- 标题
- 摘要
- 关键词
- 实验方案
- 产品
-
KRIGING-PARETO FRONT APPROACH FOR THE MULTI-OBJECTIVE EXPLORATION OF METAMATERIAL TOPOLOGIES
摘要： Metamaterials provide the opportunity for designers to create customisable artificial materials by independently tailoring the electric and magnetic response of sub-wavelength geometric structures to electromagnetic energy. Due to the increased complexity of these geometric structures, exacerbated by the increased interest in generating inhomogeneous and anisotropic metamaterials, direct optimisation of these designs using conventional approaches often becomes impractical and limited. In order to alleviate this issue, we propose an alternative optimisation approach which exploits the Kriging methodology in conjunction with an adaptive sampling plan to simultaneously optimise multiple conflicting objectives. Results show the effectiveness of the outlined algorithm in calculating a uniform spread of optimal trade-off designs, balancing the real and imaginary components of the refractive index over a wide range of values.
关键词： Kriging,Metamaterials,Refractive Index,Multi-Objective Optimisation,Pareto front
更新于2025-09-09 09:28:46
-
Optimisation of phase factor in phase retrieval for reflector antennas with active surface
摘要： The deformation of reflector surface can be obtained by Misell algorithm using two far-field amplitude patterns: one with the antenna in focus and the other with the antenna defocused. For reflector antennas with active surface, the defocused pattern can be measured by adjusting the panels to get a more effective phase factor instead of offsetting the sub-reflector. The shape and size of phase factor have a great influence on the accuracy of convergent solution. To improve the performance of Misell algorithm, this study proposed an optimisation method for phase factors under different surface deformations. This method utilises Levenberg–Marquardt algorithm to get an estimation of the aperture distribution and then applies the particle swarm optimisation to optimise the phase factor parameterised by Zernike polynomials. Numerical simulations show that the proposed method is an efficient tool to achieve applicable phase factors under different surface deformations. Using the optimised phase factor, Misell algorithm can get a better convergent performance than traditional Misell algorithm.
关键词： Misell algorithm,particle swarm optimisation,Levenberg–Marquardt algorithm,active surface,phase retrieval,reflector antennas,Zernike polynomials
更新于2025-09-09 09:28:46
-
Clusters partition and zonal voltage regulation for distribution networks with high penetration of PVs
摘要： Integration of distributed generation (DG) at large scale with high penetration challenges the radial structure of the traditional distribution networks and the effectiveness of the conventional voltage regulation methods. In this study, the clusters partitioning and voltage regulation are researched. The modified electrical distance is introduced. An effective method, based on spectral clustering algorithm, is proposed for the partitioning of the DG network via the judgement of critical load buses. Two-stage voltage regulation optimisation is realised in each sub-community. The optimal objects are the minimal voltage fluctuation and the network loss of the distributed network. The independent variables are reactive-power absorption and active-power curtailment for each controllable photovoltaic node. An advanced particle swarm optimisation algorithm is applied to the voltage regulation for the sub-communities. After a case study of the IEEE 33-bus system, a regional distribution network in Anhui province of China is analysed. Simulation results indicate that the node voltages are stabilised with the improvement of power quality employing the proposed clusters partitioning method and zonal power control scheme.
关键词： spectral clustering,particle swarm optimisation,voltage regulation,distributed generation,photovoltaic
更新于2025-09-04 15:30:14