Analysis of electrochemical noise by means of bispectral technique

Nonlinear processes are often encountered in the practice of electrochemical and corrosion measurements. Especially, activation-controlled processes are inherently nonlinear. Taking into account investigations of electrode reactions, linear approximation is a popular approach. In this introductory paper, the possibility of extension of electrochemical noise (EN) measurements to the nonlinear regime is presented. Natural consequence of focusing on nonlinear processes is application of higher-order spectral techniques. Utilization of bispectral representation enables analysis of stationarity and linearity properties of EN. The authors present algorithm enabling assessment of both quantities and also exemplary analysis of noise generated during cathodic polarization, which is important for corrosion protection. Presented at the 4th Baltic Conference on Electrochemistry, Greifswald, 13–16 March 2005. Presented at the 4th Baltic Conference on Electrochemistry, Greifswald, March 13–16, 2005.     

On the analysis of 2D nonlinear gravity waves with a fully nonlinear numerical model

A fully nonlinear numerical method, developed on the basis of Euler equations, is used to study the dynamics of nonlinear gravity waves, mainly in the aspects of the propagation of Stokes wave with disturbed sidebands, the evolution of one wave packet and the interaction of two wave groups. These cases have previously been studied with the higher order spectral method, which will be an approximately fully nonlinear scheme if the order of nonlinearity is not large enough, while the present method in the case of the 2D model has an integration scheme that is exact to the computer precision. As expected, in most cases the results are consistent between these two numerical models and it is confirmed again that this fully nonlinear numerical model is also capable of maintaining a high accuracy and good convergence, particularly in the long-term evolutionary process.     

偏度最大化多通道逆滤波语音去混响研究*

房间混响会降低语音质量和语音可懂度。高阶统计量是衡量非高斯性的重要参量,基于语音非高斯特性可实现语音去混响。本文提出一种基于高阶统计量的多通道语音去混响方法,该方法首次用多通道语音信号线性预测残差的三阶统计量偏度（Skewness）构造代价函数,以去混响重建信号线性预测残差的偏度最大化为目标自适应地更新逆滤波器;同时结合语音信号的产生模型,提出基于偏度准则的线性预测与房间脉冲响应逆滤波联合估计方法,进一步提高去混响算法性能。实验结果表明,该方法相较于已有的基于线性预测残差四阶统计量峰度（Kurtosis）的方法具有更好的去混响效果,且对噪声具有更强的鲁棒性。     

Nonlinear Behaviour of Structures Using the Volterra Series—Signal Processing and Testing Methods

To extend modal analysis to nonlinear structures, and adopting the Volterra series as a mathematical framework, we present some new routes together with progress on signal processing. The closed form expressions of higher-order transfer function on the other hand would permit one to obtain eigenvalues of various orders and eigenvectors. Existing signal processing analyzers are initially devoted to linear systems. Programs treating input and output signals of systems are tailored for one-time (or frequency) variable functions. Nonlinear systems can indeed be analyzed by one-dimensional (direct or inverse) Fourier transforms. However, the experimenter rapidly discovers their limitation when dealing with coupling phenomena that require functions with many time (or frequency) variables. In this framework, multidimensional Fourier transforms are necessary.     

Use of the 2D 1H-13C HSQC NMR Methyl Region to Evaluate the Higher Order Structural Integrity of Biopharmaceuticals

The higher-order structure (HOS) of protein therapeutics is directly related to the function and represents a critical quality attribute. Currently, the HOS of protein therapeutics is characterized by methods with low to medium structural resolution, such as Fourier transform infrared (FTIR), circular dichroism (CD), intrinsic fluorescence spectroscopy (FLD), and differential scanning calorimetry (DSC). High-resolution nuclear magnetic resonance (NMR) methods have now been introduced, representing powerful approaches for HOS characterization (HOS by NMR). NMR is a multi-attribute method with unique abilities to give information on all structural levels of proteins in solution. In this study, we have compared 2D ¹H-¹³C HSQC NMR with two established biophysical methods, i.e., near-ultraviolet circular dichroism (NUV-CD) and intrinsic fluorescence spectroscopy, for the HOS assessments for the folded and unfolded states of two monoclonal antibodies belonging to the subclasses IgG1 and IgG2. The study shows that the methyl region of the ¹H-¹³C HSQC NMR spectrum is sensitive to both the secondary and tertiary structure of proteins and therefore represents a powerful tool in assessing the overall higher-order structural integrity of biopharmaceutical molecules.