Multi-domain Optimization Research Articles

Towards digital twinning for multi-domain simulation workflows in urban design: a case study in Gothenburg

This work proposes an automated workflow using digital twinning for multi-domain environmental performance analysis of urban developments. Digital twins can potentially provide a common basis for multi-domain simulations and help overcome data availability and interoperability issues. The proposed workflow consists of five steps: (1) creating a procedural urban 3D model, (2) generating design alternatives parametrically, (3) exporting the context and each design alternative to each simulation tool, (4) running simulations for wind comfort, energy demand, and noise for each design alternative, and (5) combining and visualizing the simulation results using the digital twin. The workflow was applied to a neighbourhood in Sweden, the resultsreveal significant reduction in manual work when applying multiple simulation software for different domains. This is one step forward in streamlining the workflow for urban analysis, crucial for multi-domain optimization. In the future, further domains and simulation tools can be added to the workflow.

Multidomain Optimal Control of Variational Subpotential Mixed Evolution Inclusions

Multidomain variational optimal control of evolution mixed subpotential inclusions, are formulated and analyzed, on the basis of a perturbation conjugate duality convex analysis theory developed by the author. For Lagrangian optimality mixed conditions, fixed point existence results are demonstrated with an strongly monotone qualifying condition. Governing multidomain state systems correspond to primal evolution macro-hybrid mixed subpotential problems, whose solvability is similarly achieved. Innovative multidomain optimization existence results of primal, dual, Lagrangian mixed, as well as coupled pair state-control problems are established. Applications to underground macro-hybrid mixed control transport flow processes, illustrate the theory.

MDST: multi-domain sparse-view CT reconstruction based on convolution and swin transformer

Objective.Sparse-view computed tomography (SVCT), which can reduce the radiation doses administered to patients and hasten data acquisition, has become an area of particular interest to researchers. Most existing deep learning-based image reconstruction methods are based on convolutional neural networks (CNNs). Due to the locality of convolution and continuous sampling operations, existing approaches cannot fully model global context feature dependencies, which makes the CNN-based approaches less efficient in modeling the computed tomography (CT) images with various structural information. Approach. To overcome the above challenges, this paper develops a novel multi-domain optimization network based on convolution and swin transformer (MDST). MDST uses swin transformer block as the main building block in both projection (residual) domain and image (residual) domain sub-networks, which models global and local features of the projections and reconstructed images. MDST consists of two modules for initial reconstruction and residual-assisted reconstruction, respectively. The sparse sinogram is first expanded in the initial reconstruction module with a projection domain sub-network. Then, the sparse-view artifacts are effectively suppressed by an image domain sub-network. Finally, the residual assisted reconstruction module to correct the inconsistency of the initial reconstruction, further preserving image details. Main results. Extensive experiments on CT lymph node datasets and real walnut datasets show that MDST can effectively alleviate the loss of fine details caused by information attenuation and improve the reconstruction quality of medical images. Significance. MDST network is robust and can effectively reconstruct images with different noise level projections. Different from the current prevalent CNN-based networks, MDST uses transformer as the main backbone, which proves the potential of transformer in SVCT reconstruction.

Multi-domain optimization of cast iron components in wind turbines

The continuously rising demand for renewable energies leads to increased installations of wind turbines with higher power. While the current power-to-weight ratio of up to 20 metric tons of cast iron per megawatt is stagnating, cast iron components of modern wind turbines are facing new challenges in terms of weight, manufacturability, and castability. These challenges can be addressed by systematically using multi-domain optimization approaches to reduce component weight and increase local component utilization.In order to meet the requirements for modern cast iron components, this multi-domain approach must employ methods from casting simulation, micromechanical analysis, topology optimization, and strength assessment. Here, casting simulation is used to determine local microstructure descriptors, which are subsequently used in micromechanical shakedown analysis to estimate the local microstructure-dependent fatigue strength. In parallel to the fatigue strength estimation, topology optimization is performed iteratively in combination with a castability analysis. The component strength is evaluated using a strength assessment approach based on the previously determined local material properties in combination with the topology optimized component.In this study, the overall concept of the proposed multi-domain approach is presented and requirements for the application of such an approach are formulated. The use case of this study is a planet carrier of a wind turbine gearbox manufactured from austempered ductile cast iron ADI-GJS-1050‑6. For this use case, a weight reduction of 17% was achieved while maintaining the required stiffness, such that the microstructure variance along the component was significantly reduced. Furthermore, the potentials and limitations of the presented approach are outlined and discussed in the context of the design of heavy-section castings.

