Search

The fuel cell grid-tied power generation system usually includes a dc-dc converter and a dc-ac inverter. In a single-phase system, inherent low-order current pulsations are introduced into the system, which can have harmful effects on the fuel cell stack. For example, reducing the output voltage and output efficiency, a reduction in service life, and even accelerates the degradation rate of the membrane electrode of a proton exchange membrane fuel cell (PEMFC). In addition, dc/ac coupling power can cause distortion in the dc input current and ac grid current. To eliminate the input ripple and ensure high ac power quality on the grid side, this paper proposes a novel power decoupling control for single-phase grid-tied PEMFC systems, which uses an improved model predictive control (MPC) algorithm. With the help of the virtual vector methods, which are realized by a two-stage optimization method, excellent tracking effect and robustness can be ensured. Simulations and experimental results show that the proposed algorithm can not only completely eliminate the input current ripple and reduce the total harmonic distortion (THD) of ac current on the grid side, but also improve the transient performance of the system.

In a wind-diesel isolated microgrid, the fluctuating output power of wind turbine generator (WTG) and the perturbation of the load demand will lead to power imbalance and frequency deviation in the system. Appropriate frequency control schemes are indispensable to guarantee power quality through maintaining the power balance. In view of the outstanding performance of Equivalent-Input-Disturbance (EID) method, it is taken to the task of load frequency control (LFC) problem of this microgrid. In this paper, novel double EID controllers are proposed for the frequency control of a wind-diesel isolated microgrid. In this integrated control design, one single EID controller is applied to the pitch angle control system to smooth output power of WTG by controlling the pitch angle. At the diesel generator side, another single EID controller is applied to adjust the output power of the diesel generator to maintain the power balance of the system and finally preserve frequency in the normal range. Battery energy storage system is connected as auxiliary regulation. The simulation studies show superior flexibility and control performance of the proposed strategy compared to the conventional PI method in a wind-diesel isolated microgrid in MATLAB.

Deep neural networks have received increasing attention in human action recognition. Previous research has established that utilizing 3D convolution is a reasonable approach to learn spatio-temporal representation. Nevertheless, constructing effective 3D ConvNets usually need an expensive pre-training process that performing on a huge-scale video dataset. To avoid this burdensome situation, one major issue is to determine whether the pre-trained parameters of 2D convolution networks can be directly bootstrapped into 3D. In this paper, we devise a 2D-Inflated operation and a parallel 3D ConvNet architecture to solve this problem. The 2D-Inflated operation is used for converting pre-trained 2D ConvNets into 3D ConvNets, which avoiding video data pre-training. We further explore the optimal quantity of 3D ConvNet in the parallel architecture, and the results suggest that 6-nets architecture is an excellent solution for recognition. Another contribution of our study is two practical and valid skills, accumulated gradient descent and video sequence decomposition. Either of those techniques can promote the improvement of performance. The recognition results of UCF101 and HMDB51 reveal that, without the video data pre-training, our 3D ConvNets still can achieve competitive performance to the other generic and recent methods of using 3D ConvNets in the RGB image domain.