Reduced-order Network Research Articles

Development of a sunlight-driven thermoacoustic engine for solar energy harvesting

Development of a sunlight-driven thermoacoustic engine for solar energy harvesting

Experimental and numerical investigations of forced response of multi-element lean-premixed hydrogen flames

Experimental and numerical investigations of forced response of multi-element lean-premixed hydrogen flames

Opportunities for wave energy in bulk power system operations

Opportunities for wave energy in bulk power system operations

Data-Driven Modeling of Geometry-Adaptive Steady Heat Convection Based on Convolutional Neural Networks

Heat convection is one of the main mechanisms of heat transfer, and it involves both heat conduction and heat transportation by fluid flow; as a result, it usually requires numerical simulation for solving heat convection problems. Although the derivation of governing equations is not difficult, the solution process can be complicated and usually requires numerical discretization and iteration of differential equations. In this paper, based on neural networks, we developed a data-driven model for an extremely fast prediction of steady-state heat convection of a hot object with an arbitrary complex geometry in a two-dimensional space. According to the governing equations, the steady-state heat convection is dominated by convection and thermal diffusion terms; thus the distribution of the physical fields would exhibit stronger correlations between adjacent points. Therefore, the proposed neural network model uses convolutional neural network (CNN) layers as the encoder and deconvolutional neural network (DCNN) layers as the decoder. Compared with a fully connected (FC) network model, the CNN-based model is good for capturing and reconstructing the spatial relationships of low-rank feature spaces, such as edge intersections, parallelism, and symmetry. Furthermore, we applied the signed distance function (SDF) as the network input for representing the problem geometry, which contains more information compared with a binary image. For displaying the strong learning and generalization ability of the proposed network model, the training dataset only contains hot objects with simple geometries: triangles, quadrilaterals, pentagons, hexagons, and dodecagons, while the testing cases use arbitrary and complex geometries. According to the study, the trained network model can accurately predict the velocity and temperature field of the problems with complex geometries, which has never been seen by the network model during the model training; and the prediction speed is two orders faster than the CFD. The ability of accurate and extremely fast prediction of the network model suggests the potential of applying reduced-order network models to the applications of real-time control and fast optimization in the future.

Mode transition in a standing-wave thermoacoustic engine: A numerical study

This study investigates the mode transition phenomenon in a standing-wave thermoacoustic engine (TAE) by means of computational fluid dynamics (CFD). Firstly, the steady-state responses of the TAE at selected temperature ratios are examined via continuous wavelet transform. The bifurcation diagram and spectral map indicate that, as the temperature ratio increases, the TAE experiences a series of bifurcations, through which first-mode periodic oscillations, quasiperiodic oscillations and second-mode periodic oscillations occur. Secondly, the TAE performances in the initial decay/build-up, nonlinear saturation and steady states are studied. The onset of the first and/or second acoustic mode is identified via dynamic mode decomposition. The oscillation frequencies and growth/attenuation rates from CFD agree well with those from the reduced-order network model. Nonlinear mode competition takes place during saturation in which the growth of one acoustic mode is affected or even totally inhibited by the growth of the other. At steady state, periodic oscillations exhibit a closed loop in the phase space whilst quasiperiodic oscillations generate a torus. The time-averaged acoustic energy density, acoustic intensity and efficiency increase with increasing temperature ratio. Finally, parametric studies are conducted to investigate the effects of the gap between stack plates and stack position on mode transition. It is found that the TAE will exhibit second-mode oscillations if the stack is near the closed end or the gap is small. Results in this study indicate that mode transition could become a novel approach to match the TAE with external loads for higher electric power outputs.

Model Reduction Methods for Complex Network Systems

Network systems consist of subsystems and their interconnections and provide a powerful framework for the analysis, modeling, and control of complex systems. However, subsystems may have high-dimensional dynamics and a large number of complex interconnections, and it is therefore relevant to study reduction methods for network systems. Here, we provide an overview of reduction methods for both the topological (interconnection) structure of a network and the dynamics of the nodes while preserving structural properties of the network. We first review topological complexity reduction methods based on graph clustering and aggregation, producing a reduced-order network model. Next, we consider reduction of the nodal dynamics using extensions of classical methods while preserving the stability and synchronization properties. Finally, we present a structure-preserving generalized balancing method for simultaneously simplifying the topological structure and the order of the nodal dynamics.

On Exponential Stability of Delayed Discrete-Time Complex-Valued Inertial Neural Networks.

This article tackles the global exponential stability for a class of delayed complex-valued inertial neural networks in a discrete-time form. It is assumed that the activation function can be separated explicitly into the real part and imaginary part. Two methods are employed to deal with the stability issue. One is based on the reduced-order method. Two exponential stability criteria are obtained for the equivalent reduced-order network with the generalized matrix-measure concept. The other is directly based on the original second-order system. The main theoretical results complement each other. Some comparisons with the existing works show that the results in this article are less conservative. Two numerical examples are given to illustrate the validity of the main results.

