considered to be ideal parameters for characterizing different IGC processes. Two types of signals are detected in the rapid corro-sion stage. It can be concluded that grain boundary corrosion and grain separation are the AE sources of type 1 and type 2, respectively.     

基于特征匹配与假设检验的仿真信号确认方法

为降低仿真信号的应用风险,在使用之前需要再次确认仿真信号。根据装备试验特殊的约束与要求,讨论了仿真信号的特征匹配问题。采用参数假设检验的方法确认仿真信号的特征匹配水平,提出了一种基于特征匹配与假设检验的仿真信号确认方法,并就方法应用相关的细节问题进行了深入探讨。在目标回波仿真中的应用表明,该方法在真实信号实测样本不充分的情况下仍能适用,可为确认仿真信号提供决策支持。     

砂轮磨削过载判据及其快速诊断

基于对砂轮磨削过程中声发射信号的采集,研究了2种砂轮磨削过载判据的特征量提取方法,并提出了评价砂轮磨削过载判据的主要指标即灵敏度、稳定性和算法效率.通过选取合理的采样率与处理间隔,分析比较了2种信号处理方法对砂轮与工件过载的识别效果.结果表明,所提出的信噪特征比方法提取的特征量具有稳定性好、灵敏度和算法效率高的特点,且可以对过载点实现预判,可作为优先选择的磨削过载判据.     

不同插值的有损耗压缩法实现光学扫描全息重现的研究

采用3种不同的插值方法研究了有损耗全息压缩方法实现光学扫描全息的压缩重现过程,结果表明没有主要全息重现细节的丢失,有损耗全息压缩法可以实现压缩比达到100的全息压缩重现.     

应用于助听器反馈抑制系统的变步长归一化子带自适应滤波算法

为减小自适应滤波算法中较快的收敛速度与较低的失调量之间的矛盾,提出了一种变步长归一化子带自适应滤波(VSS-NSAF)算法,并将其用于助听器声反馈抑制系统.该算法基于滤波器系数短时平均和长时平均之间的归一化距离,将滤波器的更新状态分为收敛状态、过渡状态和平稳状态.滤波器可在不同更新状态下自适应地使用不同步长,收敛状态下使用大步长以保证快速收敛,过渡状态下则使用逐步减小的阶梯状步长以进一步降低失调量,平稳状态下使用小步长以保证系统收敛至稳态解.仿真实验结果表明,与传统的归一化最小均方误差(NLMS)算法及其他子带自适应滤波算法相比,所提算法的收敛速度更快,平均稳态失调量更低.     

基于声发射技术的单轴受压混凝土损伤细观研究

为了从细观角度了解不同强度等级,以及不同碳化龄期的混凝土的损伤演化特征,对强度等级分别为C20、C30、C40以及碳化龄期分别为0、28、56、77、90 d的混凝土棱柱体试件进行了单轴受压试验;并利用声发射(AE)技术对试验全过程进行了动态监测,获得了表征混凝土损伤演化过程的典型声发射事件累积计数(burst line cumulate,BLC)曲线;以及声发射事件数-相对应力水平关联曲线.结果显示:①BLC曲线的三个部分分别反映出材料损伤演化过程的三个阶段,曲线的两个转折点分别对应混凝土材料内部的起裂点和失稳点;②在低应力时,不同强度等级以及不同碳化龄期的混凝土试件的声发射特征没有明显区别;随着应力的增加,高强度等级和碳化龄期长的混凝土试件的声发射事件数增长较快,并且声发射活性急剧增加点明显前移.