1,720,999 research outputs found
Design of an Axial Flux Magnetic-geared Motor for Standing-up Assistance
No full textThe design procedure for a magnetic geared motor with axial flux to assist the sit-to-stand motion of elderly adults is presented in this paper. The machine uses two rotors rotating in opposite directions, and each consists on axially magnetized permanent magnets and are magnetically coupled through an iron yoke. The magnetic equivalent circuit modelling is used to calculate both the output torque of the motor set and the pull-put torque of the gear set. The dimensions for an initial design was determined by finding the intersection of these torques within the size constraint of the application. A 3-dimensional finite element analysis was used to validate this initial design and machines with similar dimensions were compared according to their pull-out torque and maximum load torque supported. The final design was tested under the simulated load conditions for a standing-up motion and proved capable of aiding a person in standing up
A Dynamically Smoothing State of Charge Estimation on Electric Vehicle Batteries
No full textWith the successful development of electric vehicle industries, the advanced utilization and the management of onboard batteries has become indispensable. As a result, the requirements on Battery Management Systems (BMS) become more and more stringent. To achieve demanding goals for energy efficiency and battery life, the BMS needs to function effectively with accurate real-time system information. Battery state-of-charge (SOC) is one of the most important system parameters, and therefore SOC online estimation is of parameters paramount importance to operations of BMS and the entire EV. In this thesis, we propose a battery SOC estimation method for EV operations. The proposed method uses neural network to obtain a preliminary estimation. After that, a corrected version of SOC is obtained by passing the preliminary estimate through a smoothing FIR filter. The neural model used was long short-term memory (LSTM) model as a member of recurrent neural network family. On the other hand, the proposed filter is a self-defining FIR filter, whose coefficients are trained using genetic algorithm. By doing this, we have the ability to achieve a SOC mean error exist within 1% and a SOC max error up to 5.17% before passing through the filter. And max error decrease to 2.44% after passing through the filter. The dynamic test data used in this paper was according to the light vehicles testing standards releasing by DiselNet
Implementation of Software Update Rollout Problem on Unbalanced Three Phase Distribution Networks
No full textWith the increasing penetration of smart devices in power grids, concerns regarding the potential impact of control software on the system buses arise.
Additionally, the communication channel between Internet of Things (IoT) devices such as smart meters and power system operators via the internet becomes susceptible to cyber-attacks.
The increasing need for software updates in power grid systems necessitates addressing related issues and mathematical models. However, existing descriptions of software update problems mainly focus on single-phase systems, while Taiwan's distribution systems are predominantly unbalanced three-phase. This thesis aims to extend the description of the software update problem to three-phase systems to address this gap.
Specifically, this thesis suggests using OpenDSS to approximate the original three-phase distribution system as a single-phase positive sequence system. The impact of the update schedule on the actual three-phase unbalanced system is then simulated. Unexpected power injections lead to further voltage imbalance in the system.
By evaluating grid code indicators, it is confirmed that the actual unbalanced three phase system remains within the specified limits despite the impact of the update schedule. This validates the proposed approach's suitability for addressing the software update problem in three-phase systems and deriving the appropriate software update schedule
Theoretical Analysis of Space Harmonic Effects of Synchronous-Reluctance Motor
No full textThe purpose of this research is to study the effects of torque and torque ripple of synchronous reluctance motor (SynRM) with the different winding arrangements and the rotor shapes, including simulation of synchronous reluctance motor with the mathematical torque equation. Then reduce the torque ripple by adjusting air-gap\ue2s harmonic proportion from move barriers. The finite element analysis is also provided to simulate to verify the derivation adequacy
Low computational method online estimation of lithium-ion battery state of health
No full textWith the increasing focus on environmental issues, the demand for electric vehicles (EVs) and energy storage systems (ESS) is gradually rising. This implies an increased reliance on batteries. Over prolonged usage, batteries experience a reduction in their actual capacity. In terms of usage, State of Health (SOH) is crucial for EVs and ESS. It indicates the remaining lifespan of the battery and helps identify potential internal failures in aging batteries.
This study proposes an online estimation model for SOH, which differs from traditional estimation methods that require complex calculations or waiting for complete charging cycles.It primarily utilizes offline aging experimental data to acquire voltage and current values during battery charging periods. It applies a first-order differential to the relationship between battery capacity and voltage, known as Incremental Capacity Analysis (ICA), and extracts appropriate feature parameters to describe the degree of aging, such as peak height and potential position of the ICA curve. Then, a piecewise linear interpolation method is used to establish the relationship between aging feature parameters and SOH. For SOH estimated, partial charging measurements of voltage and current are collected and subjected to ICA to obtain aging feature parameters for online SOH estimation .
This study utilizes both piecewise linear interpolation and neural networks to estimate SOH. The estimation results are compared using three error metrics: average error, root mean square error, and maximum error. The computational complexity between online estimations of the two methods is also compared, along with the sensitivity of input parameters. Among the computationally efficient piecewise linear interpolation methods, it achieved the best maximum error of 4.862% and demonstrated good robustness against mode
A Quasi-Convex Optimization Approach to Parameterized Model Order Reduction
No full textIn this paper an optimization based model order reduction (MOR) framework is proposed. The method involves setting up a quasiconvex program that explicitly minimizes a relaxation of the optimal H ∞ norm MOR problem. The method generates guaranteed stable and passive reduced models and it is very flexible in imposing additional constraints. The proposed optimization approach is also extended to parameterized model reduction problem (PMOR). The proposed method is compared to existing moment matching and optimization based MOR methods in several examples. A PMOR model for a large RF inductor is also constructed
Relaxed connected dominating set problem with application to secure power network design
No full text- …
