Characterizing and correcting for the effect of sensor noise in the dynamic mode decomposition

Dynamic mode decomposition (DMD) provides a practical means of extracting insightful dynamical information from fluids datasets. Like any data processing technique, DMD's usefulness is limited by its ability to extract real and accurate dynamical features from noise-corrupted data. Here we show analytically that DMD is biased to sensor noise, and quantify how this bias depends on the size and noise level of the data. We present three modifications to DMD that can be used to remove this bias: (i) a direct correction of the identified bias using known noise properties, (ii) combining the results of performing DMD forwards and backwards in time, and (iii) a total least-squares-inspired algorithm. We discuss the relative merits of each algorithm, and demonstrate the performance of these modifications on a range of synthetic, numerical, and experimental datasets. We further compare our modified DMD algorithms with other variants proposed in recent literature.


Similar Publications

It is of special interest in some conventional chemical technologies to provide either fast mixing of reagents or rapid quenching of final chemical product. One of most promising techniques to achieve it is mixing of multiple jets radially injected in crossflow (JIC) of cylindrical duct when jets impinge and form a jet flowing towards a crossflow. As result of an interaction of counter-flowing jet and confined crossflow, a recirculation flow zone (RFZ) upstream of jets injection plane (JIP) is formed (see Fig. Read More


We introduce a hybridizable discontinuous Galerkin method for the incompressible Navier--Stokes equations for which the approximate velocity field is pointwise divergence-free. The method proposed here builds on the method presented by Labeur and Wells [SIAM J. Sci. Read More


Multiple-relaxation-time (MRT) lattice Boltzmann (LB) model is an important class of LB model with lots of advantages over traditional single-relaxation-time (SRT) LB model. In addition, the computation of strain rate tensor is crucial in MRT-LB simulations of some complex flows. Up to now, there are only two formulas to compute the strain rate tensor in the MRT LB model. Read More


Scaling laws for the propulsive performance of rigid foils undergoing oscillatory heaving and pitching motions are presented. Water tunnel experiments on a nominally two-dimensional flow validate the scaling laws, with the scaled data for thrust, power, and efficiency all showing excellent collapse. The analysis indicates that the behaviour of the foils depends on both Strouhal number and reduced frequency, but for motions where the viscous drag is small the thrust closely follows a linear dependence on reduced frequency. Read More


Concentration gradients in a fluid along a reactive surface due to contrast in surface reactivity generate convective flows. These flows result from contributions by electro and diffusio osmotic phenomena. In this study we have analyzed reactive patterns that release and consume protons, analogous to bimetallic catalytic conversion of peroxide. Read More


Effects of subgrid-scale gravity waves (GWs) on the diurnal migrating tides are investigated from the mesosphere to the upper thermosphere for September equinox conditions, using a general circulation model coupled with the extended spectral nonlinear GW parameterization of Yi\u{g}it et al (2008). Simulations with GW effects cut-off above the turbopause and included in the entire thermosphere have been conducted. GWs appreciably impact the mean circulation and cool the thermosphere down by up to 12-18%. Read More


A general relation is derived between the linear and second-order nonlinear ac conductivities of an electron system at frequencies below the interparticle scattering rate. In this hydrodynamic regime the temperature dependence and the tensorial structure of the nonlinear conductivity are shown to be different from their counterparts in the more familiar kinetic regime of higher frequencies. The obtained formulas are valid for systems with an arbitrary Dirac-like dispersion, either massive or massless, and subsume known results for free-space plasmas and solid-state electron gases. Read More


Kraichnan seminal ideas on inverse cascades yielded new tools to study common phenomena in geophysical turbulent flows. In the atmosphere and the oceans, rotation and stratification result in a flow that can be approximated as two-dimensional at very large scales, but which requires considering three-dimensional effects to fully describe turbulent transport processes and non-linear phenomena. Motions can thus be classified into two classes: fast modes consisting of inertia-gravity waves, and slow quasi-geostrophic modes for which the Coriolis force and horizontal pressure gradients are close to balance. Read More


Lagrangian acceleration statistics in a fully developed turbulent channel flow at $Re_\tau = 1440$ are investigated, based on tracer particle tracking in experiments and direct numerical simulations. The evolution with wall distance of the Lagrangian velocity and acceleration time scales is analyzed. Dependency between acceleration components in the near-wall region is described using cross-correlations and joint probability density functions. Read More


We prove local well posedness in regular spaces and a Beale-Kato-Majda blow-up criterion for a recently derived stochastic model of the 3D Euler fluid equation for incompressible flow. This model describes incompressible fluid motions whose Lagrangian particle paths follow a stochastic process with cylindrical noise and also satisfy Newton's 2nd Law in every Lagrangian domain. Read More