As complementary metal-oxide semiconductor (CMOS) technology scaling reaches its limits, new compute-in-memory technologies such as “all-spin logic” (ASL) are being explored. Preliminary predictions ...indicate that ASL implemented with perpendicular magnetic anisotropy will exhibit power-delay product (PDP) and energy-delay product (EDP) compared to CMOS, supporting its candidacy as a replacement for CMOS. In recent evaluations of ASL, unrealistic parameters have been used, leading to overly-optimistic efficiency figures. This paper uses micromagnetic simulations with realistic parameters to analyze the relationships between the various device parameters and circuit parameters, and the resulting impact on PDP and EDP. This analysis indicates that the PDP and EDP of ASL is greatly inferior to CMOS with the technological parameters that are currently available. In order to overcome these challenges relating to energy efficiency, this paper also evaluates the potential to modify the device parameters to improve the energy efficiency.
The water management in a lunar base includes interaction between the In-Situ Resource Utilization (ISRU), crew habitat, and wastewater subsystems. The ISRU produces water from the lunar regolith, ...while the wastewater subsystem provides fresh water after filtering wastewater generated from the crew habitat. A model for lunar base water filtration and management for autonomous control is yet to be developed. This paper describes a model to manage the water considering the interaction between different water and wastewater tanks in the lunar base. A methodology to generate crew members' water consumption and wastewater generation profiles is also discussed. The model considers the power demand of the wastewater subsystem. It is observed that the power demand of the wastewater subsystem depends on the power availability and the desired level of different water and wastewater tanks. It is concluded that to design an autonomous energy management system, the controller should be able to generate time-varying references for different water tanks and take into account the available power.
The importance of renewable energy sources has fostered the development of new multilevel inverter (MLI) topologies to increase the reliability and robustness of inverter systems. Switched Capacitor ...MLI (SCMLI) is an alternative inverter with reduced power supplies compared to conventional inverter. It has self-voltage balancing and boosting capability. In this paper, a new three phase SCMLI topology and its closed loop control using Model Predictive Control (MPC) for output load current is presented. The proposed MLI topology is modular in nature and number of source module can be increased for more levels of output voltage. The implementation ofMPC control ofthe proposed MLI has been discussed and it was observed that the output load current follows the reference input. Further, a comparison study is presented between the open loop Phase Disposition PWM (PDPWM) control and Model predictive control. A significant reduction of 92% in voltage THD and 99% reduction in current ripple was observed in the MPC control.
In this paper, a novel structure of cascaded multilevel inverter (MLI) is proposed. Each module of the proposed MLI consists of three numbers of DC sources and eight power switching devices. Each ...module possesses the capability of generating bipolar voltage level at its output terminals. Hence, the topology is analyzed for symmetrical and asymmetrical DC source configurations. Two types of asymmetrical DC source configuration algorithms have been proposed. As compared with existing MLI topologies, the proposed MLI requires lower components such as switching devices, driver circuits to generate specific number of output voltage level. This further reduces the size and cost of the inverter system. Extensive simulation studies of 7 level symmetrical and 11 level asymmetrical structures of proposed MLI have been done in Simulink /MATLAB. The different results of simulation studies prove the effectiveness of the proposed topology.
Spintronic devices based on domain wall (DW) motion through ferromagnetic nanowire tracks have received great interest as components of neuromorphic information processing systems. Previous proposals ...for spintronic artificial neurons required external stimuli to perform the leaking functionality, one of the three fundamental functions of a leaky integrate-and-fire (LIF) neuron. The use of this external magnetic field or electrical current stimulus results in either a decrease in energy efficiency or an increase in fabrication complexity. In this work, we modify the shape of previously demonstrated three-terminal magnetic tunnel junction neurons to perform the leaking operation without any external stimuli. The trapezoidal structure causes shape-based DW drift, thus intrinsically providing the leaking functionality with no hardware cost. This LIF neuron therefore promises to advance the development of spintronic neural network crossbar arrays.
Microgrids (MGs) are building blocks of smart power systems formed by integrating local power generation resources, energy storage systems (ESSs), and power-consuming units. While MGs offer many ...benefits, including increased resilience and flexibility, there remains a need for improved control and protection techniques that can ensure their performance and automatic restoration in response to dynamic operating conditions and failure events. Recently, researchers have explored model-free emotional learning adaptive strategies based on the emotional response of human brains to control MGs. These model-free control strategies are well-suited for handling the complexity, nonlinearity, and uncertainty present in MGs and offer several advantages over traditional approaches. This article provides an overview of different emotional learning techniques applied to MG control and protection, their challenges, and future trends. In addition, we draw parallels between the hierarchical control architecture (HCA) of MGs and the emotional learning process in the human brain, discussing their operational strategies and key areas of research. Finally, the future implementations of brain emotional learning (BEL) in the control and protection of MGs are discussed, and concluding remarks on the potential of this approach are provided.