Large eddy simulation of thermally induced oscillatory flow in a thermoacoustic engine

In this paper, a comprehensive high-fidelity three-dimensional computational fluid dynamic research using large eddy simulation has been conducted to investigate a standing-wave quarter-wavelength thermoacoustic engine that consists of a hot buffer, a stack and a resonator. The performance of the thermoacoustic engine has been analysed in four aspects. Firstly, the dynamic characteristics of the engine during the initial start-up process are investigated when changing the temperature gradient imposed on the stack. Numerical results are compared with those from a system-wide reduced-order network model based on linear thermoacoustic theory. Secondly, the acoustic behaviour of the engine operating at steady state is studied. Fourier Series Model is utilized to decompose the steady-state acoustic pressure oscillations which reveals the unstable longitudinal acoustic modes excited in the engine. The stack serves as an energy source for the fundamental mode while it extracts acoustic power from the second harmonic. Thirdly, the hydrodynamic performances of the engine are inspected, and the obtained three-dimensional flow fields inside the engine enable us to probe into rich nonlinear phenomena including minor losses, mass streaming, etc. Finally, the heat transfer characteristics have been analysed by examining the mean temperature field and transversal heat fluxes along the engine. This research demonstrates that the large eddy simulation framework is effective in simulating the thermally induced oscillatory flow inside thermoacoustic engines. The multi-perspective analytical methodologies are valuable in comprehending the engine performance and provide guidelines for the design and optimization of efficient thermoacoustic engines for recovering waste thermal energy from various sources.

Emergence of Mixed Mode Oscillations in Random Networks of Diverse Excitable Neurons: The Role of Neighbors and Electrical Coupling.

In this paper, we focus on the emergence of diverse neuronal oscillations arising in a mixed population of neurons with different excitability properties. These properties produce mixed mode oscillations (MMOs) characterized by the combination of large amplitudes and alternate subthreshold or small amplitude oscillations. Considering the biophysically plausible, Izhikevich neuron model, we demonstrate that various MMOs, including MMBOs (mixed mode bursting oscillations) and synchronized tonic spiking appear in a randomly connected network of neurons, where a fraction of them is in a quiescent (silent) state and the rest in self-oscillatory (firing) states. We show that MMOs and other patterns of neural activity depend on the number of oscillatory neighbors of quiescent nodes and on electrical coupling strengths. Our results are verified by constructing a reduced-order network model and supported by systematic bifurcation diagrams as well as for a small-world network. Our results suggest that, for weak couplings, MMOs appear due to the de-synchronization of a large number of quiescent neurons in the networks. The quiescent neurons together with the firing neurons produce high frequency oscillations and bursting activity. The overarching goal is to uncover a favorable network architecture and suitable parameter spaces where Izhikevich model neurons generate diverse responses ranging from MMOs to tonic spiking.

Underlying physics of limit-cycle, beating and quasi-periodic oscillations in thermoacoustic devices

Under certain circumstances, self-excited acoustic oscillations inside thermoacoustic systems exhibit beating and quasi-periodic patterns, in contrast to constant-amplitude limit-cycle oscillations regularly studied in the literature. This paper explores the underlying physics of limit cycles, beating and quasi-periodicity in thermoacoustic devices from acoustic/vibrational and thermodynamic/hydrodynamic perspectives. Firstly, the acousto-mechanical coupling between the thermoacoustic engine and external load is investigated using a lumped element model, which is verified against a distributed parameter model. The effect of external load on the natural frequencies and mode shapes of intrinsic oscillation modes is inspected. Secondly, the thermo-acoustic coupling between the temperature and acoustic fields is analysed using a reduced-order network model based on the linear thermoacoustic theory. The effect of thermoacoustic core on stability of oscillation modes is studied. Finally, the steady-state behaviour of the device is examined by considering the unsteady linear decay/growth and nonlinear saturation processes. This study demonstrates that the linear mode selection decides whether the steady state is static (quiescent) or dynamic. Apart from linear mode selection, the dynamic steady-state responses are also affected by nonlinear mode competition during the saturation process. Simultaneous excitement of oscillation modes at different external loads may lead to distinctively different steady-state waveforms including beating and quasi-periodicity. Discussions on the steady-state energy transition/conversion indicate that, to improve the performance of the thermoacoustic device, excitation of the oscillation mode whose frequency is close to that of the external load is encouraged. It is suggested that ultra-compliant transducers should be employed for better acoustic power extraction.

