钢轨匹配中的线结构光平面标定方法研究

在利用激光无接触技术进行轨道检测时,为了彻底摆脱激光摄像式传感器对称安装的束缚,本文在大量调研国内外线结构光视觉系统标定方法的基础上,提出一种基于共面法的视觉系统结构参数的标定方法—等高线标定法。在量块上表面提取140个数据点,先对数据进行坐标变换,再利用Levenberg-Marquardt优化算法进行非线性迭代使所有点处于同一等高线上。该方法操作简单,对设备要求低,标定试验数据表明等高线标定法重复性达0.00133°,检定试验数据表明视觉系统在等高线法标定线结构光平面的基础上检测精度可达7.7 μm。其高精度性和便捷性符合铁道部门对轨道检测的要求。     

基于ANSYS仿真的阳极键合薄膜形变量的研究

阐述了一种利用阳极键合技术加工的光纤法布里-珀罗MEMS压力传感器的工作原理,建立了考虑阳极键合产生的热应力作用下边界固支的圆膜受外加压力的薄膜形变量公式.利用有限元分析软件ANSYS对与玻璃环键合后实际结构中的膜片在压力作用下的挠度变化进行模拟,采用麦夸特(Levenberg-Marquardt)算法对模拟结果进行拟合修正,曲线拟合度达99.99％,得到了键合后实际结构中硅膜在外加压力作用下的形变量拟合公式.修正公式的曲线与实际ANSYS模拟所得数据拟合曲线的误差小于0.01％.该研究对膜片型光纤压力传感器的结构参数设计和广泛应用具有重要的理论指导意义.     

A numerical method for determining a quasi solution of a backward time-fractional diffusion equation

In this study, we present existence and uniqueness theorems of a quasi solution to backward time-fractional diffusion equation. To do that, we consider a methodology, involving minimization of a least squares cost functional, to identify the unknown initial data. Firstly, we prove the continuous dependence on the initial data for the corresponding forward problem and then we obtain a stability estimate. Based on this, we give the existence theorem of a quasi solution in an appropriate class of admissible initial data. Secondly, it is shown that the cost functional is Fréchet-differentiable and its derivative can be formulated via the solution of an adjoint problem. These results help us to prove the convexity of cost functional and subsequently the uniqueness theorem of the quasi solution. In addition, in order to approximate the quasi solution, WEB-spline finite element method is used. Since the obtained system of linear equations is ill-posed, we apply the Levenberg-Marquardt regularization. Finally, a numerical example is given to show the validation of the introduced method.     

氧化铝浓度软测量技术研究

分析了铝电解生产过程中建立氧化铝浓度神经网络软测量模型的必要性和可行性,根据软测量建模理论建立了采用Levenberg-Marquardt算法的氧化铝浓度神经网络软测量模型.实际应用结果证明了该软测量模型能够有效实现氧化铝浓度的在线检测.     

基于3维模型的单视图不规则物体定位

单幅图像中物体的定位需要估计物体在3维空间中的位置和姿态,在静态场景中等价于摄像机定标,具有广泛的应用价值.针对具有3维模型,但缺失纹理信息的不规则物体的定位问题,提出一种基于轮廓匹配的定位方法.首先使用基于图像分割的方法提取输入图像中物体轮廓线,然后可以将图像轮廓线与给定位置和姿态参数下渲染3维模型的轮廓线进行匹配,匹配误差可以表示为位置与姿态参数的函数.由于该函数不能解析表达与求解,需要通过离散采样计算导数及目标函数值.位置与姿态参数的最优值可以通过LM(Levenberg-Marquardt)方法进行求解.实验结果表明,该方法可以快速收敛,并具有很高的精确性和鲁棒性.     

