Currently, grid-following (GFL) control is widely adopted in direct-drive wind farms. Its external impedance characteristic exhibits negative resistance and capacitive reactance, frequently inducing sub/super-synchronous oscillations in the direct-drive wind farm and weak grid interactive system. The positive resistance characteristic of grid-forming (GFM) control can, to a certain extent, improve the impedance characteristic of wind farms and enhance the system stability margin. However, the influence of the proportion and deployment location of GFM control within a wind farm on the mitigation of sub/super-synchronous oscillations merits further exploration. First, this paper establishes the sequence impedance models for both GFL and GFM control, analyzes the underlying causes of system oscillations from an impedance perspective, and proposes a method for calculating the stability margin of a grid-connected direct-drive wind farm system that comprehensively accounts for the generalized short-circuit ratio, the critical short-circuit ratio of the equipment, and the steady-state operational constraints of the system. Subsequently, the mitigation effects of the connection location and capacity proportion of GFM wind turbines on sub/super-synchronous oscillations are quantitatively assessed, yielding feasible ranges of the short-circuit ratio under various operating conditions that ensure stable operation of the direct-drive wind farm. The system stability is further examined via Nyquist curve analysis. Finally, the effectiveness of the proposed method is validated by electromagnetic transient simulations in MATLAB/Simulink.

To enhance the reliability of fusion power protection systems, an integrated "fault-thermal-structure" co-design framework for the bypass circuit in the Comprehensive Research Facility for Fusion Technology (CRAFT) converter is proposed in this study. System-level models for three critical fault scenarios-magnet quench, load short-circuit, and mistake trigger-are established using PSCAD simulations to quantify the associated fault currents. Junction temperature limits of 125 degrees C for 5STP 50Q1800 thyristors are validated through a thermal resistance network, determining six parallel thyristors as the optimal configuration with 119 degrees C junction temperature under worst-case conditions. Novel metrics, Current Range Value (CRV) and absolute peak current dispersion (|D|), are proposed to overcome limitations of the conventional current-sharing coefficient (K). Subsequently, intelligent parametric optimization is implemented through ANSYS Q3D/Optimetrics, culminating in an asymmetric busbar structure. Validation under severe mistake trigger without electronic protection demonstrates a peak branch current reduction from 18.33 kA to 17.78 kA, a CRV reduction by 89.8% from 1470 A to 150 A, and |D| confinement below 1.1%. Similar enhancements are observed across other fault scenarios. A systematic methodology for fusion bypass design, integrating fault analysis, thermal verification, and structural optimization for improved current sharing and reliability, is established by this research.

This paper presents a symplectic Hamiltonian direct discontinuous Galerkin (DDG) method for approximating wave propagation problems, including the linear and semilinear wave equations. Within an auxiliary-variable-free DG framework, we prove that the symmetry of the numerical flux bilinear form is equivalent to the existence of a discrete Hamiltonian structure. It follows that methods such as the symmetric interior penalty method and the symmetric DDG (SDDG) method admit a discrete Hamiltonian structure, whereas schemes including the Baumann–Oden, DDG, and BR2 methods do not possess this property. Exploiting this structure, we construct fully discrete symplectic schemes by combining the SDDG spatial discretization with symplectic time integrators. We further derive error estimates for the SDDG method applied to semilinear wave equations, showing the optimal convergence rate for the displacement and the suboptimal convergence rate for the velocity. Numerical experiments validate the theoretical convergence rates and demonstrate that the symplectic Hamiltonian DDG method achieves superior long-time energy conservation and accuracy.

In this work, we develop a multidomain hybrid discontinuous Galerkin (DG) method and finite difference (FD) method for solving two-dimensional compressible Navier-Stokes equations on the hybrid meshes. The direct discontinuous Galerkin (DDG) method and central difference (CD) scheme are utilized to discretize the viscous fluxes respectively. This approach combines the flexibility for the complex geometries of the DG method on the unstructured meshes, and the computational efficiency of the FD method on Cartesian grids. At the artificial interfaces between the DG subdomain and FD subdomain, the square ghost cells are generated and the weighted essentially non-oscillatory (WENO) interpolation is employed to reconstruct the degrees of freedom of these ghost cells. To ensure the accuracy in smooth regions and the correct position of the shock wave, the troubled cell indicator is adopted to determine the nonconservative or conservative coupling modes. The construction process of the numerical fluxes at the artificial interfaces is described specifically and the corresponding WENO interpolation coefficients are given in detail. Numerous numerical results demonstrate that the multidomain hybrid method achieves high-order accuracy in smooth regions, robustness in shock simulations, flexibility in handling complex geometries, and significant computational cost savings compared to the traditional DG method on the hybrid meshes.