Numerical investigation of synthetic jets driven by thermoacoustic standing waves

Abstract We have carried out a preliminary study of the physical processes leading to the formation of a jet into external quiescent surroundings driven by thermoacoustic standing waves. The standing waves are initiated in a thermoacoustic engine (TAE) utilizing the thermoacoustic effect, and the synthetic jet is produced via a jet ejector where a sudden change in the cross-section is employed. We investigate the characteristics of the proposed synthetic jet actuator using both reduced-order network model and computational fluid dynamics (CFD) simulations. The network technique, which is based on linear thermoacoustic theory, can predict the onset of thermoacoustic instability in the frequency domain. The CFD code solves the fully coupled nonlinear compressible flow equations and enables the time-domain analysis of complex flow patterns, which facilitates comprehension of the jet formation process. Both theoretical analysis and numerical simulations reveal that spontaneous, self-excited oscillations inside the TAE will happen when the temperature ratio is greater than the onset temperature ratio for thermoacoustic instability. CFD simulations further identified the transition from no jet to a clear synthetic jet, which determines the onset temperature ratio for jet formation and the threshold value of a non-dimensional parameter. Finally, we carried out a parametric study to investigate the influence of resonator length and orifice diameter on the onset characteristics as well as other performance parameters including the acoustic intensity, space-averaged mean momentum flux, space-averaged velocity and the jet effectiveness. The proposed thermoacoustically-driven synthetic jet actuator may outperform conventional actuators driven by a piston, loudspeaker or piezoelectric transducers on occasions where the surface temperature is high, and therefore has the potential to be utilized for self-cooling purposes.

Model reduction of synchronized homogeneous Lur’e networks with incrementally sector-bounded nonlinearities

Abstract This paper proposes a model order reduction scheme that reduces the complexity of diffusively coupled homogeneous Lur’e systems. We aim to reduce the dimension of each subsystem and meanwhile preserve the synchronization property of the overall network. Using the Laplacian spectral radius, we characterize the robust synchronization of the Lur’e network by a linear matrix inequality (LMI), whose solutions then are treated as generalized Gramians for the balanced truncation of the linear component of each Lur’e subsystem. It is verified that, with the same communication topology, the resulting reduced-order network system is still robustly synchronized, and an a priori bound on the approximation error is guaranteed to compare the behaviors of the full-order and reduced-order Lur’e subsystems.

An adaptive approach for prediction of propellant feedline dynamics in fluid network

PurposeThis paper aims to propose an adaptive unstructured finite volume procedure for efficient prediction of propellant feedline dynamics in fluid network.Design/methodology/approachThe adaptive strategy is based on feedback control of errors defined by changes in key variables in two subsequent time steps.FindingsAs an evaluation of the proposed approach, two feedline dynamics problems are formulated and solved. First problem involves prediction of pressure surges in a pipeline that has entrapped air and the second is a conjugate heat transfer problem involving prediction of chill down of cryogenic transfer line. Numerical predictions with the adaptive strategy are compared with available experimental data and are found to be in good agreement. The adaptive strategy is found to be efficient and robust for predicting feedline dynamics in flow network at reduced CPU time.Originality/valueThis study uses an adaptive reduced-order network modeling approach for fluid network.

Model reduction of linear multi-agent systems by clustering with varvec{mathcal {H}_2} and varvec{mathcal {H}_infty } error bounds

In the recent paper (Monshizadeh et al. in IEEE Trans Control Netw Syst 1(2):145–154, 2014. https://doi.org/10.1109/TCNS.2014.2311883), model reduction of leader–follower multi-agent networks by clustering was studied. For such multi-agent networks, a reduced order network is obtained by partitioning the set of nodes in the graph into disjoint sets, called clusters, and associating with each cluster a single, new, node in a reduced network graph. In Monshizadeh et al. (2014), this method was studied for the special case that the agents have single integrator dynamics. For a special class of graph partitions, called almost equitable partitions, an explicit formula was derived for the mathcal {H}_2 model reduction error. In the present paper, we will extend and generalize the results from Monshizadeh et al. (2014) in a number of directions. Firstly, we will establish an a priori upper bound for the mathcal {H}_2 model reduction error in case that the agent dynamics is an arbitrary multivariable input–state–output system. Secondly, for the single integrator case, we will derive an explicit formula for the mathcal {H}_infty model reduction error. Thirdly, we will prove an a priori upper bound for the mathcal {H}_infty model reduction error in case that the agent dynamics is a symmetric multivariable input–state–output system. Finally, we will consider the problem of obtaining a priori upper bounds if we cluster using arbitrary, possibly non almost equitable, partitions.

Stability and synchronization preserving model reduction of multi-agent systems

Abstract In this paper, stability and synchronization preserving model reduction schemes are developed for linear multi-agent systems. The multi-agent systems that are considered here are composed of general, yet identical linear subsystems, and the communication topology is assumed to be time-independent. First, under the assumption that the agents have stable internal dynamics and the network is stable, the dynamic order of the agents is reduced such that the corresponding reduced order network is again stable. Then, starting from a synchronized network where agents are allowed to have unstable dynamics, a reduced order model for the network which preserves synchronization is obtained. In addition, model reduction error bounds are established to compare the behavior of the original network to that of the reduced order model. The proposed results are illustrated through a numerical example.

