Pub Date : 2025-10-07DOI: 10.1109/JMMCT.2025.3618590
Mohamed Kheir;Kun Qian;Mubashra Nabi;Thomas Ebel
A novel meshless electromagnetic (EM) simulation framework based on Physics-Informed Neural Networks (PINNs), enhanced by the integration of Kolmogorov–Arnold Networks (KANs) is presented. The proposed method addresses both electrostatic and electrodynamic problems governed by Laplace and Helmholtz equations, respectively. A modular and interpretable simulation architecture is developed using KAN-PINNs which enables accurate field learning in multilayered printed circuit boards (PCBs) with spatially varying permittivity. Three canonical structures: an electrostatic box, a parallel-plate transmission line (TL) and a stripline TL are modeled to demonstrate the validity of the method. The results are validated against commercial full-wave solvers showing excellent agreement with normalized root-mean-square errors (RMSE) below 0.1. Moreover, a 4-layer PCB structure is simulated at 100 MHz to demonstrate the method's capability at higher frequencies. For this case, the model achieves a normalized RMSE of 0.153 while reducing simulation time by a factor of three compared to numerical solvers. The proposed framework provides a scalable and fully mesh-free alternative to traditional electromagnetic solvers. This introduces new potential for efficiently simulating complex PCB structures in EMI applications.
{"title":"Modular Meshless Electromagnetic Simulation Using KAN-Based Physics-Informed Neural Networks","authors":"Mohamed Kheir;Kun Qian;Mubashra Nabi;Thomas Ebel","doi":"10.1109/JMMCT.2025.3618590","DOIUrl":"https://doi.org/10.1109/JMMCT.2025.3618590","url":null,"abstract":"A novel meshless electromagnetic (EM) simulation framework based on Physics-Informed Neural Networks (PINNs), enhanced by the integration of Kolmogorov–Arnold Networks (KANs) is presented. The proposed method addresses both electrostatic and electrodynamic problems governed by Laplace and Helmholtz equations, respectively. A modular and interpretable simulation architecture is developed using KAN-PINNs which enables accurate field learning in multilayered printed circuit boards (PCBs) with spatially varying permittivity. Three canonical structures: an electrostatic box, a parallel-plate transmission line (TL) and a stripline TL are modeled to demonstrate the validity of the method. The results are validated against commercial full-wave solvers showing excellent agreement with normalized root-mean-square errors (RMSE) below 0.1. Moreover, a 4-layer PCB structure is simulated at 100 MHz to demonstrate the method's capability at higher frequencies. For this case, the model achieves a normalized RMSE of 0.153 while reducing simulation time by a factor of three compared to numerical solvers. The proposed framework provides a scalable and fully mesh-free alternative to traditional electromagnetic solvers. This introduces new potential for efficiently simulating complex PCB structures in EMI applications.","PeriodicalId":52176,"journal":{"name":"IEEE Journal on Multiscale and Multiphysics Computational Techniques","volume":"10 ","pages":"452-458"},"PeriodicalIF":1.5,"publicationDate":"2025-10-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145352081","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-09-24DOI: 10.1109/JMMCT.2025.3613962
Ranjith Kumar R;Parthasarathy Ramanujam
Reinforcement learning (RL) provides a data-driven approach for optimizing design parameters in X and Ku band applications, including radar systems and electromagnetic shielding. Conventional frequency selective surface (FSS) design methods face several challenges such as inefficient optimization and time-consuming iterative processes. To overcome these challenges, an RL-based design optimization is proposed for FSS structures to achieve enhanced selectivity, wider bandwidth, and reduced fabrication complexity. The soft actor-critic (SAC) algorithm, an advanced RL approach, is implemented in the proposed FSS design. The structure features a single-layered square grid pattern with interconnected loops to achieve wide stopband characteristics. The FSS has a guided wavelength of 0.66$lambda _{g}$ × 0.66$lambda _{g}$ × 0.1$lambda _{g}$. The design exhibits a stopband bandwidth of 10.42 GHz, ranging from 7.88 GHz to 18.30 GHz, with a shielding effectiveness (SE) of 68dB and angular stability up to 60$^{circ }$. The equivalent circuit model (ECM) analysis, simulated response, and measured results demonstrate that the SAC-based RL approach outperforms benchmark schemes, achieving the lowest mean squared error (MSE) of 0.2341.
