非定常动态演化伴随优化设计方法

研究了基于定常流动解和伴随方程定常解基础上的传统的伴随方法.在此基础上对定态飞行气动外形的优化设计提出了基于非定常流动控制方程瞬态解和非定常伴随方程瞬态解的新的优化方法,称之为动态演化伴随方法.这种新的优化方法保留了传统伴随方法适用于具有大数量设计变量的气动优化问题,而且比传统的伴随方法可节省大量的计算时间.大量算例计算结果表明,新方法与传统方法具有相同的精度.     

基于控制理论和N-S方程的二维叶栅气动优化算法

基于控制理论的气动优化方法的计算量与设计变量无关,可快速精确地完成控制变量的灵敏度分析.本文以出口熵增最小为目标函数,详细推导了二维 N-S 方程伴随系统的偏微分方程组及其相应的边界条件和敏感性导数的表达式.以拟牛顿算法为优化求解器,利用 CFD 方法求解流动变量,采用时间推进方法求解伴随方程,建立了基于黏性伴随方法和 N-S 方程的二维叶栅气动优化设计算法.     

应用控制理论的叶栅气动反设计

基于控制理论的气动设计方法作为一种基于梯度的优化方法,通过引入伴随系统计算目标函数的敏感性导数,大大降低设计成本.本文将基于控制理论的气动设计方法应用到透平叶栅的气动反问题中,应用Euler方程研究了二维叶栅的压力反设计问题,并讨论了该方法具体实施中的关键问题,包括采用非均匀B样条进行二维叶栅造型;应用Thompson时间相关边界条件理论进行伴随方程特征分析;研究伴随方程的数值求解方法,构造伴随方程的耗散通量.通过算例证明了该气动设计方法适用性好,速度快,可以大大节约计算成本.     

透平叶栅三维粘性气动反问题的控制理论方法

将基于控制理论的形状优化设计方法应用于粘性可压流动条件下的透平叶栅三维气动反设计,详细推导了三维N-S方程伴随系统的偏微分方程组及其各类边界条件.讨论了伴随系统的解的适定性条件,并由此给出应用N-S方程进行气动优化的目标函数的选取限制.研究了伴随方程的数值求解技术,给出敏感性导数的最终计算式,结合拟牛顿算法发展了三维透平叶栅粘性反问题的气动设计方法.     

海水温度垂直分布预报数据同化的离散伴随算子法

以一维水温模型为例,利用伴随算子法进行海洋观测数据同化,以便给水温的数值预报提供一个较准确的初始场。讨论了离散伴随算子法的思想,最优化过程,计算水温方程和伴随方程的差分格式,并图示描述水温场的初始猜测和同化后的分布。     

复杂约束环境下基于控制理论的透平叶栅气动优化

本文应用控制理论,基于网格节点位置坐标直接变分法建立了N-S方程的伴随系统.以叶栅通道内熵增最小为目标函数,并以流量为约束条件,详细推导了具有约束条件的二维N-S方程伴随系统的偏微分方程组及其相应的边界条件和敏感性导数的表达式。建立了基于黏性连续伴随方程和N-S方程的二维叶栅气动优化设计系统,并成功地应用于某一跨音速叶栅的优化设计。对计算结果的分析表明,该方法能够适用于透平叶栅气动优化设计。     

透平叶栅三维形状反问题研究

随着CFD技术的发展,基于伴随方法的求解Euler和NS方程的气动优化设计已成为流体力学形状反问题研究中的热门领域.本文应用该方法对透平叶栅进行三维气动优化设计,详细推导了Euler方程伴随系统的偏微分方程组及其各类边界条件,首次给出了透平内流伴随方程边界条件的具体形式,并给出伴随变量的物理意义.结合拟牛顿算法发展了三维透平叶栅形状反问题气动优化算法,并给出了算法的流程.     