Based on recent work by He et al. (Computer Physics Communications 286 (2023): 108660), this work develops a robust and efficient workflow for 3D aerodynamic shape optimization (ASO) based on the discontinuous Galerkin method (DGM) with solution remapping technique. It is theoretically found and numerically validated through 2D cases that the DGMs can provide more precise adjoint-enabled gradients even on a coarse mesh as compared with the FVMs under coequal computational costs. A number of 2D and 3D intricate cases are tested to illustrate the potential advantages of the DGM-based ASO workflow in terms of computational efficiency and final optimization results. The results demonstrate that the solution remapping technique can save around 50%$$ 50\% $$ of the computational time. Moreover, the DGM-based workflow extends the capability to explore the design space, leading to aerodynamic shapes with superior performance.

Current unbalance caused by impedance disparities among multiple parallel branches can lead to accelerated aging of some power devices and localized overheating, posing a threat to the normal operation of fusion power supplies. A method for improving the current sharing performance of fusion converters with paralleled branches based on structural parameter optimization is proposed in this paper. Firstly, a high-precision converter bridge arm structure is constructed using ANSYS Q3D, and sub-circuits of segmented modules are extracted; subsequently, a circuit model is built in Simplorer, and the current of each branch is obtained through joint simulation with Simplorer; then, with branch current as the matching parameter, an adaptive target optimization algorithm is utilized to iterate and optimize structural parameters, thereby acquiring a set of bridge arm structures with optimal current sharing performance; finally, fine-tuning is conducted considering the actual spatial limitation. To validate the superiority of the proposed method, the current sharing effects of the structures before and after optimization under both steady-state and short-circuit conditions are analyzed and compared. The results indicate that the optimized structure has shown significant improvements in terms of current sharing coefficient, over-current ratio, and maximum turn-off error time. This method could be used for the current sharing design of fusion power supplies and related multiple parallel devices.

Phylodynamic inference (PI) quantifies population dynamics and evolutionary trajectories using phylogenetic trees. Single-cell lineage tracing enables phylogenetic tree reconstruction for thousands of cells in multicellular organisms, facilitating PI at the cellular level. However, cell differentiation and somatic evolution challenge the direct application of existing PI frameworks to somatic tissues. We introduce scPhyloX, a computational framework modeling structured cell populations by leveraging single-cell phylogenetic trees to infer tissue development and tumor evolution dynamics. A key advancement is its ability to infer time-varying parameters, capturing dynamic biological processes. Simulations demonstrate scPhyloX’s accuracy in scenarios including tissue development, disease treatment, and tumor growth. Application to three real datasets reveals insights into somatic dynamics: cycling stem cell overshoot in fly organ development, clonal expansion of multipotent hematopoietic progenitors during human aging, and pronounced subclonal selection in early colorectal tumorigenesis. scPhyloX thus provides a computational approach for investigating somatic tissue development and evolution.

This paper implements the CUDA and hybrid CUDA/MPI parallel computation based on GPGPU heterogeneous parallel strategies for the direct discontinuous method (DDG) on 3D unstructured grids. The direct discontinuous Galerkin method inherits the compactness of the discontinuous Galerkin (DG) method, making it well-suited for large-scale parallelization. Firstly, we present the full single-GPU implementation of the three-dimensional (3D) DDG method with cell-level parallelism and face-level parallelism. Herein, all the numerical operators including volume integration, face integration (numerical fluxes), conservation variables calculation, and time iteration, are implemented by designing the corresponding kernel functions. Especially, we implement several key memory access optimization strategies, which are crucial for performance improvement. Operators merging and shared memory utilizing reduces the number of global access. Such memory Coalescing and data structure reconstruction apparently enhances the efficiency of global memory access. To align with data access pattern, we employ atomic operations to eliminate data race conditions. Furthermore, we propose a full hybrid GPU/CPU heterogeneous parallel strategy to implement multi-GPU parallelization of the DDG method, where asynchronization optimization is introduced to fully overlap communication and computation and basically eliminates the communication overhead. Finally, several numerical tests are conducted on Tesla V100 Cards to show performance of the parallelization. In addition, we utilize the NVIDIA performance testing tool, nvprof, to evaluate multiple metrics of the kernel functions and conduct a detailed analysis of the results. In the tests of parallel scalability, the weak scaling efficiency achieves 97% from 4 to 32 GPU cards, and the strong scaling efficiency is 90% from 1 to 8 GPU cards.