{"title":"Design and Optimization of Wide-Band FSS Using Reinforcement Learning for X and Ku Band Radar Shielding Applications","authors":"Ranjith Kumar R;Parthasarathy Ramanujam","doi":"10.1109/JMMCT.2025.3613962","DOIUrl":"https://doi.org/10.1109/JMMCT.2025.3613962","url":null,"abstract":"Reinforcement learning (RL) provides a data-driven approach for optimizing design parameters in X and Ku band applications, including radar systems and electromagnetic shielding. Conventional frequency selective surface (FSS) design methods face several challenges such as inefficient optimization and time-consuming iterative processes. To overcome these challenges, an RL-based design optimization is proposed for FSS structures to achieve enhanced selectivity, wider bandwidth, and reduced fabrication complexity. The soft actor-critic (SAC) algorithm, an advanced RL approach, is implemented in the proposed FSS design. The structure features a single-layered square grid pattern with interconnected loops to achieve wide stopband characteristics. The FSS has a guided wavelength of 0.66<inline-formula><tex-math>$lambda _{g}$</tex-math></inline-formula> × 0.66<inline-formula><tex-math>$lambda _{g}$</tex-math></inline-formula> × 0.1<inline-formula><tex-math>$lambda _{g}$</tex-math></inline-formula>. The design exhibits a stopband bandwidth of 10.42 GHz, ranging from 7.88 GHz to 18.30 GHz, with a shielding effectiveness (SE) of 68dB and angular stability up to 60<inline-formula><tex-math>$^{circ }$</tex-math></inline-formula>. The equivalent circuit model (ECM) analysis, simulated response, and measured results demonstrate that the SAC-based RL approach outperforms benchmark schemes, achieving the lowest mean squared error (MSE) of 0.2341.","PeriodicalId":52176,"journal":{"name":"IEEE Journal on Multiscale and Multiphysics Computational Techniques","volume":"10 ","pages":"473-482"},"PeriodicalIF":1.5,"publicationDate":"2025-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145510230","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-09-23DOI: 10.1109/JMMCT.2025.3613377
Vishakha Harlapur;Salil Kulkarni
Thomson coil is a fast-acting electromagnetic actuator, making it a preferred choice for hybrid circuit breakers. Modeling of Thomson coil actuator involves multi-physics analysis consisting of electric circuit, electromagnetism and structural mechanics and is typically carried out using Finite Element Analysis. These simulations are time consuming and therefore not practical during the preliminary design stage or during optimization studies. To address this challenge, a data-driven modeling approach is presented in this paper. First, a set of dimensionless parameters are identified using the Buckingham Pi theorem and then a regression model is developed to establish an approximate relationship between different design variables and the quantity of interest. The use of dimensionless variables not only reduces the number of fitting parameters but also helps to achieve size based scaling. Using the developed model, optimized values of design variables are obtained. This is followed by a sensitivity analysis to evaluate the effect of design parameters on the quantity of interest.
