Electrochemical reaction and transport processes in porous electrodes are optimized with a mesoscale topology porous structure evolution scheme, which is simulated by a mesoscale non-dimensional lattice Boltzmann method (NDLBM). The evolutional porosity distribution fields are simulated for both single-component (SC) and multi-component (MC) electrochemical reaction models. The dynamic balance of positive and negative direction reactions in porous electrodes is also investigated to capture the transient fields of mesoscale reaction rates and electric fluxes during cyclic charge-discharge (CCD) applications. Compared to experimental data of CCD potentials, the mean errors are 11.63% and 5.25% for the SC and MC models, respectively. Relative to pre-designed channel structures, the topology optimized structure (TOS) increases the maximum transient chemical reaction rate by more than 54.8%, and more than doubles the electric fluxes during structure evolutions. With the quasi-steady TOS fixed during the CCD, the cyclic averaged enhancement ratios of the electric power reach 23.6% and 32.7% for the SC and MC models, respectively. Under equivalent charge conditions, both TOS models achieve a maximum enhancement ratio about 25% for the electric energy density. This study provides a tool for optimizing the design and evaluating the CCD performance of TOS electrodes.

A physics informed neural network (PINN) framework is built and applied for efficient prediction of convective oscillating pipe flows. Two PINN models, i.e., the Navier-Stokes equations informed neural network (NS-NN) and the Reynolds averaged Navier-Stokes equations informed neural network (RANS-NN), are proposed based on the physics-guided loss functions formulated from the governing equation residuals. A four-module computational code is developed for comparative studies. Code validations for Taylor-Green vortex flow fields show that the prediction errors are less than 0.19%, 0.60%, and 4.88% for outputs, first and second gradients, respectively. For both training and validation sets, the loss functions of the RANS-NN model versus epoch numbers decay at a rate of 4% faster than the NN and NS-NN models. The parametric studies with combined Reynolds numbers in the range of 102 <= Re <= 104 reveal different prediction performance across three flow regimes. The RANS-NN is more efficient, accurate and robust in high Re ranges, especially in the turbulent regime with the error reduction ratios relative to the NN model of at least 9.24% and 5.04% for velocity and temperature fields, respectively. The present PINN is based on interpolation schemes, and its temporal extrapolation performance degrades faster than the parameter extrapolation performance.

A physics informed neural network (PINN) framework is constructed for predicting the transient fields of natural convective nanofluids immersed in oscillating magnetic fields. A PINN code with inverse inference of dynamic parameters is developed. For pure fluid without nanoparticles, prediction errors of temperature and velocity fields by the PINN are 0.5% to 1.1% smaller than the conventional neural network (NN). Across all training and validation datasets for both pure fluid and nanofluid cases, the convergence speed of the PINN is about 10% faster than that of the NN at 90% accuracy level. Due to combined effects of parameters on multiple inverse coefficients, the recovered nanoparticle volume fractions of nanofluid properties and Hartmann numbers of magnetic field intensities exhibit larger inference errors of 1.5% and 3.0%, respectively, compared to the two transport parameters, i.e., Rayleigh numbers and Prandtl numbers with inference errors less than 1.0%. The convergence speed of the nanoparticle volume fraction is considerably slower than that of all other parameters, because it affects all effective properties and governing parameters. Investigations with Hartmann numbers ranging from 0 to 1000 show varying performance in different heat transfer regimes. In the convection dominating range with Hartmann numbers less than 100, the PINN significantly outperforms the NN, achieving reductions at least 4.81% and 4.21% in the testing datasets for velocity and temperature fields, respectively. The PINN improves the fidelity of forecasting and inverse inference, particularly for stronger convective nanofluid flows.