含有平面结构场景的捆绑调整

捆绑调整是计算机视觉中三维结构恢复过程的重要步骤. 捆绑调整通常将空间中点与点坐标的调整视为相互独立的过程,但是在包含有自然物和人工物的场景中,由于存在多余的自由度,这种调整方法会导致优化结果偏离真值. 提出了一种带有共面约束和平面夹角约束的捆绑调整,用于优化带有平面的场景. 借助新的参数化方法,共面约束和夹角约束可以方便地进行表示,并且带有这两类约束的捆绑调整求解过程,仍然是一个无约束的非线性最小二乘问题. 实验结果表明,这种带有先验信息的捆绑调整提供了对结构的更准确估计. 由于先验信息的加入,增强型法方程的维度变高,借助了稀疏的求解技术和预条件子方法,大大降低了求解时间. 最后,为了在实际应用中能够自动寻找出夹角约束,提出了一种基于最大完全图的贪心方法,该方法尽可能多地保留了夹角约束.     

基于LM算法的企业资信网络评估模型研究

针对当前企业资信评估方法的不足,提出将基于Levenberg-Marquard(LM)算法的前向多层神经网络用于企业资信评估,并通过MATLAB神经网络工具对其进行模拟计算。实验结果表明,该方法稳定、快捷、预测准确,对企业资信评估有着良好的性能。     

Short-term electricity prices forecasting in a competitive market: A neural network approach

This paper proposes a neural network approach for forecasting short-term electricity prices. Almost until the end of last century, electricity supply was considered a public service and any price forecasting which was undertaken tended to be over the longer term, concerning future fuel prices and technical improvements. Nowadays, short-term forecasts have become increasingly important since the rise of the competitive electricity markets. In this new competitive framework, short-term price forecasting is required by producers and consumers to derive their bidding strategies to the electricity market. Accurate forecasting tools are essential for producers to maximize their profits, avowing profit losses over the misjudgement of future price movements, and for consumers to maximize their utilities. A three-layered feedforward neural network, trained by the Levenberg-Marquardt algorithm, is used for forecasting next-week electricity prices. We evaluate the accuracy of the price forecasting attained with the proposed neural network approach, reporting the results from the electricity markets of mainland Spain and California.     

基于Levenberg-Marquardt算法的碳化硅MOSFET建模研究

为了更快速准确地对碳化硅MOSFET功率器件进行开关行为预测与分析,需要建立其静态和动态行为模型. 静态模型包括不同温度下的转移特性曲线和输出特性曲线,以及寄生非线性电容曲线等. 提出了一种基于EKV公式改进的曲线拟合公式和一种新的非线性电容拟合公式,利用Levenberg-Marquardt算法进行参数拟合,建模速度快,模型误差小. 动态模型在考虑封装寄生电感和寄生非线性电容等非理想条件下,分别建立器件导通和关断过程每一个阶段的栅源极电压、漏源极电压、肖特基二极管电压、栅极电流和漏极电流的电路微分方程组,再以每一个阶段结束时的状态变量作为下一个阶段的初始条件. 采用4阶龙格库塔法求解上述微分方程组的数值解,并与LTspice仿真波形进行对比分析. 结果表明,上述建模方法能较好地描述器件的动态行为特性.     

Preparation of Fe-Mn/K/Al2O3 Fischer-Tropsch Catalyst and Its Catalytic Kinetics for the Hydrogenation of Carbon Monoxide

A K promoted iron-manganese catalyst was prepared by sol-gel method, and subsequently was tested for hydrogenation of carbon monoxide to light olefins. The kinetic experiments on a well-characterized Fe-Mn/K/Al₂O₃ catalyst were performed in a fixed-bed micro-reactor in a temperature range of 280-380 ℃, pressure range of 0.1-1.2 MPa, H₂/CO feed molar ratio range of 1-2.1 and a space velocity range of 2000-7200 h^-1. Considering the mechanism of the process and Langmuir-Hinshelwood-Hogan-Watson (LHHW) approach, unassisted CO dissociation and H-assisted CO dissociation mechanisms were defined. The best models were obtained using non-linear regression analysis and Levenberg-Marquardt algorithm. Consequently, 4 models were considered as the preferred models based on the carbide mechanism. Finally, a model was proposed as a best model that assumed the following kinetically relevant steps in the iron-Fischer-Tropsch (FT) synthesis: (1) CO dissociation occurred without hydrogen interaction and was not a rate-limiting step; (2) the first hydrogen addition to surface carbon was the rate-determining steps. The activation energy and adsorption enthalpy were calculated 40.0 and -30.2 kJ·mol^-1, respectively.