模糊系统辨识方法在TIG焊过程建模中的应用

运用模糊集合理论，通过系统输入／输出量测值进行相关分析来建立脉冲ＴＩＧ焊动态过程的模糊模型，这是基于模糊关系模型系统辨识方法，并给出了模型精度的测试结果，结果表明，用该方法对ＴＩＧ焊过程建模是有效的。     

Identification of Dynamic Parameters Based on Pseudo-Parallel Ant Colony Optimization Algorithm

IntroductionReal ants are capable of finding the shortest pathfrom a food source to the nest. Inspired by this factand the behavior of ant colonies, a novel optimizationalgorithm called ant system (AS) was first developedby Dorigo in1992[1]. In the following years, diversemodifications of the AS algorithm were made andapplied to many different types of optimization prob-lems, and satisfactory results were obtained. Re-cently, the AS algorithm has been extended to an al-gorithm for solving d…     

谈谈如何辨识危险源——实施《职业健康安全管理体系》的几点体会

通过对危险源概念的阐述及原因的分析，结合上海港复兴船务公司实际情况，以“活动分析”为关键点，讨论了危险源的辨识方法，旨在推动危险源辨识，从而达到安全生产目的。     

GM(0,N)模型参数估计的新方法

从灰色理论建模特点出发，提出了GM（0，N）模型参数辨识的一种新方法-累积法，实例分析表明，累积法较传统的参数估计方法在预测精度上有很大的提高，而且计算量小。     

电液伺服系统的神经网络辨识与自适应控制

对一实际电液伺服系统进行了神经网络仿真研究，仿真结果表明，神经网络用于电液控制系统的辨识与自适应控制是可行的，可以解决电液控制系统中非线性及其控制等问题，它对于增强系统的自适应能力，鲁棒性和稳态精度等都有令人满意的控制效果。     

On-line detecting of transformer winding deformation based on parameter identification of leakage inductance

Transformers are required to demonstrate the ability to withstand short circuit currents. Over currents caused by short circuit can give rise to windings deformation. In this paper, a novel method is proposed to monitor the state of transformer windings, which is achieved through on-line detecting the leakage inductance of the windings. Specifically, the mathematical model is established for on-line identifying the leakage inductance of the windings by applying least square algorithm （LSA） to the equivalent circuit equations. The effect of measurement and model inaccuracy on the identification error is analyzed, and the corrected model is also given to decrease these adverse effect on the results. Finally, dynamic test is carried out to verify our method. The test results clearly show that our method is very accurate even under the fluctuation of load or power factor. Therefore, our method can be effectively used to on-line detect the windings deformation.     

基于改进遗传算法的递归神经网络非线性系统辨识

将递归内时延神经网络应用于非线性动力学系统辨识中，描述了其动力学方程，并引入改进遗传算法作为其学习算法，通过非线性动力学SISO和MIMO系统的辨识仿真研究，验证了内时延递归网络结构和改进遗传算法的有效性。     

一种操纵性水动力导数的实用估算方法

目前有很多诸如最小二乘、卡尔曼滤波等系统辨识方法用于船舶操纵性水动力导数的估算．但涉及神经网络的甚少．随着神经网络技术的日益成熟、完善，神经网络在船舶操纵性领域的应用将成为可能．文中介绍了一种基于线性船舶操纵运动方程的水动力导数估算方法，建立了一种基于神经网络的系统辨识模型，并将其应用于分析实船试验数据．结果表明了该方法的有效性．     

高速铁路噪声源区划及各区域声源贡献量分析

研究高速铁路噪声源区划方法并分析各区域声源贡献量,对高速铁路噪声治理有重要意义。基于高速铁路噪声源辨识现场测试,分析得到噪声源的位置和幅值。将噪声源按高度划分为轮轨区、车体下部、车体上部、集电系统和桥梁结构等5个区域,进一步将车体上部沿线路方向划分为车头区和非车头区,将集电系统区域沿线路方向划分为受电弓区和接触网区。根据声波能量叠加原理计算每个区域噪声源辐射功率,研究各个区域声源贡献量。分析结果表明,列车以300 km/h运行时,轮轨区噪声占48%,车体下部噪声占25%,合计占总噪声的73%,对高速铁路辐射噪声起主导作用。