基于概率法的电力系统安全分析

为应对电力系统安全分析中的停机问题，基于概率法的方式，将常用的确定停机计算与加入了概率法的概率停机进行比较，研究了二者的区别与其在长期投资方向的不同。在进行电力系统停机分析时，通常会分别从确定停机与概率停机的角度出发，对其应急状态下的潮流进行计算。但前者的方法可能导致极低概率的停机事件被忽略，进而影响长期的资金投资。通过加入概率法的计算，使得对单个停机事件的判定由其具体的频率来确定，增加了系统运行的稳定性。     

基于误差估值累加开环校正的诱导式欺骗检测方法

诱导式卫星欺骗干扰可诱导航空器逐渐偏离预定航迹，难以被发现，因此及时有效地检测干扰是飞行安全的保障。在现有紧组合导航体制基础上，设计了一种基于误差估值累加开环校正的紧组合导航结构，并证明了其性能与传统闭环校正紧组合导航性能等效。在此结构中，将紧组合导航系统与自适应序贯概率比检测方法结合，提出了一种基于误差估值累加开环校正的诱导式欺骗检测方法，融合紧组合导航信息与其他不受欺骗影响的导航信息，构建欺骗检测统计量进行诱导式欺骗检测。仿真结果表明，开环校正结构可避免随时间累加的惯性导航系统误差所导致的组合导航滤波器发散问题，同时欺骗检测方法可进一步提高算法对“最坏”情形下微小诱导式欺骗的检测效果。     

Combination of dynamic features with a new mask to optimize neural network speech enhancement

Concerning the problem that the Neural Network speech enhancement algorithm cannot fully represent the nonlinear structure of speech due to feature selection,which leads to speech distortion.This paper proposes the combination of dynamic features with a new mask to optimize neural network speech enhancement.First,three features of noisy speech are extracted and spliced to obtain static features.Then,the first and second difference derivatives are obtained to capture the instantaneous signals of speech and fuse them into dynamic features.The combination of dynamic and static features completes internal complementarity of features and reduced speech distortion.Second,in order to enhance the intelligibility and clarity of speech at the same time,an adaptive mask is proposed,which can adjust the energy ratio of speech and noise as well as the ratio of the traditional mask and the square root mask.The Gammatone channel weight is used to modify the mask value in each channel to simulate the human auditory system and further improve the speech intelligibility.Finally,the simulation of multiple voices under different noise backgrounds shows that compared with different literature algorithms,the algorithm has a higher SNR,subjective speech quality and short-term objective intelligibility,which verifies the effectiveness of the algorithm.     

Advances in statistical mechanics of rock masses and its engineering applications

To efficiently link the continuum mechanics for rocks with the structural statistics of rock masses,a theoretical and methodological system called the statistical mechanics of rock masses(SMRM)was developed in the past three decades.In SMRM,equivalent continuum models of stressestrain relationship,strength and failure probability for jointed rock masses were established,which were based on the geometric probability models characterising the rock mass structure.This follows the statistical physics,the continuum mechanics,the fracture mechanics and the weakest link hypothesis.A general constitutive model and complete stressestrain models under compressive and shear conditions were also developed as the derivatives of the SMRM theory.An SMRM calculation system was then developed to provide fast and precise solutions for parameter estimations of rock masses,such as full-direction rock quality designation(RQD),elastic modulus,Coulomb compressive strength,rock mass quality rating,and Poisson’s ratio and shear strength.The constitutive equations involved in SMRM were integrated into a FLAC3D based numerical module to apply for engineering rock masses.It is also capable of analysing the complete deformation of rock masses and active reinforcement of engineering rock masses.Examples of engineering applications of SMRM were presented,including a rock mass at QBT hydropower station in northwestern China,a dam slope of Zongo II hydropower station in D.R.Congo,an open-pit mine in Dexing,China,an underground powerhouse of Jinping I hydropower station in southwestern China,and a typical circular tunnel in Lanzhou-Chongqing railway,China.These applications verified the reliability of the SMRM and demonstrated its applicability to broad engineering issues associated with jointed rock masses.     

Simulation studies on a rotating ring disk electrode: Effects of ionic migration with kinetic complication-disk results

In continuation to my previous work (Guha S. AIChE J. 2013;59(4):1390-1399), in this work, effects of ionic migration are evaluated for disk region of a rotating ring disk electrode system by numerically solving complex differential equations, developed for mass transfer along with kinetic complication in presence of ionic migration under limiting current condition. The system for simulation is 0.01 M Fe₂(SO₄)₃ solution with H₂SO₄ as supporting electrolyte. Simulation cases are presence and absence of ionic migration with kinetic complication (oxidation of Fe²⁺ to Fe³⁺ under O₂ pressure). Results show that concentration boundary layer thickness of reactant Fe³⁺ reduces appreciably and steady-state disk current reduces substantially in presence of migration. Simulated steady-state disk current in absence of migration case agrees well with published data. Results indicate higher Fe²⁺ concentration in presence of migration and thereby higher rate of oxidation of Fe²⁺ to Fe³⁺ at all rate constant values.     

Tibetan Multi-Dialect Speech and Dialect Identity Recognition

Tibetan language has very limited resource for conventional automatic speech recognition so far. It lacks of enough data, sub-word unit, lexicons and word inventories for some dialects. And speech content recognition and dialect classification have been treated as two independent tasks and modeled respectively in most prior works. But the two tasks are highly correlated. In this paper, we present a multi-task WaveNet model to perform simultaneous Tibetan multi-dialect speech recognition and dialect identification. It avoids processing the pronunciation dictionary and word segmentation for new dialects, while, in the meantime, allows training speech recognition and dialect identification in a single model. The experimental results show our method can simultaneously recognize speech content for different Tibetan dialects and identify the dialect with high accuracy using a unified model. The dialect information used in output for training can improve multi-dialect speech recognition accuracy, and the low-resource dialects got higher speech content recognition rate and dialect classification accuracy by multi-dialect and multi-task recognition model than task-specific models.     

Attitudinal choice models with applications in human decision making

A new family of attitudinal discrete choice models is proposed by considering the attitudinal character and the weight vector, both of which are specific to a decision maker (DM). Given the attribute values of different alternatives, the proposed models give varying choice probabilities, as per the DM's-specific attitudinal character and the weight vector. It is also shown that the conventional discrete choice models are the special cases of the proposed attitudinal models. The proposed choice models are also generalized through an additional parameter to add to their capabilities. An application on real data is included to demonstrate their usefulness in the real world.     

针对多种处理痕迹的数字语音取证算法

现有的数字语音取证研究主要集中于对单一的某种操作进行检测，无法对不相关的操作进行判断。针对该问题，提出了一种能够同时检测经过变调、低通滤波、高通滤波和加噪这四种操作的数字语音取证方法。首先，计算语音的归一化梅尔频率倒谱系数（MFCC）统计矩特征；然后通过多个二分类器对特征进行训练，并组合投票得到多分类器；最后使用该多分类器对待测语音进行分类。在TIMIT以及UME语音库上的实验结果表明，归一化MFCC统计矩特征在库内实验中均达到了97%以上的检测率，且在对MP3压缩鲁棒性测试的实验中，检测率仍能保持在96%以上。