Design and Optimization Approach for the Magnetic Components of a Five-Level Current Source Rectifier

This work presents a design and optimization method of the five-level current source converter magnetics components, this is a contribution of this work, since the magnetics volume information of this topology lacks in the literature. The magnetics components are: the interphase transformers, the input and the output filter inductors. The converter volume is minimized when the magnetic components multi-domain optimization is done. The inductance values are defined with given constraints such as Electromagnetic Compatibility standards for differential mode filters, conduction and switching losses of the semiconductors and maximum current variation. This optimization is shown for two different converter switching frequencies: 45 kHz and 150 kHz. The copper losses, iron losses and temperature rise are calculated and used for selecting the smallest volume of each magnetic component with given specifications. A 6-kW prototype with the optimized magnetics was constructed

A thermal and acoustic co-simulation method for the multi-domain optimization of nearly zero energy buildings

In the last years, the Nearly Zero Energy Building (or NZEB) concept has become one of the main targets to drive the built environment towards a sustainable future. Advanced methodologies and tools have been developed to support the complex NZEB design problem considering multiple objectives and the renovated interest for building occupants’ wellbeing has highlighted the need for integrating energy and comfort aspects for a holistic vision of the future NZEB. This work intends to advance in the ongoing discussion about comfort-driven NZEB design and develop a new multi-tool multi-objective simulation-based optimization methodology to integrate energy, cost and acoustic performance in the search for the NZEB optimal design. The proposed approach relies on the coupling between an energy model in TRNSYS®, a sound insulation model developed in Matlab® and the PSO algorithm to drive the optimization through the GenOpt® tool. After theoretical elaboration of the so-defined optimization problem, the method was applied to a French single-family case study. The applied methodology identifies several non-obvious solutions that are near-optimal for all the three objectives and shows that the traditional cost-optimized NZEB design would lead to low sound insulation performance, revealing the potential for a full exploitation of synergies between thermal and acoustic domains in high-performance buildings, reinforcing the need for further investigation in the field.

Integrated Analysis and Optimization of the Large Airborne Radome-Enclosed Antenna System

In order to realize integrally analysis and optimization of the large airborne radome-enclosed antenna system, a novel optimization strategy is proposed based on an overlapping domain decomposition method by using higher-order MoM and out-of-core solver (HO-OC-DDM), and combining with adaptive mutation particle swarm optimization (AMPSO). The introduction of parallel out-of-core solver and DDM can effectively break the random access memory (RAM) limit. This strategy can decompose difficult-to-solve global optimization problems into multi-domain optimization problems by using domain decomposition method. Finally, take airborne Yagi antenna system as an example, the numerical results show that the design of large airborne radome-enclosed antenna system based on the proposed strategy is convenient and effective.

Multi-domain optimization of the eigenstructure of controlled underactuated vibrating systems

The paper proposes a multi-domain approach to the optimization of the dynamic response of an underactuated vibrating linear system through eigenstructure assignment, by exploiting the concurrent design of the mechanical properties, the regulator and state observers. The approach relies on handling simultaneously mechanical design and controller synthesis in order to enlarge the set of the achievable performances. The underlying novel idea is that structural properties of controlled mechanical systems should be designed considering the presence of the controller through a concurrent approach: this can considerably improve the optimization possibilities. The method is, first, developed theoretically. Starting from the definition of the set of feasible system responses, defined through the feasible mode shapes, an original formulation of the optimality criterion is proposed to properly shape the allowable subspace through the optimal modification of the design variables. A proper choice of the modifications of the elastic and inertial parameters, indeed, changes the space of the allowable eigenvectors that can be achieved through active control and allows obtaining the desired performances. The problem is then solved through a rank-minimization with constraints on the design variables: a convex optimization problem is formulated through the “semidefinite embedding lemma” and the “trace heuristics”. Finally, experimental validation is provided through the assignment of a mode shape and of the related eigenfrequency to a cantilever beam controlled by a piezoelectric actuator, in order to obtain a region of the beam with negligible oscillations and the other one with large oscillations. The results prove the effectiveness of the proposed approach that outperforms active control and mechanical design when used alone.

Thermodynamic loss mechanisms and strategies for efficient hot-electron photoconversion

There are currently extensive interests on the hot-electron-mediated photoconversion. However, the device quantum yield is fundamentally low due to the existences of various hot-electron loss channels; moreover, the nanostructured plasmonic/metamaterial are generally required, which bring challenges to the low-cost and large-scale fabrication. In this study, we focus on distinguishing the thermodynamic losses in hot-electron devices and presenting the possible route-maps for performance improvement. It is shown that a number of optical, electrical, and material factors, which lead to the substantial losses of hot electrons during the generation, transport, and emission processes. These loss mechanisms involve extensively the photon absorption, resistive dissipation, electron/electron or electron/phonon thermalization, carrier diffusion, Schottky barrier, and electron-momentum conservation. We further exemplify several planar hot-electron systems to show the possibilities of breaking these limitations for high efficiency. The planar hot-electron devices are based on Tamm plasmons, microcavity with double barriers, and optimized system by controlling the barrier and electron density of state, respectively. Results indicate that these designs can significantly improve the efficiencies of hot-electron generation, transport, and collection. It reveals that the external quantum efficiency of the system after the multi-domain optimization can be up to 60% in the near-infrared band. This study will motivate deeper understanding on the physical mechanisms, which restrict the performance of hot-electron device, and provide the solutions to improve the performance of hot-electron devices.

