• 제목/요약/키워드: feed-forward architecture

검색결과 35건 처리시간 0.026초

Experimental calibration of forward and inverse neural networks for rotary type magnetorheological damper

  • Bhowmik, Subrata;Weber, Felix;Hogsberg, Jan
    • Structural Engineering and Mechanics
    • /
    • 제46권5호
    • /
    • pp.673-693
    • /
    • 2013
  • This paper presents a systematic design and training procedure for the feed-forward back-propagation neural network (NN) modeling of both forward and inverse behavior of a rotary magnetorheological (MR) damper based on experimental data. For the forward damper model, with damper force as output, an optimization procedure demonstrates accurate training of the NN architecture with only current and velocity as input states. For the inverse damper model, with current as output, the absolute value of velocity and force are used as input states to avoid negative current spikes when tracking a desired damper force. The forward and inverse damper models are trained and validated experimentally, combining a limited number of harmonic displacement records, and constant and half-sinusoidal current records. In general the validation shows accurate results for both forward and inverse damper models, where the observed modeling errors for the inverse model can be related to knocking effects in the measured force due to the bearing plays between hydraulic piston and MR damper rod. Finally, the validated models are used to emulate pure viscous damping. Comparison of numerical and experimental results demonstrates good agreement in the post-yield region of the MR damper, while the main error of the inverse NN occurs in the pre-yield region where the inverse NN overestimates the current to track the desired viscous force.

Differential Power Processing System for the Capacitor Voltage Balancing of Cost-effective Photovoltaic Multi-level Inverters

  • Jeon, Young-Tae;Kim, Kyoung-Tak;Park, Joung-Hu
    • Journal of Power Electronics
    • /
    • 제17권4호
    • /
    • pp.1037-1047
    • /
    • 2017
  • The Differential Power Processing (DPP) converter is a promising multi-module photovoltaic inverter architecture recently proposed for photovoltaic systems. In this paper, a DPP converter architecture, in which each PV-panel has its own DPP converter in shunt, performs distributed maximum power point tracking (DMPPT) control. It maintains a high energy conversion efficiency, even under partial shading conditions. The system architecture only deals with the power differences among the PV panels, which reduces the power capacity of the converters. Therefore, the DPP systems can easily overcome the conventional disadvantages of PCS such as centralized, string, and module integrated converter (MIC) topologies. Among the various types of the DPP systems, the feed-forward method has been selected for both its voltage balancing and power transfer to a modified H-bridge inverter that needs charge balancing of the input capacitors. The modified H-bridge multi-level inverter had some advantages such as a low part count and cost competitiveness when compared to conventional multi-level inverters. Therefore, it is frequently used in photovoltaic (PV) power conditioning system (PCS). However, its simplified switching network draws input current asymmetrically. Therefore, input capacitors in series suffer from a problem due to a charge imbalance. This paper validates the operating principle and feasibility of the proposed topology through the simulation and experimental results. They show that the input-capacitor voltages maintain the voltage balance with the PV MPPT control operating with a 140-W hardware prototype.

광대역 시스템을 위한 저전력 시그마-델타 변조기 (Design of the Low-Power Continuous-Time Sigma-Delta Modulator for Wideband Applications)

  • 김근모;박창준;이상훈;김상길;김주성
    • 전기전자학회논문지
    • /
    • 제21권4호
    • /
    • pp.331-337
    • /
    • 2017
  • 본 논문에서는 20MHz 대역폭, 저잡음, 저전력의 3차 저역 통과 시그마-델타 모듈레이터를 개발한다. 본 시스템의 대역폭은 LTE 및 그 외 다른 광대역 무선통신 표준을 만족할 수 있다. Feed-forward 구조의 3차 저역 통과 필터를 통해 저전력 및 저복잡도를 실현한다. 개발된 시스템은 빠른 데이터 변환을 실현하기 위해 3bit-flash 타입의 양자화 회로를 사용하였다. Current-steering DAC의 경우 추가적인 회로 없이 높은 정확도와 낮은 전력 소모의 이유로 고안되었다. DAC의 입력 전압이 변할 경우 생기는 glitch들을 없애기 위해 cross-coupled 트랜시스터를 사용하여 glitch 상쇄(cancellation)를 실현하였다. 개발된 시스템은 32.65mW의 저전력 구현과 함께 65.9dB의 peak SNDR, 20MHz의 대역폭을 실현한다. 600mVp-p의 입력 two-tone 신호 입력 인가후의 IM3는 69dBc를 실현하였으며 TSMC의 0.18-um CMOS 공정을 이용하여 설계되었다.