Principal Component Analysis (PCA) is a popular technique for dimension reduction, but the conventional PCA approach includes many irrelevant variables in the estimates of the principal components (PCs). This makes it difficult to interpret the PCs, especially with a large number of variables. To enhance interpretability, various sparse PCA approaches have been proposed. These approaches mainly impose convex penalties such as the L1$$ {L}_1 $$ penalty to generate sparse estimates of PCs. However, achieving a sparser solution requires selecting a larger regularization parameter, which increases the bias in the estimation of PC loadings and negatively impacts estimation accuracy. To address this bias issue, this paper extends the sparse PCA method to the case with nonconvex penalties. To simplify the resulting optimization problem, an iterative procedure based on the local linear approximation is developed for estimating the sparse PC loadings. The simulation studies and a real example show that the proposed approach provides more accurate PC estimates than the existing PCA approaches in the presence of rowwise and/or cellwise data contamination.

Natural convective nanofluid flows immersed in oscillating magnetic fields are simulated with a sub-continuous nondimensional lattice Boltzmann model. The effective electrical conductivity model is built including coupled effects of nanoparticle concentrations and two Knudsen numbers. Effects of directions, frequencies, and strength amplitudes of the magnetic fields are studied in wide ranges of Hartmann numbers (0.1 <= Ha integral,perpendicular to <= 600) and Rayleigh numbers (10(3)<= Ra integral,perpendicular to <= 10(7)). To achieve higher values of cycle averaged Nusselt numbers Nuintegral,perpendicular to, optimal magnetic directions are along or opposite from the gravity directions. Effects of the magnetic frequency f(B) are negligible, in the conduction dominating lower Rayleigh number regime of Ra integral,perpendicular to<10(4). In the convection dominating regime, Nuintegral,perpendicular to increase with Ra-integral,Ra-perpendicular to in orders of Ra-integral,perpendicular to(0.48) and Ra-integral,perpendicular to(0.45) for vertical and horizontal magnetic directions, respectively, and maximum values of Nu(integral,perpendicular to) appear at the optimal magnetic frequency of f(B)=1/5 (c)(s)*MaL(L/U-L) for all magnetic directions. With Ra-integral,Ra-perpendicular to as high as 106, the oscillating amplitudes of the transient wall mean Nusselt numbers Nu(integral,perpendicular to) increase with increasing Ha integral,perpendicular to, but the cycle averaged Nusselt numbers Nuintegral,perpendicular to decrease from 9.35 to 1.42 with increasing Ha integral,perpendicular to in the transient regime of 5 <= Ha(integral,perpendicular to)<= 500. Meanwhile, heat transfer patterns transit back from convection to conduction dominating patterns with increasing Ha integral,perpendicular to, as illustrated by transient streamlines and isotherms.

A design of microchip cooling with natural convection driven multi-phase coolant flows is proposed by using thermoelectric modules as heat sinks. A new multi-component multi-phase (MCMP) thermal energy transport model is built for convective heat transfer with solid–liquid–gas phase changes. Different from previous lattice Boltzmann models based on temperature distributions, the present model is based on MCMP energy density distributions, which track sensible and latent heat transport for each phase. The local MCMP heat fluxes are related to mesoscale relaxation times and phase volume fractions. Transient solid–liquid–gas interfaces and local temperatures are obtained through the dynamical balance of mesoscale forces and fluxes among all phases together with MCMP enthalpy–temperature distributions. Comparisons of the on and off states of the heat sink modules show that the chip temperature can drop from about 300 to 110 °C, with the coolant liquid–gas phase change temperature of 80 °C at equivalent heat source and sink intensities of 100 W/cm2. Results also show that increasing heat sink intensity to 300 W/cm2 or decreasing gas–liquid phase change temperature to 60 °C can further decrease the chip temperature to 90 °C. The effects of sizes of chips, modules, and cooling units are also illustrated. It is observed that gas bubbles attached on chips impede heat transfer due to the low gas phase thermal conductivity, which means future improvements should also focus on removing or degrading attached gas bubbles. This model provides a quantitative tool for immersion chip cooling performances.

Prediction of non-equilibrium flows is critical to space flight. In the present work we demonstrate that the recently developed spectral multiple-relaxation-time (SMRT) lattice Boltzmann (LB) model is theoretically equivalent to Grad's eigen-system [Grad, H., In Thermodynamics of Gases (1958)] where the eigen-functions obtained by tensor decomposition of the Hermite polynomials are also those of the linearized Boltzmann equation. Numerical results of shock structure simulation using Maxwell molecular model agree very well with those of a high-resolution fast spectral method code up to Mach 7 provided that the relaxation times of the irreducible tensor components match their theoretical values. If a reduced set of relaxation times are used such as in the Shakhov model and lumped-sum relaxation of Hermite modes, non-negligible discrepancies starts to occur as Mach number is raised, indicating the necessity of the fine-grained relaxation model. Together with the proven advantages of LB, the LB-SMRT scheme offers a competitive alternative for non-equilibrium flow simulation.