{"title":"Data Driven Modeling and Design Optimization of Thomson Coil Using Dimensionless Parameters","authors":"Vishakha Harlapur;Salil Kulkarni","doi":"10.1109/JMMCT.2025.3613377","DOIUrl":"https://doi.org/10.1109/JMMCT.2025.3613377","url":null,"abstract":"Thomson coil is a fast-acting electromagnetic actuator, making it a preferred choice for hybrid circuit breakers. Modeling of Thomson coil actuator involves multi-physics analysis consisting of electric circuit, electromagnetism and structural mechanics and is typically carried out using Finite Element Analysis. These simulations are time consuming and therefore not practical during the preliminary design stage or during optimization studies. To address this challenge, a data-driven modeling approach is presented in this paper. First, a set of dimensionless parameters are identified using the Buckingham Pi theorem and then a regression model is developed to establish an approximate relationship between different design variables and the quantity of interest. The use of dimensionless variables not only reduces the number of fitting parameters but also helps to achieve size based scaling. Using the developed model, optimized values of design variables are obtained. This is followed by a sensitivity analysis to evaluate the effect of design parameters on the quantity of interest.","PeriodicalId":52176,"journal":{"name":"IEEE Journal on Multiscale and Multiphysics Computational Techniques","volume":"10 ","pages":"443-451"},"PeriodicalIF":1.5,"publicationDate":"2025-09-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145315384","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-09-22DOI: 10.1109/JMMCT.2025.3608780
Dan Jiao
{"title":"Editorial: Introducing Explaining the Unexplained","authors":"Dan Jiao","doi":"10.1109/JMMCT.2025.3608780","DOIUrl":"https://doi.org/10.1109/JMMCT.2025.3608780","url":null,"abstract":"","PeriodicalId":52176,"journal":{"name":"IEEE Journal on Multiscale and Multiphysics Computational Techniques","volume":"10 ","pages":"433-433"},"PeriodicalIF":1.5,"publicationDate":"2025-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11175564","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145141694","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-09-09DOI: 10.1109/JMMCT.2025.3608139
Mingyu Lu;Charan Litchfield
A 5-MHz sinusoidal signal, a 500-Hz sinusoidal signal, and a DC signal are compared among each other experimentally when they are turned on, after they are established over a piece of long co-axial cable, and when they are turned off. The experimental results do not demonstrate any fundamental differences among the 5-MHz signal, 500-Hz signal, and DC signal in terms of propagation over the co-axial cable. Based on the experimental results, the well-known formulations of AC wave propagation are extended to DC wave propagation. The experimental and theoretical studies of this paper indicate that DC electrical power is transported by electromagnetic wave propagation in practical DC circuits.
{"title":"Is DC Power Transmitted by Electromagnetic Waves?","authors":"Mingyu Lu;Charan Litchfield","doi":"10.1109/JMMCT.2025.3608139","DOIUrl":"https://doi.org/10.1109/JMMCT.2025.3608139","url":null,"abstract":"A 5-MHz sinusoidal signal, a 500-Hz sinusoidal signal, and a DC signal are compared among each other experimentally when they are turned on, after they are established over a piece of long co-axial cable, and when they are turned off. The experimental results do not demonstrate any fundamental differences among the 5-MHz signal, 500-Hz signal, and DC signal in terms of propagation over the co-axial cable. Based on the experimental results, the well-known formulations of AC wave propagation are extended to DC wave propagation. The experimental and theoretical studies of this paper indicate that DC electrical power is transported by electromagnetic wave propagation in practical DC circuits.","PeriodicalId":52176,"journal":{"name":"IEEE Journal on Multiscale and Multiphysics Computational Techniques","volume":"10 ","pages":"434-442"},"PeriodicalIF":1.5,"publicationDate":"2025-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145141781","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Surge over-voltages may induce magnetic saturation, flux instability in power components and undermining reliability. To address trade-off between computational efficiency and accuracy of the fixed-step finite element method (FEM) under transients, this paper presents an adaptive time-stepping FEM (ATS-FEM) driven by higher-order truncation-error estimation, with Schur complement preconditioning integrated to optimize memory usage for accelerating parallel matrix solution. Three typical magnetic components often used in strong magnetic launch and propulsion systems are simulated and validated in comparison with that of commercial software. It is shown that our developed ATS-FEM can dynamically adjust the time steps but with high numerical accuracy maintained, and it also has the capability for capturing localized saturation, radial gradients, and permeability drops in high-current regions of the magnetic components.