Physics-informed Neural Network (PINN) faces significant challenges when approximating solutions to conservation laws, particularly in ensuring conservation and accurately resolving discontinuities. To address these limitations, we propose Conservation Law-informed Neural Network (CLINN), a novel framework that incorporates the boundedness constraint, implicit solution form, and Rankine-Hugoniot condition of scalar conservation laws into the loss function, thereby enforcing exact conservation properties. Furthermore, we integrate a residual-based adaptive refinement (RAR) strategy to dynamically prioritize training near discontinuities, substantially improving the network's ability to capture sharp gradients. Numerical experiments are conducted on benchmark problems, including the inviscid Burgers equation, the Lighthill-Whitham-Richards (LWR) traffic flow model, and the Buckley-Leverett problem. Results demonstrate that CLINN achieves superior accuracy in resolving solution profiles and discontinuity locations while reducing numeral oscillations. Compared to conventional PINN, CLINN yields a maximum reduction of 99.2

A class of reconstruction-based high-order conservative central difference schemes (CD) is developed for solving the compressible Navier-Stokes equations in this paper. The discretization of the viscous and heat fluxes in the Navier-Stokes equations involves a two-step process, where the external first derivatives of the viscous terms are disposed in a reconstruction way, and then the interpolation operation is carried out to calculate the internal first derivatives within the same stencil. Two approaches to the interpolation implementation are discussed: one is founded on the conservative variables, while the other is based on the primitive variables. This design can maintain compactness and consistence as in the stencil of the weighted essentially non-oscillatory (WENO) schemes for the inviscid terms. Under the present framework, a sixth-order central difference scheme for the viscous terms is designed with a stencil width that falls within the range of the fifth-order WENO scheme for the inviscid terms. The accuracy for both linear and nonlinear diffusion equations are demonstrated theoretically and the spectral properties are verified via Fourier analysis. Numerous compressible viscous results validate that the present central difference schemes are high-order accurate in smooth regions, easy to implement, robust for the viscous shock simulations and computationally cost-effective.

Fe-Cr-Ni alloys, widely used in industrial applications and serving as crucial high-temperature materials, require accurate interatomic potential functions for understanding their microstructure and properties. In this study, we developed a machine learning potential (Deep Potential, DP) with high precision for Fe-Cr-Ni alloys using the Deep Potential for Molecular Dynamics (DeePMD) framework combined with an extensive density functional theory (DFT) database. By integrating training databases generated through first-principles calculations covering pure metals, binary, and ternary alloy systems, a deep neural network was employed to encode atomic-level energies, forces, and local environmental features, enabling accurate prediction of multi-body potentials for complex alloy systems. Model validation demonstrates that the root-mean-square errors for energy and force are controlled within 3.27 meV/atom and 72.4 meV/& Aring;, respectively, achieving high consistency with DFT results. Further reliability verification was conducted through energy-volume curve analysis, radial distribution function (RDF) evaluations, and mechanical parameter calculations. The yield strength of Fe33.3Cr33.3Ni33.3 (Fe1) alloy shows a simulation-to-experiment deviation <= 5 %, while the elastic modulus of Fe60Cr25Ni15 (Fe2) alloy (101.32 GPa) exhibits less than 1 % deviation from the DFT-calculated value (102.22 GPa). This potential function maintains DFT-level accuracy while ensuring computational efficiency for large-scale molecular dynamics simulations. It provides theoretical support for Fe-Cr-Ni alloys in nuclear materials, high-temperature corrosion applications, and expands the applicability of machine learning potentials in multi-principal element alloy design.

A high-order compact finite difference scheme for solving natural convection problems using velocity-vorticity formulation of the incompressible Navier-Stokes equations is presented. The basic idea of the method is to regard all controlling equations as the Poisson-type. We construct a fourth-order finite difference scheme for the velocityvorticity equation based on the nine-point stencils for each Poisson-type equation. Next we give an example with an exact solution to verify that the scheme has the fourth-order accuracy. Finally, numerical solutions for the model problem of natural convection in a square heating cavity are presented to show the reliability and effectiveness of this method.