A wind farm is usually equipped with multiple wind turbines of the same type. These wind turbines often work under same complex conditions. Accurate performance degradation monitoring is crucial for ensuring the reliable operation of wind farms and reducing maintenance costs. Motivated by this, this article develops a new wind turbine performance degradation monitoring scheme, which is based on pairwise comparison of the probability power curves of different wind turbines in a wind farm. Firstly, covariate matching is used to eliminate the inherent differences in meteorological variables of different turbines within the same data segment. Next, two probabilistic wind power curves, the quantile power curve and density power curve, model the functional relationship between the meteorological variables and wind power output. Then, deviation vectors are generated by calculating the deviation of probabilistic power curves between each pair of wind turbines. Finally, a directional Hotelling T2 control chart is proposed to monitor the deviation vectors. We apply the new method on the real data of a wind farm in East Britain. Results show that the proposed monitoring technique can monitor wind turbine performance degradation more precisely and comprehensively than the existing approaches.

An improved topology optimization lattice Boltzmann model for porous structures is proposed for advection- diffusion chemical reaction systems. Nonlinear porous medium drag models related to the mesoscale porosities and velocities are introduced for the momentum and mass transport coupled with the topology structure evolution. The evolution of the temporal-spatially varying porosity field is governed by a mesoscale porosity transfer equation. At each evolution step, the porous structure is sharpened by the porosity adjustment of a transient deviation of the objective function relative to porosity, and is slightly smoothed by the porosity diffusion simultaneously. Adjustable weights for chemical reaction, mass diffusion, kinetic energy, and viscous shears are applied for calculation of the transient deviation fields, which control the evolution speeds of the topology structures due to local chemical reaction, mass diffusion, flow convection, and interface drag, respectively. The effects of governing parameters and objective weights on structure patterns are illustrated. Three topology structure patterns with natural shapes are observed: (i) trunk shaped convective flow induced channels, (ii) fractal-branch shaped diffusion enhanced sub-channels, and (iii) leaf shaped chemical reaction fins. The trunk-branch-leaf topology structures result from the dynamic balance of the mesoscale convection, diffusion, and reaction.

A new mesoscale two-phase flow non-dimensional lattice Boltzmann model (NDLBM) is developed for dynamical simulations of emergence and dispersion of self-sustained vortex and energy structures in microbial suspensions. Microbial fluids are treated as two-phase fluids with interaction forces between the biofluid phase and the base fluid phase, and the bioforces are based on transient local relative velocities. Microbial concentrations, sizes, shapes, densities, viscosities, surface tensions, active indexes, self repelling coefficients, and the nonlinear drag between the two phases are included in the present model. Base fluid properties and temperatures are also included. Both macroscopic and mesoscopic governing parameters are expressed with physical means for each coefficient. Compared to previous single-phase fluid mixture absolute velocity based models, the present two-phase fluid model based on the relative velocity shows faster convergence and more stable results. Parametric studies have been performed for intensities and group sizes of active flow structures in wide ranges of physical properties and governing parameters. Vorticity and energy structure patterns and scales are illustrated to show the underlying mechanism.

An improved mesoscale two-phase flow non-dimensional lattice Boltzmann model (NDLBM) is developed for simulating dynamic balance effects of varying surface tension and interaction coefficients. The mesoscale surface tension coefficient is related to the ratio of surface tension strength to the product of the mesoscopic length scale and the equilibrium particle group size. Relative attraction and repulsion strengths among fluid particles of both different and same phases are modeled by positive and negative interaction coefficients, respectively. Both macroscopic and mesoscopic governing parameters are expressed with physical properties for each coefficient. Three basic types (liquid crystal, viscous fluid, and chain rod vorticity patterns) and two combined types (lattice viscous and stretched viscous vorticity patterns) are observed. Phase volume fraction distributions and vorticity structure patterns are illustrated to show the underlying mechanism of the balance effects of the mesoscale repulsion and attraction. The full map is shown with division curves of different pattern regimes of key mesoscale coefficients, and the division curves quantitatively show how the coefficients affect the transition of flow patterns.

