Journal of Applied Sciences

Parameter Identification of Photovoltaic Models Based on Adaptive Differential Evolution with Decomposition

YAN Zhen, LI Shuijia, GONG Wenyin

2022, 40(5): 713-726. doi:10.3969/j.issn.0255-8297.2022.05.001

Asbtract ( 1562 )

PDF (680KB) ( 110 )

References | Related Articles | Metrics

In order to quickly, accurately and reliably identify the parameters of photovoltaic (PV) models under different environmental conditions, an improved adaptive differential evolution algorithm based on improved adaptive differential evolution with decomposition (IADE-D) is proposed. In IADE-D, first, an unknown parameter decomposition technique is proposed to reduce the dimension of a problem and thus reduce the complexity of the problem. Second, an improved adaptive differential evolution algorithm is employed to solve the decomposed unknown parameters. In order to verify the effectiveness of the proposed algorithm, it is used for the single diode-based PV panel model parameter identification, namely, multi-crystalline KC200GT. Simulation results show that the IADE-D algorithm proposed in this paper is more competitive in terms of the accuracy and reliability than some of the advanced algorithms proposed recently. Therefore, IADE-D can be considered as an effective method for parameter identification of PV models.

A Multi-distribution Evolutionary Algorithm with Differential Evolution

XU Yongjian, CHEN Yu, XIE Chengwang

2022, 40(5): 727-738. doi:10.3969/j.issn.0255-8297.2022.05.002

Asbtract ( 1637 )

PDF (992KB) ( 105 )

References | Related Articles | Metrics

A multi-distribution evolutionary algorithm with differential evolution (MDEA_DE) is proposed by incorporating the strong global convergence of distribution estimation algorithm and the fast convergence of differential evolution. To improve the global convergence ability, MDEA_DE employs a population-based multi-distribution evolution mechanism, and three Gaussian distributions are utilized to generate diverse population with solutions of high quality. Meanwhile, a search space regulation strategy is proposed to improve sampling precision of the Gaussian distributions, and local exploitation ability is enhanced by an improved differential evolution search in the solution space. Experimental results for selected benchmark problems demonstrate that MDEA_DE converges efficiently to the globally optimal solutions of complicated optimization problems by striking a good balance between global exploration and local exploitation.

Multi-model Multi-objective Optimization Algorithms with a New Environmental Selection Strategy

ZHANG Guochen, LIU Pengfei, SUN Chaoli

2022, 40(5): 739-748. doi:10.3969/j.issn.0255-8297.2022.05.003

Asbtract ( 1785 )

PDF (1185KB) ( 158 )

References | Related Articles | Metrics

In order to achieve the optimal solutions for multi-model multi-objective problems, in this paper a new environmental selection strategy is proposed for differential evolution approach. First, non-dominated solutions are kept to ensure the convergence in objective space; second, a population with good distribution in objective space is obtained by using its correlation with reference vectors; and then the next parent population is selected by simultaneously considering the convergence performance in objective space and the diversity performance in decision space. Experimental results on 11 multi-model multiobjective test problems show that the proposed method is efficient in solving multi-model multi-objective problems.

FWM Effect in 5G Fronthaul Transmission System Based on WDM-PON Architecture

YU Peihua, LI Zhengxuan, XU Yan, SONG Yingxiong

2022, 40(5): 749-757. doi:10.3969/j.issn.0255-8297.2022.05.004

Asbtract ( 1615 )

PDF (2240KB) ( 107 )

References | Related Articles | Metrics

Wavelength-division multiplexing passive optical network (WDM-PON) is considered as a mainstream solution for 5G fronthaul transmission system to cope with the rapidly increasing demand of bandwidth. However, nonlinearity effects, especially the fourwave mixing (FWM) effect in WDM-PON systems will impair system performances in the transmission distance and channel capacity. In this paper, the influence of FWM effect on multi-wavelength system with 25 Gb·s^-1 ·λ^-1 non-return-to-zero (NRZ) signal over 25 km standard single-mode fiber (SSMF) transmission is explored by experiment and simulation. Study results show that in a 200 GHz spaced 12-channel system, the bit error rate (BER) of all channels, except for the first and last channels, cannot reach the threshold 1.0× 10^-3 of forward error correction (FEC). This implies that FWM crosstalk should be given full consideration while making wavelength plans for 5G fronthaul system, including the wavelength distribution and channel spacing selection.

Network Traffic Classification Based on LSTM and Feature Generation

WANG Shuai, DONG Yuning, LI Tao

2022, 40(5): 758-769. doi:10.3969/j.issn.0255-8297.2022.05.005

Asbtract ( 1751 )

PDF (976KB) ( 202 )

References | Related Articles | Metrics

This paper proposes a network traffic classification method that combines feature generation and long short term memory (LSTM) model. This method analyzes and compares the classification performances of different feature generation methods using matrix multiplication feature generation method. The accuracy of original data and feature data on the classification problem is tested experimentally, and the results of convolutional neural network (CNN) and the proposed method are compared on network flow classification. The kernel function is used in the statistical feature, so that it can adapt to the LSTM input dimension and obtain better classification results. Experimental results on real network flow data show that the proposed method can achieve 93.9% accuracy in classification, and 99.2% in coarse grained classification task, and this performance is significantly better than that of existing methods.