Unsupervised Real-time Obstacle Avoidance Technique based on a Hybrid Fuzzy Method for AUVs

  • Anwary, Arif Reza;Lee, Young-Il;Jung, Hee;Kim, Yong-Gi
    • International Journal of Fuzzy Logic and Intelligent Systems
    • /
    • 제8권1호
    • /
    • pp.82-86
    • /
    • 2008
  • The article presents ARTMAP and Fuzzy BK-Product approach underwater obstacle avoidance for the Autonomous underwater Vehicles (AUV). The AUV moves an unstructured area of underwater and could be met with obstacles in its way. The AUVs are equipped with complex sensorial systems like camera, aquatic sonar system, and transducers. A Neural integrated Fuzzy BK-Product controller, which integrates Fuzzy logic representation of the human thinking procedure with the learning capabilities of neural-networks (ARTMAP), is developed for obstacle avoidance in the case of unstructured areas. In this paper, ARTMAP-Fuzzy BK-Product controller architecture comprises of two distinct elements, are 1) Fuzzy Logic Membership Function and 2) Feed-Forward ART component. Feed-Forward ART component is used to understanding the unstructured underwater environment and Fuzzy BK-Product interpolates the Fuzzy rule set and after the defuzzyfication, the output is used to take the decision for safety direction to go for avoiding the obstacle collision with the AUV. An on-line reinforcement learning method is introduced which adapts the performance of the fuzzy units continuously to any changes in the environment and make decision for the optimal path from source to destination.

Developing an approach for fast estimation of range of ion in interaction with material using the Geant4 toolkit in combination with the neural network

  • Khalil Moshkbar-Bakhshayesh;Soroush Mohtashami
    • Nuclear Engineering and Technology
    • /
    • 제54권11호
    • /
    • pp.4209-4214
    • /
    • 2022
  • Precise modelling of the interaction of ions with materials is important for many applications including material characterization, ion implantation in devices, thermonuclear fusion, hadron therapy, secondary particle production (e.g. neutron), etc. In this study, a new approach using the Geant4 toolkit in combination with the Bayesian regularization (BR) learning algorithm of the feed-forward neural network (FFNN) is developed to estimate the range of ions in materials accurately and quickly. The different incident ions at different energies are interacted with the target materials. The Geant4 is utilized to model the interactions and to calculate the range of the ions. Afterward, the appropriate architecture of the FFNN-BR with the relevant input features is utilized to learn the modelled ranges and to estimate the new ranges for the new cases. The notable achievements of the proposed approach are: 1- The range of ions in different materials is given as quickly as possible and the time required for estimating the ranges can be neglected (i.e. less than 0.01 s by a typical personal computer). 2- The proposed approach can generalize its ability for estimating the new untrained cases. 3- There is no need for a pre-made lookup table for the estimation of the range values.

Identification of a suitable ANN architecture in predicting strain in tie section of concrete deep beams

  • Mohammadhassani, Mohammad;Nezamabadi-pour, Hossein;Suhatril, Meldi;Shariati, Mahdi
    • Structural Engineering and Mechanics
    • /
    • 제46권6호
    • /
    • pp.853-868
    • /
    • 2013
  • The comparison of the effectiveness of artificial neural network (ANN) and linear regression (LR) in the prediction of strain in tie section using experimental data from eight high-strength-self-compact-concrete (HSSCC) deep beams are presented here. Prior to the aforementioned, a suitable ANN architecture was identified. The format of the network architecture was ten input parameters, two hidden layers, and one output. The feed forward back propagation neural network of eleven and ten neurons in first and second TRAINLM training function was highly accurate and generated more precise tie strain diagrams compared to classical LR. The ANN's MSE values are 90 times smaller than the LR's. The correlation coefficient value from ANN is 0.9995 which is indicative of a high level of confidence.