The instability of wind power has a serious impact on the power grid system, and an accurate wind power forecasting is greatly demanded in practice. Due to the time-varying nature, the distribution of weather variables such as wind speed, wind direction and temperature and/or the functional relationship of the power output relating to these weather variables are likely to change over time, which are often known as "concept drifts". However, most of the prediction models usually fail to adapt to such concept drifts. This would deteriorate the prediction accuracy, especially for the short-term prediction with high accuracy requirements. Motivated by this, this paper proposes an adaptive short-term wind power prediction method, aimed at automatically detecting the occurrence of concept drifts and updating the forecast model accordingly. The proposed method consists of three main steps, extract sample-related and sequence-related features of the weather data, cluster the data based on the similarity of the features extracted, and develop a new power prediction model with automatic drift detection and adaptation for each cluster. The comparison results show that the prediction performance of the proposed method is superior to that of the competitive methods.

Metal foams with larger interface areas and higher effective conductivities are widely used to increase heat transfer in engineering applications. To reveal the quantitative relation between the pore scale and the macroscale convective thermal transport, foams immersed in a water cooling channel for two high temperature flat plates were simulated by the mesoscale non-dimensional lattice Boltzmann method. The generated structure parameters are global porosities from 0.70 to 0.95 and pore sizes from 6% to 16% of the square channel height. The flow simulation parameters are macroscopic Reynolds numbers from 50 to 1500 and pore scale Reynolds numbers from 3 to 240. Three-dimensional spirals inside foam pores and behind foam structures, which induce the third spatial direction convection, are observed to be positively related to pore sizes, Reynolds numbers, and the volume fractions of fluid and solid phases, respectively. The pore scale drag and heat transfer coefficient constants are inversely correlated from the pressure drop and heat flux of mesoscale simulation results with deviations of 13.2% and 12.5%, respectively. To-gether with the pore scale geometry factor and the global porosity, the correlations of pore scale and macroscopic scale drag and heat transfer are bridged by microscopic coefficient constants (0.0 0 04, 4.59, 0.47) and (0.0 0 01, 0.096, 0.50, 0.33), respectively. The cross scale correlations provide a practical tool for further engineering applications of enhancing convective cooling through porous structures.(c) 2023 Elsevier Ltd. All rights reserved.

Mesoscopic scale simulations on heat transport in porous structures are shown with the coupled structure generation and transport simulation methods based on probability distribution functions in Gaussian quadrature grid and velocity sets (GQGVS). The mesoscopic scale method for various structure generation named a controllable structure generation scheme (CSGS) may generate not only multiple-phase isotropic homogenous random structures, but also shape-constrained anisotropic heterogeneous structures through relatively simple constraint integer indices with a high accuracy of global phase volume fractions. A nondimensional lattice Boltzmann method (NDLBM)-based solver for coupled transport equations is applied for direct mesoscopic scale simulations of thermal diffusion and natural convection in generated structures described by the local phase volume fraction matrix. Variable timesteps with transient mesoscopic March numbers and stretched space steps with local coordinate transformation gradients are discussed. The effective parallel NDLBM scheme with minimum information transferred during the data communication between subdomains is introduced. Two- and three-dimensional thermal diffusion and natural convection through various porous structures including random, sphere, fiber, and porous pellet packed structures are illustrated to provide a better mesoscopic scale understanding of the mechanism of heat transport in complex structures.