When the d-T fusion reaction takes place in a Tokamak, the high-energy neutrons emitted, with an energy of 14.1 MeV, can alter the electrical properties of high-power thyristors in Quench Protection System (QPS). The Reverse Recovery Characteristics (RRC) of High-power Thyristor (HP-SCR) is one of the crucial issues affecting the reliability of QPS. In this paper, the change in the RRC of HP-SCR under 14.1 MeV neutron irradiation is deeply studied. Firstly, the microscopic material damage mechanism of HP-SCR induced by neutron irradiation and its relationship with RRC are deeply analyzed. Secondly, a highly efficient neutron irradiation experiment is designed and conducted to effectively validate the correctness of the theoretical analysis regarding changes in RRC. Finally, the effects of neutron irradiation on the QPS, consisting of multiple thyristors arranged in series, are analyzed and discussed through system-level simulations. The study offers important recommendations for the maintenance and upgrade of QPS, which will greatly enhance the safety of Tokamak devices.

Geological carbon sequestration (GCS) and gas energy storage in water-bearing formations, one of the most promising methods of Carbon Capture, Utilization, and Storage (CCUS), offers a permanent and reliable solution for mitigating climate change. Although reservoir simulation is widely employed to evaluate GCS projects, its high computational cost, driven by the complexity of phase equilibrium calculations, poses a challenge to obtain timely assessments. To address this, we propose an innovative Adaptive Saturated Composition (ASC) algorithm to bypass unnecessary stability analyses, therefore reducing the computational expenses of gases-brine phase equilibrium calculations in GCS simulation. The ASC algorithm relies on pre-calculated binary gas-brine data and is extended to multicomponent systems (CH4, N2, H2S, CO2) through a mixing rule. We performed extensive tests under varying underground conditions, demonstrating that the proposed ASC algorithm offers superior efficiency, accuracy and scalability compared to existing algorithms. It eliminates over 99.9 % of unnecessary stability analyses in binary systems and reduces their frequency by over 93 % in multicomponent gas systems. Remarkably, the ASC algorithm achieves accurate detection of phase transitions across all tested scenarios. Moreover, it is widely applicable in most of water-bearing geological gas storage scenarios, including natural gas storage, geological CO2/air energy storage, and CO2-based geothermal systems.

In the design realm of fusion power supplies, structural components play a pivotal role in ensuring the safety of fusion devices. To verify the reliability of the converter structure design at the Comprehensive Research Facility for Fusion Technology (CRAFT), meticulous analysis of the converter's dynamic impact is carefully performed based on the worst fault current (400 kA), firstly. Subsequently, the thermal stress analysis based on the maximum allowable steady-state temperature is finished, and the equivalent thermal stress, thermal deformation, maximum shear stress of a single bridge arm and the whole converter are studied. Furthermore, a simple research method involving the current-sharing characteristics of a bridge arm with multi-thyristor parallel connection is proposed using a combination of Simplorer with Q3D in ANSYS. The results show that the current-sharing characteristics are excellent. Finally, the structural design has been meticulously tailored to meet the established requirements.

In this paper, an Efficiency Optimization Strategy (EOS) is proposed to address the issue of low efficiency in the Dual Active Bridge (DAB) DC-DC converter across certain power ranges under wide voltage conditions, with the aim of further enhancing the converter's efficiency over a broad operating range. First, in the low power range, a phase-shift control strategy is introduced, which enables wide-range Zero Voltage Switching (ZVS) and near-optimal inductor current RMS values. Through this strategy, ZVS is ensured for all switches under light load conditions, while under medium load conditions, ZVS is lost for only two switches. Subsequently, in the high power range, the optimization target is smoothly transitioned to the optimal RMS current value by utilizing the natural ZVS characteristics of the DAB converter. The operating range of the EOS is effectively extended, further reducing current stress and RMS current values, thereby achieving global efficiency optimization of the DAB converter. Finally, an experimental platform is constructed for verification, and the correctness and effectiveness of the theoretical analysis are confirmed by the experimental results.

The capillary pressure (Pc) between the aqueous phase and the carbon dioxide (CO2)-rich phase plays a pivotal role in geological CO2 storage (GCS) due to its ability to augment the swept volume. A substantial number of experimental studies showed that Pc strongly depended on the formation pressure. However, the capillary pressure fluctuation, arising from the pressurization of CO2 injection, has not been considered in previous GCS numerical simulation studies due to the difficulty in quantifying its impact. In this paper, we developed a fully implicit parallel compositional GCS simulator (PCOMP-GCS) based on our in-house parallel platform. The simulator is equipped with fast and robust multiple-phase flash algorithms to accurately describe the phase behavior of a depleted gas-CO2-brine system. The validation results demonstrate the reliability and capability of our simulator in emulating the CO2 sequestration processes. Meanwhile, a generalized pressure-saturation-dependent capillary pressure model is proposed to assess the Pc under various underground pressure and saturation. The new model integrates macroscopic and microscopic experimental data through Young–Laplace’s equation, enabling it to consider both the effects of pressurization and saturation variations on Pc. Compared to traditional capillary pressure measurements, the data used in this model is highly cost-effective and time-efficient. Moreover, a comprehensive case study is conducted to quantitatively investigate the impact of Pc fluctuation, induced by the injection pressurization, on a field scale CO2 storage problem. This case study results reveal that the CO2 distribution, trapping mechanisms, storage capacity and leakage quantity have significant changes when the influence of pressurization on Pc is taken into consideration.