Modeling and simulation of electrification delivery in functionalized textiles in electromagnetic fields

Abstract This work investigates the deformation of electrified textiles in the presence of an externally supplied magnetic field ( B ext ). The electrification is delivered by running current ( J ) through the fibers from an external power source. Of primary interest is to ascertain the resulting electromagnetic forces imposed on the fabric, and the subsequent deformation, due to the terms J × B ext and P E , where P is the charge density, E is the electric field and the current given by J = σ ( E + v × B ext ) , where σ is the fabric conductivity, and v is the fabric velocity. As the fabric deforms, the current changes direction and magnitude, due to the fact that it flows through the fabric. The charge density is dictated by Gauss’ law, ∇ · D = P , where D = ϵ E , ϵ is the electrical permittivity and D is the electric field flux. In order to simulate such a system, one must solve a set of coupled equations governing the charge distribution, current flow and system dynamics. The deformation of the fabric, as well as the charge distribution and current flow, are dictated by solving the coupled system of differential equations for the motion of lumped masses, which are coupled through the fiber-segments under the action of electromagnetically-induced forces acting on a reduced order network model. In the work, reduced order models are developed for (a) Gauss’ law ( ∇ · D = P ), (b) the conservation of current/charge, ∇ · J + ∂ P ∂ t = 0 , and (c) the system dynamics, ∇ · T + f = ρ d v dt , where T is the Cauchy stress and f represents the induced body forces, which are proportional to P E + J × B ext . A temporally-adaptive, recursive, staggering scheme is developed to solve this strongly coupled system of equations. We also consider the effects of progressive fiber damage/rupture during the deformation process, which leads to changes (reduction) in the electrical conductivity and permittivity throughout the network. Numerical examples are given, as well as extensions to thermal effects, which are induced by the current-induced Joule-heating.

Analytic modelling methodology for analysis of energy consumption for ISO 18000-7 RFID networks

Active Radio Frequency Identification (RFID) networks consist of battery powered RFID tags. These tags have a lifetime limited by the on-board battery. Energy consumption of such a network is a critical metric. Future RFID networks are projected to continue to increase in size, eventually containing thousands or millions of nodes. Current simulation methods require evaluation of all entities resulting in extremely long execution time for large networks such as RFID networks. Simulations are a useful tool for designers to select the best design alternative for a given network. This paper presents a general framework and method that reduces the order of a given network providing a relatively simple tractable model. The energy consumption of this reduced order network can then be found quickly and efficiently. This order reduction method allows the designer to quickly analyse the design space and selects the best alternative for the network in question.

A computational framework for agglomeration in thermochemically reacting granular flows

A computational framework is developed which couples a series of models, each describing vastly different physical events, in order to characterize particle growth (agglomeration) in thermochemically reacting granular flows. The modelling is purposely simplified to expose the dominant mechanisms which control agglomeration. The overall system is comprised of relatively simple coupled submodels describing impact, heat production, bonding and fragmentation, each of which can be replaced by more elaborate descriptions, if and when they are available. Inverse problems, solved with a genetic algorithm, are then constructed to ascertain system parameters which maximize agglomeration likelihood within a range of admissible data.

AUXILIARY CONTROL OF STATIC VAR SYSTEM TO IMPROVE THE DYNAMIC PERFORMANCE OF LONG TRANSMISSION LINES

ABSTRACT A static VAR system has been considered located at the middle of a long transmission line, and the effect of the SVS auxiliary controllers on the system dynamic performances has been examined over a wide range of operating conditions. The performance of auxiliary controllers has been studied when a reduced order network model, represented by a lumped parameter T-circuit, is used. The performance is compared with a system having a network model represented by single lumped parameter π-circuit on either side of SVS. It is observed that the size of the system matrix reduces significantly when the reduced order model is used. Also The accuracy of the eigenvalues of the system Is not much affected. The application of the auxiliary controllers, considerably, increases the system damping and unstable system modes are stabilized.

Inertial and slow coherency aggregation algorithms for power system dynamic model reduction

This paper presents new aggregation algorithms for obtaining reduced order power networks when coherent generators are aggregated. The generation terminal bus aggregation algorithm in the EPRI DYNRED software tends to stiffen the reduced order network during the aggregation process, thus increasing the frequencies of inter-area modes. The inertial and slow coherency aggregations will decrease the stiffening effect and produce, for the same coherent machine groups, aggregate networks with improved inter-area mode approximations. This paper contains new procedures to construct these aggregate networks and demonstrates the benefits of these new aggregate networks on a 48-machine power system using eigenvalues and nonlinear simulations.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>