SUMMARYWe propose an improved framework for dynamic mode decomposition (DMD) of 2-D flows for problems originating from meteorology when a large time step acts like a filter in obtaining the significant Koopman modes, therefore, the classic DMD method is not effective. This study is motivated by the need to further clarify the connection between Koopman modes and proper orthogonal decomposition (POD) dynamic modes. We apply DMD and POD to derive reduced order models (ROM) of the shallow water equations. Key innovations for the DMD-based ROM introduced in this paper are the use of the Moore-Penrose pseudoinverse in the DMD computation that produced an accurate result and a novel selection method for the DMD modes and associated amplitudes and Ritz values. A quantitative comparison of the spatial modes computed from the two decompositions is performed, and a rigorous error analysis for the ROM models obtained is presented.

SUMMARYThis paper focuses on a new framework for reduced order modelling of non-intrusive data with application to 2D flows. To overcome the shortcomings of intrusive model order reduction usually derived by combining the POD and the Galerkin projection methods, we developed a novel technique based on Randomized Dynamic Mode Decomposition as a fast and accurate option in model order reduction of non-intrusive data originating from Saint-Venant systems. Combining efficiently the Randomized Dynamic Mode Decomposition algorithm with Radial Basis Function interpolation, we produced an efficient tool in developing the linear model of a complex flow field described by non-intrusive (or experimental) data. The rank of the reduced DMD model is given as the unique solution of a constrained optimization problem. We emphasize the excellent behavior of the non-intrusive reduced order models by performing a qualitative analysis. In addition, we gain a significantly reduction of CPU time in computation of the reduced order models (ROMs) for non-intrusive numerical data.

SUMMARYIn this paper, a new vector-filtering criterion for dynamic modes selection is proposed that is able to extract dynamically relevant flow features from dynamic mode decomposition of time-resolved experimental or numerical data. We employ a novel modes selection criterion in parallel with the classic selection based on modes amplitudes, in order to analyze which of these procedures better highlight the coherent structures of the flow dynamics. Numerical tests are performed on two distinct problems. The efficiency of the proposed criterion is proved in retaining the most influential modes and reducing the size of the dynamic mode decomposition model. By applying the proposed filtering mode technique, the flow reconstruction error is shown to be significantly reduced.

SUMMARYThe approximation of reduced linear evolution operator (propagator) via dynamic mode decomposition (DMD) is addressed for both linear and nonlinear events. The 2D unsteady supersonic underexpanded jet, impinging the flat plate in nonlinear oscillating mode, is used as the first test problem for both modes. Large memory savings for the propagator approximation are demonstrated. Corresponding prospects for the estimation of receptivity and singular vectors are discussed. The shallow water equations are used as the second large-scale test problem. Excellent results are obtained for the proposed optimized DMD method of the shallow water equations when compared with recent POD-based/discrete empirical interpolationbased model reduction results in the literature.

In this paper, a new Domain Decomposition Non-Intrusive Reduced Order Model (DDNIROM) is developed for turbulent flows. The method works by partitioning the computational domain into a number of subdomains in such a way that the summation of weights associated with the finite element nodes within each subdomain is approximately equal, and the communication between subdomains is minimised. With suitably chosen weights, it is expected that there will be approximately equal accuracy associated with each subdomain. This accuracy is maximised by allowing the partitioning to occur through areas of the domain that have relatively little flow activity, which, in this case, is characterised by the pointwise maximum Reynolds stresses. A Gaussian Process Regression (GPR) machine learning method is used to construct a set of local approximation functions (hypersurfaces) for each subdomain. Each local hypersurface represents not only the fluid dynamics over the subdomain it belongs to, but also the interactions of the

We investigate the sensitivity of reduced order models (ROMs) to training data resolution as well as sampling rate. In particular, we consider proper orthogonal decomposition (POD), coupled with Galerkin projection (POD-GP), as an intrusive reduced order modeling technique. For non-intrusive ROMs, we consider two frameworks. The first is using dynamic mode decomposition (DMD), and the second is based on artificial neural networks (ANNs). For ANN, we utilized a residual deep neural network, and for DMD we have studied two versions of DMD approaches; one with hard thresholding and the other with sorted bases selection. Also, we highlight the differences between mean-subtracting the data (i.e., centering) and using the data without mean-subtraction. We tested these ROMs using a system of 2D shallow water equations for four different numerical experiments, adopting combinations of sampling rates and resolutions. For these cases, we found that the DMD basis obtained with hard threshodling is sensitive to sampling rate, performing worse at very high rates. The sorted DMD algorithm helps to mitigate this problem and yields more stabilized and converging solution. Furthermore, we demonstrate that both DMD approaches with mean subtraction provide significantly more accurate results than DMD with mean-subtracting the data. On the other hand, POD is relatively insensitive to sampling rate and yields better representation of the flow field. Meanwhile, resolution has little effect on both POD and DMD performances. Therefore, we preferred to train the neural network in POD space, since it is more stable and better represents our dataset. Numerical results reveal that an artificial neural network on POD subspace (POD-ANN) performs remarkably better than POD-GP and DMD in capturing system dynamics, even with a small number of modes.

No abstract

We perform a combined analytical-numerical study of hydrodynamic sensitivity of a concentrated vortex in collision with a roughness wall. Models of the hydrodynamic instability and wall-bounded flow sensitivity are provided analytically by means of parabolization of the Navier-Stokes equations. The differential quadrature method was originally applied as a numerical technique to solve sets of linear algebraic equations by use of a uniform grid. We extend this technique to multidimensional clustered grids and we develop a two-level differential quadrature algorithm to compute the fluctuating flow and the sensitivity field. Extensive numerical results are presented for a two-dimensional Lamb–Oseen problem and both the solution accuracy and convergence rate are assessed. A variational flow control method is also investigated, that proved to be useful for mitigating the flow instabilities by setting the base flow to optimal parameters.