Nanofluid cooling of a concentrated photovoltaic thermal (CPVT) receiver was simulated by a sub-continuous lattice Boltzmann model with the effective thermal conductivity (ETC) and the effective viscosity (EV) nonlinearly related to both nanoparticle concentration and size. Al2O3-water nanofluid cooling efficiencies for various solar irradiance are compared with those of pure water cooling. Flow and temperature fields are simulated for nanofluids with the nanoparticle concentration from 1% to 10%, particle size less than 120 nm, and flow rate over a range of 0.17-3.34 L/min (i.e., the inlet velocity from 1/10 to 2 times of the natural convection velocity scale). In dimensionless form, the parameters are described by concentration, Knudsen number and Richardson number. The enhancement ratios of Nusselt numbers, drag coefficients, and power coefficients due to the application of nanofluids compared to water are presented. An objective enhancement function is defined as the ratio of the Nusselt number to the power coefficient. The maximum enhancement ratio is 1.14 for nanoparticle concentration at 8%, Knudsen number at 0.1 (Al2O3 nanoparticle size 6 nm), and Richardson number 10 (the inlet velocity about 1/3 of the natural convection velocity scale), respectively. This study provides a practical tool for optimal nanofluid cooling enhancement of CPVT solar receivers. (c) 2021 Elsevier Ltd. All rights reserved.

Quartet structure generation set (QSGS) method is improved to numerically reconstruct three-dimensional heterogeneous porous wicks with uniform pore size distribution. The capillary pumping processes of the reconstructed random porous wicks are simulated at pore scale by using a three-dimensional two-phase lattice Boltzmann model. The evolutions of two-phase interface and the variations of the imbibed liquid volume fraction with time are analyzed under the conditions of different porosity, pore structure, and surface wettability. The comparisons between the LBM results and those predicted by a macroscopic scale homogenous model are also conducted. It is found that due to the pore scale effects, the two-phase interface in a random porous media is very irregular, especially at the earlier stages of capillary pumping process when the liquid penetration is faster, and the imbibed liquid does not increase exponentially with time as predicted by the macroscopic scale model. Meanwhile, the liquid penetration rate does not decrease monotonously, but exhibits different degrees of fluctuations. The pore scale effects are more prominent in the cases of lower porosity, smaller pore size and better surface wettability. In the parametric range of the present study, the capillary performance increases with the decreasing of porosity, average pore radius (when porosity is fixed) and contact angle.

This paper develops a new hybrid forecasting model for the hourly power output of multiple turbines (or plants), aimed at making use of both the spatial and temporal correlations of weather factors and power outputs from wind turbines (or plants). To account for the time-varying nature of local pattern and wind propagation, a new clustering algorithm named Kmeans–Hierarchical Clustering (KHC) is first used to adaptively cluster the wind turbines (or plants). Then, singular value decomposition (SVD) is employed to extract the leading components of the wind power of the wind turbines (or plants) in the same cluster. Finally, the support vector regression (SVR) models are built for the leading components and the predictions of the components are transformed into the predictions of the power output of each turbine (or plant) with the inverse SVD. The comparison results show that the proposed model is able to improve the forecast accuracy over the existing short-term forecasting models.

To investigate effects of particle sizes on natural convection of nanofluids, nanoparticle sizes together with mean free paths in solid and fluid phases are used to define two Knudsen numbers Kn(d,s) and Kn(d, f) , respectively. These two Knudsen numbers are further applied to build the new effective conductivity and viscosity models, which include non-linear effects of both Knudsen numbers and volume fractions. A sub continuous non-dimensional lattice Boltzmann model is built to compute the velocity and temperature fields. Effective thermal conductivity and base fluid conductivity based Nusselt numbers are compared. Heat flux enhancement ratios relative to pure base fluid are calculated. Nusselt numbers and enhancement ratios in wide ranges of key governing parameters over 10(-6) <= K-nd,K- f <= 10(4) , 10(3) <= Ra-f,Ra-L <= 10(6) , and 10(-3) <= phi(s) <= 10(-1) are plotted. The optimal governing parameters and corresponding nanoparticle sizes are obtained from the full maps. The changes of heat transfer patterns with varying Knudsen numbers are also illustrated by temperature and streamline fields. Results agree with various experimental and simulation results. Also, the present sub-continuous models and simulation results provide a correlation with all parameters for practical engineering applications of nanofluids with effects of particle sizes. (c) 2022 Elsevier Ltd. All rights reserved.