Passive shimming is cheaper, simpler and more effective than active shimming. However it is very tedious to realize passive shimming because the position determination of shimming elements is concerned with the experi...
详细信息
Passive shimming is cheaper, simpler and more effective than active shimming. However it is very tedious to realize passive shimming because the position determination of shimming elements is concerned with the experience of the designer. This paper presents a useful and simple method to design the passive shimming system for homogeneous permanent magnets. An optimal model is set up to determine the location and size of the ferromagnetic shimming elements. To obtain a global solution of the optimization problem with less computational effort, a hybrid optimal algorithm is presented. First, the improved genetic algorithm is used as the initial optimal tool. Then the optimal shim's size is computed by steepest descent algorithm combined with design sensitivity analysis starting from the proper initial solution. Finally, the shim's position is chosen as the design variables to obtain the global optimal solution. A numerical implementation shows the proposed method can design the shimming system intelligently and effectively.
Short-term load forecasting(STLF) is of great importance for the safety and stabilization of grids. Based on the historical load data of meritorious power of some area in Guizhou power system, three BP neural networks...
详细信息
Short-term load forecasting(STLF) is of great importance for the safety and stabilization of grids. Based on the historical load data of meritorious power of some area in Guizhou power system, three BP neural networks in steepest descent algorithm back propogation neural network (SDBP), Levenberg -Marquardt algorithm back propogation neural network ( LMBP) and Bayesian regularization algorithm back propogation neural network ( BRBP ) models in 24 hours ahead prediction are compared. Since the traditional BP algorithm has some drawbacks such as slow training convergence speed and possibility of local minimizing the optimized function, an optimized L-M algorithm, which can improve the stability of convergence and accelerate the training speed of neural network has been applied to carry out load forecasting work to reduce the mean relative error. Bayesian regularization also be applied which can overcome and improve the generalization of neural network. The prediction precision of BRBP are superior to LMBP and SDBP,while BRBP has poor training speed than others.
To reduce the backflow power of dual-active- bridge (DAB) DC-DC converter and improve the operating efficiency, this paper proposes a gradient descentalgorithm (GD- EPS) based extended-phase-shift (EPS) control. Firs...
详细信息
ISBN:
(数字)9781728116754
ISBN:
(纸本)9781728116761
To reduce the backflow power of dual-active- bridge (DAB) DC-DC converter and improve the operating efficiency, this paper proposes a gradient descentalgorithm (GD- EPS) based extended-phase-shift (EPS) control. Firstly, the transmission and backflow power models of DAB converter with EPS and single-phase-shift (SPS) control are established, and the influencing factors of backflow power are analyzed. Secondly, according to the backflow power model, the gradient descentalgorithm for solving the optimal solution of backflow power is obtained. On this basis, the output voltage closed-loop control is added to derive the DAB overall optimal control scheme. At last, the simulation models of DAB converter with GD-EPS and SPS control are built and compared, it is verified that the GD-EPS control can optimize the backflow power better than the SPS control when the load and input voltage change.
With the development of the missile digital control and network transmission technologies, the data bus transmit scheme has been paid more and more attentions. However, communication networks inevitably introduce time...
详细信息
With the development of the missile digital control and network transmission technologies, the data bus transmit scheme has been paid more and more attentions. However, communication networks inevitably introduce time delay. To compensate the time delay effects, an application of fuzzy controller to such system is presented. The main contents are adjusting PID parameters using fuzzy logic method. Moreover, an optimization method is introduced to tune the consequent parameters on-line such that a certain performance index is minimized. Simulation results show the effectiveness of these methods.
This paper describes techniques to implement gradient-descent-based machine learning algorithms on crossbar arrays made of memristors or other analog memory devices. We introduce the Unregulated Step descent (USD) alg...
详细信息
ISBN:
(纸本)9781479919611
This paper describes techniques to implement gradient-descent-based machine learning algorithms on crossbar arrays made of memristors or other analog memory devices. We introduce the Unregulated Step descent (USD) algorithm, which is an approximation of the steepest descent algorithm, and discuss how it addresses various hardware implementation issues. We discuss the effect of device parameters and their variability on performance of the algorithm by using artificially generated and real-world datasets. In addition to providing insights on the effect of device parameters on learning, we illustrate how the USD algorithm partially offsets the effect of device variability. Finally, we discuss how the USD algorithm can be implemented in crossbar arrays using a simple 4-phase training scheme. The method allows parallel update of crossbar memory elements and reduces the hardware cost and complexity of the training architecture significantly.
According tothe modern control and optimal estimation theory, a mathematical model of the pilot attention resource allocation is developed. This model analyzes the inherent characteristics of pilotinput and output lin...
详细信息
According tothe modern control and optimal estimation theory, a mathematical model of the pilot attention resource allocation is developed. This model analyzes the inherent characteristics of pilotinput and output links in the pilot-vehicle closed-loop system, and the optimal control model of pilotis established by simplification and mathematical modeling. The control task chosen in this paper concerns the longitudinal motion of a hovering aircraft. Tosolve the pilot optimal attention allocation problem in the instrument-monitoring task of the pilot-vehicle closedloop system an objective function is set up based on flight control system state equation. In the process of solving optimization problem, the constrained optimization problem is converted intoan unconstrained problem, and is solved effectively with the method of steepest descent algorithm in the MATLAB.
The total variation (TV) regularization method is very attractive for various image processing applications. In order to apply the TV approach to motion pictures, it is required to reduce the computational time of the...
详细信息
ISBN:
(纸本)9781424458561
The total variation (TV) regularization method is very attractive for various image processing applications. In order to apply the TV approach to motion pictures, it is required to reduce the computational time of the iterative signal processing of the TV regularization. In this paper, we propose the diagonal TV criterion instead of conventional isotropic or anisotropic TV criteria. The experimental results show that we can obtain almost half of the iteration number both in the steepest descent algorithm and Chambolle 's algorithm by utilizing the diagonal TV compared with the conventional TV criteria.
This paper presents a novel digital blind calibration method for time interleaved analog to digital converters (TIADCs). A simple cost function based on the cross-correlation of channel statistics is used to derive a ...
详细信息
ISBN:
(纸本)9781479903573
This paper presents a novel digital blind calibration method for time interleaved analog to digital converters (TIADCs). A simple cost function based on the cross-correlation of channel statistics is used to derive a steepest descent algorithm for the compensation of timing mismatch errors. Instead of calibrating the timing mismatches independently for each channel, only one adaptation channel needs to be calibrated within a closed loop. The calibration of the rest of the channels can be coordinated according to a scaling relationship established during an initialization stage. As a result, both the computational complexity and convergence speed of the proposed algorithm can be improved significantly with little loss in calibration performance.
暂无评论