{"title":"An Adaptive Time-Stepping Finite Element Method With Schur-Complement Preconditioning for Surge Simulation of Magnetic Components","authors":"Zhe Chen;Yanning Chen;Yi-Yao Wang;Hao-Xuan Zhang;Yin-Da Wang;Rongchuan Bai;Zhengwei Du;Yingzong Liang;Fang Liu;Hao Xie;Wen-Yan Yin","doi":"10.1109/JMMCT.2025.3606993","DOIUrl":"https://doi.org/10.1109/JMMCT.2025.3606993","url":null,"abstract":"Surge over-voltages may induce magnetic saturation, flux instability in power components and undermining reliability. To address trade-off between computational efficiency and accuracy of the fixed-step finite element method (FEM) under transients, this paper presents an adaptive time-stepping FEM (ATS-FEM) driven by higher-order truncation-error estimation, with Schur complement preconditioning integrated to optimize memory usage for accelerating parallel matrix solution. Three typical magnetic components often used in strong magnetic launch and propulsion systems are simulated and validated in comparison with that of commercial software. It is shown that our developed ATS-FEM can dynamically adjust the time steps but with high numerical accuracy maintained, and it also has the capability for capturing localized saturation, radial gradients, and permeability drops in high-current regions of the magnetic components.","PeriodicalId":52176,"journal":{"name":"IEEE Journal on Multiscale and Multiphysics Computational Techniques","volume":"10 ","pages":"421-432"},"PeriodicalIF":1.5,"publicationDate":"2025-09-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145090144","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-08-28DOI: 10.1109/JMMCT.2025.3603902
Kiran Ravindran;Abhijith B. Narendranath;Kalarickaparambil Joseph Vinoy
Numerical electromagnetic computations must often accommodate random geometric representations while handling biological tissues, and engineered components with manufacturing tolerances. Meshless time-domain radial point interpolation method (RPIM) offers advantages to quantitatively analyze such geometric uncertainties using polynomial chaos expansion (PCE). Formulations for geometric uncertainties may require variations in mesh or node distribution for each analyzed sample, leading to high computational requirement for re-meshing. The proposed geometric stochastic RPIM (G-SRPIM) overcomes this with a single domain model by expressing the shape function matrix of RPIM in a stochastic framework. The uncertainty is quantified in G-SRPIM through a novel way by which its random support domain moment matrices are organized in a block structure, and inverted using Schur's complement and Neumann approximation, exploiting the underlying symmetry. The proposed method is validated by analyzing a parallel plate waveguide with a slit exhibiting random variations, a realistic 3D bio-electromagnetic problem involving a section of human head, and an iris filter with random variations in its iris dimensions. Standard deviation upto $45 %$ of the average inter-node distance is tested without jeopardizing the stability. The accuracy of our approach is compared with Monte-Carlo (MC) simulations on a deterministic RPIM using the Kolmogorov-Smirnov (KS) test. Additionally, results are compared with MC simulation on CST Studio Suite 2018 and stochastic collocation (SC). The proposed method exhibits superior execution time compared to SC and MC-based non-intrusive implementations, underscoring its efficiency and reliability in handling geometric uncertainties in microwave components.
在处理生物组织和具有制造公差的工程部件时，数值电磁计算必须经常适应随机几何表示。无网格时域径向点插值法（RPIM）具有利用多项式混沌展开（PCE）定量分析几何不确定性的优势。几何不确定性的公式可能需要每个分析样本的网格或节点分布的变化，导致重新网格划分的高计算需求。提出的几何随机RPIM （G-SRPIM）通过在随机框架中表示RPIM的形状函数矩阵，克服了这一问题。在G-SRPIM中，不确定性是通过一种新颖的方法来量化的，通过这种方法，它的随机支持域矩矩阵被组织成一个块结构，并使用舒尔补和诺依曼近似来反演，利用潜在的对称性。通过分析具有随机变化的狭缝平行板波导、涉及人体头部部分的现实三维生物电磁问题以及虹膜尺寸随机变化的虹膜滤波器，验证了所提方法的有效性。在不影响稳定性的情况下，测试平均节点间距离的标准偏差可达45%。我们的方法的准确性与蒙特卡罗（MC）模拟的确定性RPIM使用Kolmogorov-Smirnov （KS）测试进行了比较。此外，还将结果与CST Studio Suite 2018上的MC模拟和随机配置（SC）进行了比较。与基于SC和mc的非侵入式实现相比，该方法具有更好的执行时间，强调了其在处理微波元件几何不确定性方面的效率和可靠性。
{"title":"A Meshless Time-Domain Method for Geometric Uncertainty Quantification","authors":"Kiran Ravindran;Abhijith B. Narendranath;Kalarickaparambil Joseph Vinoy","doi":"10.1109/JMMCT.2025.3603902","DOIUrl":"https://doi.org/10.1109/JMMCT.2025.3603902","url":null,"abstract":"Numerical electromagnetic computations must often accommodate random geometric representations while handling biological tissues, and engineered components with manufacturing tolerances. Meshless time-domain radial point interpolation method (RPIM) offers advantages to quantitatively analyze such geometric uncertainties using polynomial chaos expansion (PCE). Formulations for geometric uncertainties may require variations in mesh or node distribution for