Abstract Storing CO2 in deep saline aquifers is one of the most promising methods for achieving carbon neutrality. Numerical reservoir simulation can assist researchers and engineers in comprehensively understanding and effectively managing the CO2 sequestration process, thereby ensuring its safety and effectiveness. Most of the current CO2 storage simulators use a CO2-brine fluid model which only has water and CO2 components. When impure CO2 is injected, the subsurface fluids form a multi-component gas-brine system, whose phase behavior cannot be accurately described by a CO2-brine fluid model. The current gas-water compositional models include a very limited number of gas components. In this study, we developed a fully implicit parallel CO2 storage simulator for distributed memory computers based on our in-house parallel platform. This simulator uses a fully compositional fluid model which treats water as a component and incorporates it into phase equilibrium calculations. The gas components include not only CO2 but also other common components in captured impure CO2, such as CH4, N2 and H2S. The cubic Peng-Robinson equation of state was used to predict the fugacity and PVT properties of CO2-rich phase. The fugacity in aqueous phase and brine properties are calculated by Henry's law and empirical correlations respectively. Meanwhile, the effect of dissolved gas influence on aqueous viscosity is considered which is ignored by current commercial simulator. The finite difference (volume) method is applied to discretize the compositional fluid model. Numerical experiments show that our simulator is scalable, stable and validated to simulate large-scale CO2 storage problems with hundreds of millions of grid blocks on a parallel supercomputer cluster.

Phylodynamics inference (PI) is a powerful approach for quantifying population dynamics and evolutionary trajectories of natural species based on phylogenetic trees. The emergence of single-cell lineage tracing technologies now enables the reconstruction of phylogenetic trees for thousands of individual cells within a multicellular organism, opening avenues for employing PI methodologies at the cellular level. However, the intricate process of cell differentiation poses challenges for directly applying current PI frameworks in somatic tissues. Here, we introduce a novel computational approach called single-cell phylodynamic explorer (scPhyloX), designed to model structured cell populations in various cell states, by leveraging single-cell phylogenetic trees to infer dynamics of tissue development and tumor evolution. Our comprehensive simulations demonstrate the high accuracy of scPhyloX across various biological scenarios. Application of scPhyloX to three real datasets of single-cell lineage tracing unveils novel insights into somatic dynamics, such as the overshoot of cycling stem cell populations in fly organ development, clonal expansion of multipotent progenitors of hematopoiesis during human aging, and pronounced subclonal selection in early colorectal tumorigenesis. Thus, scPhyloX is an innovative computational method for investigating the development and evolution of somatic tissues.

Storing CO2 in deep aquifers and depleted gas reservoirs is an effective way to achieve carbon neutrality. However, the numerical simulation of CO2 storage in these formations is challenging due to the complexity of gases-brine systems. The number of gas species included in the gases-brine fluid models of existing simulators cannot meet the rapidly evolving CO2 sequestration scenarios. To address this intricate issue, we developed a three-dimensional fully implicit parallel CO2 geological storage simulator (PRSI-CGCS) on distributed-memory computers based on our in-house parallel platform. This simulator uses a compositional fluid model with a diverse range of gas species, including CO2, C1 ~ C3, N2, H2S, as well as newly added gases H2 and O2, which may be encountered in geological CO2 storages. Besides, we provide more suitable scaling factors for different gases in the stability analysis bypassing (SAB) method to accelerate the gases-brine phase equilibrium calculations. PRSI-CGCS does not incorporate energy conservation equations, and salt precipitation or dissolution is also not considered. Numerical experiments show that our simulator is scalable, robust and validated to simulate large-scale CO2 storage problems with hundreds of millions of grid blocks on a parallel supercomputer cluster. Besides, after our modification on scaling factors, the SAB method can reduce the number of stability analyses by 61.39 % to 88.71 %, thereby reducing simulation time. Furthermore, case studies indicate that injecting O2 and H2 along with CO2 reduces the stability or capacity of CO2 storage and increases the pressure required for injection. However, this impact is not significant when the impurity content is less than 10 %.