Sliding mode control based on neural network for the vibration reduction of flexible structures

  • Huang, Yong-An;Deng, Zi-Chen;Li, Wen-Cheng
    • Structural Engineering and Mechanics
    • /
    • 제26권4호
    • /
    • pp.377-392
    • /
    • 2007
  • A discrete sliding mode control (SMC) method based on hybrid model of neural network and nominal model is proposed to reduce the vibration of flexible structures, which is a robust active controller developed by using a sliding manifold approach. Since the thick boundary layer will reduce the virtue of SMC, the multilayer feed-forward neural network is adopted to model the uncertainty part. The neural network is trained by Levenberg-Marquardt backpropagation. The design objective of the sliding mode surface is based on the quadratic optimal cost function. In course of running, the input signal of SMC come from the hybrid model of the nominal model and the neural network. The simulation shows that the proposed control scheme is very effective for large uncertainty systems.

On-Chip 학습기능을 가진 확률연산 펄스형 디지털 신경망의 구현 (Implementation of A Pulse-mode Digital Neural Network with On-chip Learning Using Stochastic Computation)

  • 위재우;이종호
    • 대한전기학회:학술대회논문집
    • /
    • 대한전기학회 1998년도 하계학술대회 논문집 G
    • /
    • pp.2296-2298
    • /
    • 1998
  • In this paper, an on-chip learning pulse-mode digital neural network with a massively parallel yet compact and flexible network architecture is suggested. Algebraic neural operations are replaced by stochastic processes using pseudo-random sequences and simple logic gates are used as basic computing elements. Using Back-propagation algorithm both feed-forward and learning phases are efficiently implemented with simple logical gates. RNG architecture using LFSR and barrel shifter are adopted to avoid some correlation between pulse trains. Suggested network is designed in digital circuit and its performance is verified by computer simulation.

  • PDF

다관절 머니퓰레이터의 사용자 중심 제어기 설계 (User-Oriented Controller Design for Multi-Axis Manipulators)

  • 손현석;강대훈;이장명
    • 대한임베디드공학회논문지
    • /
    • 제3권2호
    • /
    • pp.49-56
    • /
    • 2008
  • This paper proposes a PC-based open architecture controller for a multi-axis robotic manipulator. The designed controller can be applied for various multi-axes robotic manipulators since the motion controller is implemented on a PC with its peripheral devices. The accuracy of the controller based on the computed torque method has been measured with the dynamic model of manipulator. Since the controller is implemented in the PC-based architecture, it is free from the user circumstances and the operating environment. Dynamics of the manipulator have been compensated by the feed forward path in the inner loop and the resulting linear outer loop has been controlled by PD algorithm. Using the specialized language, it can be more efficient in programming and in driving of the multi-axis robot. Unlike the conventional controller that is used to control only a specific robot, this controller can be easily changed for various types of robots. This paper proposes a PC-based controller that has a simple architecture with its simple interface circuits than general commercial controllers. The maintenance and the performance of the controller can be easily improved for a specific robot. In fact, using a Samsung multi-axis robot, AT1, the controller performance and convenience of the PC-based controller have been verified by comparing to the commercial one.

  • PDF

An Artificial Neural Networks Application for the Automatic Detection of Severity of Stator Inter Coil Fault in Three Phase Induction Motor

  • Rajamany, Gayatridevi;Srinivasan, Sekar
    • Journal of Electrical Engineering and Technology
    • /
    • 제12권6호
    • /
    • pp.2219-2226
    • /
    • 2017
  • This paper deals with artificial neural network approach for automatic detection of severity level of stator winding fault in induction motor. The problem is faced through modelling and simulation of induction motor with inter coil shorting in stator winding. The sum of the absolute values of difference in the peak values of phase currents from each half cycle has been chosen as the main input to the classifier. Sample values from workspace of Simulink model, which are verified with experiment setup practically, have been imported to neural network architecture. Consideration of a single input extracted from time domain simplifies and advances the fault detection technique. The output of the feed forward back propagation neural network classifies the short circuit fault level of the stator winding.