each analyzed sample, leading to high computational requirement for re-meshing. The proposed geometric stochastic RPIM (G-SRPIM) overcomes this with a single domain model by expressing the shape function matrix of RPIM in a stochastic framework. The uncertainty is quantified in G-SRPIM through a novel way by which its random support domain moment matrices are organized in a block structure, and inverted using Schur's complement and Neumann approximation, exploiting the underlying symmetry. The proposed method is validated by analyzing a parallel plate waveguide with a slit exhibiting random variations, a realistic 3D bio-electromagnetic problem involving a section of human head, and an iris filter with random variations in its iris dimensions. Standard deviation upto <inline-formula><tex-math>$45 %$</tex-math></inline-formula> of the average inter-node distance is tested without jeopardizing the stability. The accuracy of our approach is compared with Monte-Carlo (MC) simulations on a deterministic RPIM using the Kolmogorov-Smirnov (KS) test. Additionally, results are compared with MC simulation on CST Studio Suite 2018 and stochastic collocation (SC). The proposed method exhibits superior execution time compared to SC and MC-based non-intrusive implementations, underscoring its efficiency and reliability in handling geometric uncertainties in microwave components.","PeriodicalId":52176,"journal":{"name":"IEEE Journal on Multiscale and Multiphysics Computational Techniques","volume":"10 ","pages":"396-406"},"PeriodicalIF":1.5,"publicationDate":"2025-08-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145036818","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-08-26DOI: 10.1109/JMMCT.2025.3602986
Ting Zang;Gaobiao Xiao
This paper presents an efficient optimization algorithm for synthesizing the discrete array factor, which extends the optimization domain to the invisible region to mitigate aliasing effect, thereby achieving well-controlled radiation patterns. By further lowering the level of the sidelobes in part of the visible region, the algorithm allows to shape the radiation patterns of sparse arrays with desired characteristics, such as uniform main lobe ripples and low sidelobe levels. Some evanescent modes have been added to compensate for the additional degrees of freedom caused by the increased optimization range, so that the number of the extreme points to be controlled is still approximately equal to the number of degrees of freedom (NDF), maintaining the monotonic convergence property of the algorithm. Numerical examples and FEKO simulation results validate the effectiveness and the accuracy of the proposed method.
{"title":"An Efficient Method for Synthesizing Sparse Arrays With Well-Controlled Discrete Array Factors","authors":"Ting Zang;Gaobiao Xiao","doi":"10.1109/JMMCT.2025.3602986","DOIUrl":"https://doi.org/10.1109/JMMCT.2025.3602986","url":null,"abstract":"This paper presents an efficient optimization algorithm for synthesizing the discrete array factor, which extends the optimization domain to the invisible region to mitigate aliasing effect, thereby achieving well-controlled radiation patterns. By further lowering the level of the sidelobes in part of the visible region, the algorithm allows to shape the radiation patterns of sparse arrays with desired characteristics, such as uniform main lobe ripples and low sidelobe levels. Some evanescent modes have been added to compensate for the additional degrees of freedom caused by the increased optimization range, so that the number of the extreme points to be controlled is still approximately equal to the number of degrees of freedom (NDF), maintaining the monotonic convergence property of the algorithm. Numerical examples and FEKO simulation results validate the effectiveness and the accuracy of the proposed method.","PeriodicalId":52176,"journal":{"name":"IEEE Journal on Multiscale and Multiphysics Computational Techniques","volume":"10 ","pages":"388-395"},"PeriodicalIF":1.5,"publicationDate":"2025-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144998059","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
An efficient hybrid approach based on combining the bidirectional recurrent neural network with knowledge-based neural network is presented to predict jitter in a chain of CMOS inverters in the presence of multiple noise sources. The new method achieves a reasonable accuracy and provides for efficient training using input data obtained from both a circuit simulator as well as analytical relations. The proposed approach can also estimate jitter for each inverter in the chain by only employing the accurate training data associated with the first inverter, resulting in a significant increase in speed compared to conventional approaches.