基于定涡黏性假设的连续伴随优化方法研究

基于网格节点位置坐标变分并结合通量雅克比矩阵技术,完成了定涡黏性假设下带湍流模型连续伴随系统的建立,极大降低了RANS方程下复杂伴随系统的推导难度,简化了伴随系统表达式,节约了计算资源。同时采用定涡黏性假设耦合SpaIart-Allmaras湍流模型,实现了二维叶栅壁面压力反设计的数值验证,针对变量梯度、优化过程及结果与全湍流系统进行了分析对比;并以通道内熵增为优化目标完成了某亚音速叶栅基于该假设的气动优化及叶栅吹风试验,试验结果验证了优化设计结果的可靠性。     

基于梯度优化物理信息神经网络求解复杂非线性问题

近年来,物理信息神经网络(PINNs)因其仅通过少量数据就能快速获得高精度的数据驱动解而受到越来越多的关注.然而,尽管该模型在部分非线性问题中有着很好的结果,但它还是有一些不足的地方,如它的不平衡的反向传播梯度计算导致模型训练期间梯度值剧烈振荡,这容易导致预测精度不稳定.基于此,本文通过梯度统计平衡了模型训练期间损失函数中不同项之间的相互作用,提出了一种梯度优化物理信息神经网络(GOPINNs),该网络结构对梯度波动更具鲁棒性.然后以Camassa-Holm(CH)方程、导数非线性薛定谔方程为例,利用GOPINNs模拟了CH方程的peakon解和导数非线性薛定谔方程的有理波解、怪波解.数值结果表明,GOPINNs可以有效地平滑计算过程中损失函数的梯度,并获得了比原始PINNs精度更高的解.总之,本文的工作为优化神经网络的学习性能提供了新的见解,并在求解复杂的CH方程和导数非线性薛定谔方程时用时更少,节约了超过三分之一的时间,并且将预测精度提高了将近10倍.     

基于跟驰模型的列车运行优化控制模拟研究

列车运行的优化控制是降低运输成本,提高运输服务质量以及实现轨道交通可持续发展的重要方式.本文在传统优化速度跟驰模型的基础上,以能量节约为目标提出了一种改进的模拟模型,用以模拟分析城市轨道交通系统中列车运行的优化控制.所提出的模型是通过在经典的优化速度跟驰模型(见Phys.Rev.E 51 1035 Bando等,1995)中引入新的目标优化速度函数来实现在复杂限速条件下列车运行的优化控制.数值模拟则是以北京市地铁亦庄线为例,利用亦庄线实测数据开展研究.结果表明,所提出的模型能够很好地描述复杂限速条件下列车运行的动态特性,模拟测量得到的结果和亦庄线的实测数据较为符合,由此说明所提出模型的有效性.进一步,通过分析列车运行时空图,列车运行的速度变化及运行时间等,讨论了复杂环境下列车流的时空演化特性.     

Data assimilation for the Navier–Stokes- equations

The well-posedness of the data assimilation problem for the Navier–Stokes-α equations on a bounded three-dimensional domain is investigated. The data assimilation procedures under consideration are the adjoint method of variational data assimilation (4D-Var) and the method of continuous data assimilation. Concerning the adjoint method the existence of optimal initial conditions with respect to an observation-dependent cost functional is proven, the optimizers are characterized by a first-order necessary condition involving the adjoint linearized Navier–Stokes-α equations and conditions for the uniqueness of the initial conditions are given. Well-posedness of the continuous data assimilation problem is proven and convergence rates in terms of observational resolution are provided.     

SUPG-Stabilized Virtual Element Method for Optimal Control Problem Governed by a Convection Dominated Diffusion Equation

In this paper, the streamline upwind/Petrov Galerkin (SUPG) stabilized virtual element method (VEM) for optimal control problem governed by a convection dominated diffusion equation is investigated. The virtual element discrete scheme is constructed based on the first-optimize-then-discretize strategy and SUPG stabilized virtual element approximation of the state equation and adjoint state equation. An a priori error estimate is derived for both the state, adjoint state, and the control. Numerical experiments are carried out to illustrate the theoretical findings.     

Identification of an optimal derivatives approximation by variational data assimilation

Variational data assimilation technique applied to identification of optimal approximations of derivatives near boundary is discussed in frames of one-dimensional wave equation. Simplicity of the equation and of its numerical scheme allows us to discuss in detail as the development of the adjoint model and assimilation results. It is shown what kind of errors can be corrected by this control and how these errors are corrected. This study is carried out in view of using this control to identify optimal numerical schemes in coastal regions of ocean models.     