Optimization of Low-Power Brushless PM-Machines for Automotive Applications With Focus on High-Volume Mass Production

This paper deals with the optimization for performance and cost of the low-power permanent magnet synchronous machines used in automotive applications with focus on particularities of high-volume mass production and the constraints that arise. The described optimization loop is used to compare different 12/8 and 12/10 designs of interior permanent magnet (IPM) type. After showing that the 12/10 IPM exhibits the best performance–cost ratio, results obtained at different stages in the process (i.e., sensitivity analysis and constrained single-/multiobjective optimizations), for both no-load and load case, are provided. Finally, different aspects for reducing the complexity of the multidomain optimization problem (e.g., 3-D effects, iron losses, and effects of tolerances on the motor characteristics) are addressed.

Weight Minimization of LCL Filters for High-Power Converters: Impact of PWM Method on Power Loss and Power Density

For high-power grid-connected voltage source converters (VSC), weight reduction of the LCL filter is one of the main concerns, where a LCL filter can weigh hundreds of kilograms. It is a multidomain optimization issue to find the minimum weight of a LCL filter considering pulse-width modulation (PWM) method, control bandwidth, filter design, and magnetic components with given constraints, such as harmonic standards, semiconductor power losses, and cooling conditions. The comprehensive design procedure and analytical limits of a LCL filter for a 690-V/1.2-MW VSC is provided. In addition, a prototype system of 400 V/40 kW has been built to verify the design methodology and to investigate the impact of different PWM methods on the LCL filter.

The Algorithm Research of Complex Multi-Domain Constraint Models

This paper is mainly discussed in areas under a unified multi-domain framework, the unified constraint expression of optimization simulation model and critical theory in the process of solving. Establish the multi-domain optimization simulation model of complex product, study the complex coupling between multi-domain constraint models, analyze the parameters sensitivity of unified constraint expression model and reduce the restraint scale. Using the principle of global satisfaction as far as possible, to achieve coordination of multi-domain constraints based on the component level, solving the Large-scale multi-domain unified optimization simulation model with the homotopy iteration method.

Crashworthiness Optimization for a Two-Layered Front Rail Considering Front Oblique Impact

This paper is concerned with the crashworthiness design of the front rail on a vehicle chassis frame structure considering uncertain crash directions. The front rail, as a typical thin-walled structure, is sensitive to the loading direction for crashworthiness requirements. Multi-step multi-domain optimization method was applied to obtain a new single-layered front rail (SFR) which has better adaptability to crash loading directions. To further improve the front rail crashworthiness, a two-layered front rail (TFR) is proposed. Response surface method is conducted to pursue the optimal thickness pair of the two-layered front rail. Numerical simulations are carried out with Altair/Hypermesh and LS-DYNA to compare the crashworthiness performances of the standard rectangle front rail, SFR and TFR. The results show that TFR keeps the same adaptability of loading directions as SFR meanwhile it can absorb more crash energy during front angle impacts.

Research on Multi-Domain Unified Modeling and Optimization of a Micro Mirror

Referring to the natures of nonlinear dynamics and coupled energy domains presented by Micro-Opto-Electro-Mechanical Systems (MOEMS), a multi energy domain unified simulation language (modelica) based coupled modeling, simulation and optimization methodology is put forward. Then by choosing a micro mirror as the example, corresponding multi energy domain unified modeling and simulation processes of such a coupled micro device show the efficiency and convenience of this advanced modeling method. The dynamic input/output curves comparison also shows that the simulation result is accurate enough for system level design. At last, by using genetic optimization algorithm, the characteristics input/output curves of the micro mirror are converged to meet the ideal design requirement curves after 700 generations evolution. It proves the accuracy and efficiency of this system level simulation and optimization method.

Reduced-complexity representation of the coherent point-spread function in the presence of aberrations and arbitrarily large defocus

We introduce a method to analyze the diffraction integral for evaluating the point-spread function. Our method is based on the use of higher-order Airy functions along with Zernike and Taylor expansions. Our approach is applicable when we are considering a finite, arbitrary number of aberrations and arbitrarily large defocus simultaneously. We present an upper bound for the complexity and the convergence rate of this method. We also compare the cost and accuracy of this method with those of traditional ones and show the efficiency of our method through these comparisons. In particular, we rigorously show that this method is constructed in a way that the complexity of the analysis (i.e., the number of terms needed for expressing the light disturbance) does not increase as either defocus or resolution of interest increases. This has applications in several fields such as biological microscopy, lithography, and multidomain optimization in optical systems.