{"title":"Knowledge-Based Bidirectional Recurrent Neural Network Approach for Efficient Prediction of Jitter in a Chain of CMOS Inverters","authors":"Ahsan Javaid;Ramachandra Achar;Jai Narayan Tripathi","doi":"10.1109/JMMCT.2025.3602632","DOIUrl":"https://doi.org/10.1109/JMMCT.2025.3602632","url":null,"abstract":"An efficient hybrid approach based on combining the bidirectional recurrent neural network with knowledge-based neural network is presented to predict jitter in a chain of CMOS inverters in the presence of multiple noise sources. The new method achieves a reasonable accuracy and provides for efficient training using input data obtained from both a circuit simulator as well as analytical relations. The proposed approach can also estimate jitter for each inverter in the chain by only employing the accurate training data associated with the first inverter, resulting in a significant increase in speed compared to conventional approaches.","PeriodicalId":52176,"journal":{"name":"IEEE Journal on Multiscale and Multiphysics Computational Techniques","volume":"10 ","pages":"407-420"},"PeriodicalIF":1.5,"publicationDate":"2025-08-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145036817","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-08-01DOI: 10.1109/JMMCT.2025.3593872
Suyash Kushwaha;Chintu Bhaskara Rao;Shamini P R;Sourajeet Roy;Rohit Sharma
In this paper, novel copper graphene heterogeneous interconnect structures are proposed which retain the ease of fabrication while having far better electrical performance when compared to the conventional copper interconnects. In the nanoscale regime, signal integrity (SI) of the copper interconnects degrades significantly. To address the signal integrity issues, these heterogeneous interconnects are developed at 7 nm technology nodes which are further used to make the crossbar arrays for neuromorphic computing. The proposed copper graphene heterogeneous interconnects were designed by stacking the layers of copper and multilayer graphene nanoribbons (MLGNRs) one over the other and a detailed signal integrity analysis is done based on the quantities like the per unit length Resistance, Insertion Loss (IL), Return Loss (RL), eye diagrams, surface charge density and volume current density. The results shows that the proposed interconnects outperformed the copper interconnects based on each and every SI quantity. Finally, in the application example, the best performing heterogeneous interconnects are used to create larger crossbar arrays with sizes 64 × 64, 128 × 128. Further, the key performance matrices such as the delay time, the rise time and the fall time are analyzed and compared with the conventional crossbars made from the copper interconnects. The results in application example proved that the heterogeneous interconnects performs better than the copper interconnects for neuromorphic computing.
{"title":"Performance Enhanced Copper-Graphene Hetero Interconnect Structures in Crossbar Arrays for Neuromorphic Computing","authors":"Suyash Kushwaha;Chintu Bhaskara Rao;Shamini P R;Sourajeet Roy;Rohit Sharma","doi":"10.1109/JMMCT.2025.3593872","DOIUrl":"https://doi.org/10.1109/JMMCT.2025.3593872","url":null,"abstract":"In this paper, novel copper graphene heterogeneous interconnect structures are proposed which retain the ease of fabrication while having far better electrical performance when compared to the conventional copper interconnects. In the nanoscale regime, signal integrity (SI) of the copper interconnects degrades significantly. To address the signal integrity issues, these heterogeneous interconnects are developed at 7 nm technology nodes which are further used to make the crossbar arrays for neuromorphic computing. The proposed copper graphene heterogeneous interconnects were designed by stacking the layers of copper and multilayer graphene nanoribbons (MLGNRs) one over the other and a detailed signal integrity analysis is done based on the quantities like the per unit length Resistance, Insertion Loss (IL), Return Loss (RL), eye diagrams, surface charge density and volume current density. The results shows that the proposed interconnects outperformed the copper interconnects based on each and every SI quantity. Finally, in the application example, the best performing heterogeneous interconnects are used to create larger crossbar arrays with sizes 64 × 64, 128 × 128. Further, the key performance matrices such as the delay time, the rise time and the fall time are analyzed and compared with the conventional crossbars made from the copper interconnects. The results in application example proved that the heterogeneous interconnects performs better than the copper interconnects for neuromorphic computing.","PeriodicalId":52176,"journal":{"name":"IEEE Journal on Multiscale and Multiphysics Computational Techniques","volume":"10 ","pages":"379-387"},"PeriodicalIF":1.5,"publicationDate":"2025-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144904842","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：