Optimal nonlocal boundary control of the wide-angle parabolic equation for inversion of a waveguide acoustic field

This paper applies the concept of optimal boundary control for solving inverse problems in shallow water acoustics. To treat the controllability problem, a continuous analytic adjoint model is derived for the Claerbout wide-angle parabolic equation (PE) using a generalized nonlocal impedance boundary condition at the water-bottom interface. While the potential of adjoint methodology has been recently demonstrated for ocean acoustic tomography, this approach combines the advantages of exact transparent boundary conditions for the wide-angle PE with the concept of adjoint-based optimal control. In contrast to meta-heuristic approaches the inversion procedure itself is directly controlled by the waveguide physics and, in a numerical implementation based on conjugate gradient optimization, many fewer iterations are required for assessment of an environment that is supported by the underlying subbottom model. Furthermore, since regularization schemes are particularly important to enhance the performance of full-field acoustic inversion, special attention is devoted to the application of penalization methods to the adjoint optimization formalism. Regularization incorporates additional information about the desired solution in order to stabilize ill-posed inverse problems and identify useful solutions, a feature that is of particular importance for inversion of field data sampled on a vertical receiver array in the presence of measurement noise and modeling uncertainty. Results with test data show that the acoustic field and the bottom properties embedded in the control parameters can be efficiently retrieved.     

升力系数不变时跨音速机翼阻力优化设计

基于共轭方程的优化设计理论,应用三维欧拉方程进行了升力系数不变时跨音速机翼阻力优化设计研究,根据给定的目标函数推导了在物理空间上表述的共轭方程及边界条件,研究了共轭方程的数值求解方法及目标函数对设计变量的敏感性导数求解问题,发展了一种跨音速机翼阻力优化设计方法,应用该设计方法进行了跨音速机翼阻力优化设计研究,优化后机翼表面的激波强度减弱很多,有效减少了波阻.     

Computation of the analysis error covariance in variational data assimilation problems with nonlinear dynamics

The problem of variational data assimilation for a nonlinear evolution model is formulated as an optimal control problem to find the initial condition function. The data contain errors (observation and background errors), hence there will be errors in the optimal solution. For mildly nonlinear dynamics, the covariance matrix of the optimal solution error can often be approximated by the inverse Hessian of the cost functional. Here we focus on highly nonlinear dynamics, in which case this approximation may not be valid. The equation relating the optimal solution error and the errors of the input data is used to construct an approximation of the optimal solution error covariance. Two new methods for computing this covariance are presented: the fully nonlinear ensemble method with sampling error compensation and the ‘effective inverse Hessian’ method. The second method relies on the efficient computation of the inverse Hessian by the quasi-Newton BFGS method with preconditioning. Numerical examples are presented for the model governed by Burgers equation with a nonlinear viscous term.     

Theoretical framework for geoacoustic inversion by adjoint method

Traditional geoacoustic inversions are generally solved by matched-field processing in combination with metaheuristic global searching algorithms which usually need massive computations. This paper proposes a new physical framework for geoacoustic retrievals. A parabolic approximation of wave equation with non-local boundary condition is used as the forward propagation model. The expressions of the corresponding tangent linear model and the adjoint operator are derived, respectively, by variational method. The analytical expressions for the gradient of the cost function with respect to the control variables can be formulated by the adjoint operator, which in turn can be used for optimization by the gradient-based method.     

Optimal tracking control with zero steady-state error for time-delay systems with sinusoidal disturbances

The problem of optimal tracking control with zero steady-state error for linear time-delay systems with sinusoidal disturbances is considered. Based on the internal model principle, a disturbance compensator is constructed such that the system with external sinusoidal disturbances is transformed into an augmented system without disturbances. By introducing a sensitivity parameter and expanding power series around it, the optimal tracking control problem can be simplified into the problem of solving an infinite sum of linear optimal control series without time-delay and disturbance. The obtained optimal tracking control law with zero steady-state error consists of accurate linear state feedback terms and a time-delay compensating term, which is an infinite sum of an adjoint vector series. The accurate linear terms can be obtained by solving a Riccati matrix equation and a Sylvester equation, respectively. The compensation term can be approximately obtained through a recursive algorithm. A numerical simulation shows that the algorithm is effective and easily implemented, and the designed tracking controller is robust with respect to the sinusoidal disturbances.