Demodulation of LoRa Synchronous Superposition Signals

LEI Fang, CHEN Bo, Lü Jingzhao, LI Pingan, QIN Hong

2022, 40(5): 770-778. doi:10.3969/j.issn.0255-8297.2022.05.006

Asbtract ( 1709 )

PDF (586KB) ( 107 )

References | Related Articles | Metrics

In order to improve the transmission rate of LoRa technology, a new demodulation algorithm for LoRa synchronous superimposed signals is proposed. In this paper, we first theoretically find that the core problem of demodulation of LoRa superposition signals in the same expansion frequency channel is to distinguish the signals of different nodes, and if the power difference of two nodes is greater than 5 dB, the data of the two nodes can be demodulated. Moreover, the disadvantages of the method are also indicated. Second, we propose a algorithm of using the LoRa modulation symbol as the highest marker bit to distinguish the data of two nodes, indicate the application scenario of the algorithm, and prove that it has the same complexity as the demodulation algorithm of LoRa technique. Simulation results show that the new algorithm will not increase the bit error rate of the LoRa demodulation algorithm, and compared with the demodulation algorithm of LoRa technology, the new algorithm greatly improves the transmission rate of LoRa packets and reduces the bit energy consumption of transmitting effective information.

Recognition Method of Lateral Buckling of Submarine Pipeline Based on Distributed Optical Fiber Sensing

WU Wenjing, FENG Xin

2022, 40(5): 779-789. doi:10.3969/j.issn.0255-8297.2022.05.007

Asbtract ( 1621 )

PDF (8546KB) ( 136 )

References | Related Articles | Metrics

In view of the problem that the traditional subsea pipeline lateral buckling identification method cannot truly reflect the structural state of the subsea pipeline, this paper takes the subsea pipeline with initial defects as the research object, and uses the Euler-Bernoulli curve calculation method to obtain the distributed structural response of submarine pipelines based on distributed optical fiber sensing. The method establishes a pipeline buckling displacement reconstruction algorithm, and realizes the quantitative identification of the front and rear buckling behaviors of submarine pipelines under the condition of unknown loads. Taking the initial defect amplitude as variable parameter, the feasibility of the proposed method is verified by the finite element analysis and the model test of the initial defect submarine pipeline. Simulation results show that the reconstruction algorithm proposed in this paper can quantitatively identify the generation and development process of lateral buckling of submarine pipelines, and provide a feasible method for solving the problem of lateral buckling identification of submarine pipelines under unknown loads.

Signal and Interference Analysis of UFMC System with Timing Offset

WANG Jingjing, WEN Jiangang, ZOU Yuanping, WANG Anding, HUA Jingyu

2022, 40(5): 790-800. doi:10.3969/j.issn.0255-8297.2022.05.008

Asbtract ( 1775 )

PDF (2516KB) ( 82 )

References | Related Articles | Metrics

As the actual arrival time of a signal arriving universal a filtered multi-carrier (UFMC) receiver is inconsistent with the expected arrival time, timing offset (TO) occurs, which introduces inter symbol interference (ISI), destroys the orthogonality of subcarriers, and decreases the communication quality. In order to analyze the influence of TO on UFMC systems, this paper analyzes the signal and interference of UFMC systems with TO, derives corresponding theoretical expressions, numerically calculates and compares the signal to interference ratio (SIR) under different waveform filter lengths and different TOs, and simulates and compares the symbol error rate (SER) of UFMC systems in different situations. Simulation results show that the value of TO is positively correlated with the signal interference intensity, and the filter length is inversely correlated with the SER of systems.

Linearly Polarized Modes Amplifier Based on PbS doped Few-Mode Fiber

YANG Jinhong, SHANG Yana, WEI Huimei, PANG Fufei, WEN Jianxiang, DONG Yanhua, CHEN Na, CHEN Zhenyi

2022, 40(5): 801-808. doi:10.3969/j.issn.0255-8297.2022.05.009

Asbtract ( 1750 )

PDF (4077KB) ( 71 )

References | Related Articles | Metrics

In this paper, a new type of few-mode fiber, lead sulfide (PbS) doped fewmode fiber is fabricated using modified chemical vapor deposition (MCVD) technique. The fluorescence characteristics of the fiber is characterized. A few-mode fiber amplification system based on the spatial light modulator is built, and the mode gain characteristics of the fiber are measured. The experimental results show that the fiber exhibits ultrawideband near-infrared luminescence when pumped by a 980 nm laser, and the spectral range is from 1 050 nm to 1 650 nm. The LP₀₁ and LP₁₁ mode amplifications are realized with the average modal on-off gain of 4.5 dB and the differential modal gain is less than 0.6 dB within the band from 1 540 to 1 560 nm. The optical fiber provides the feasibility to realize broadband mode amplification in the mode division multiplexing system.

Inversion of Forest Canopy Height in Transmission Line Corridor Using ICESat-2 ATLAS & JL-1

LIN Haoyu, WU Zhaocong, SUN Xiaohu

2022, 40(5): 809-819. doi:10.3969/j.issn.0255-8297.2022.05.010

Asbtract ( 1634 )

PDF (7119KB) ( 147 )

References | Related Articles | Metrics

Aiming at the problem that ICESat-2 ATL08 product contains a lot of noise photons and gross errors of tree height, a method of photon filtering and gross errors elimination was designed by using a series of characteristics such as photon number, beam energy and ground coverage; The correlations of reflection parameters in each band with NDVI, RVI, SAVI, MSAVI, PVI vegetation indexes extracted from high-resolution multispectral data and with laser altimetry data were studied, and the effectiveness and accuracy of multiple regression and random forest model inversion were compared and analyzed. Study results showed that the random forest model can obtain higher inversion accuracy of tree height. The inversion experiment carried out in the transmission corridor of Lin' an District, Hangzhou City, Zhejiang Province obtained the result of RMSE=2.84 m. The tree height hidden danger map produced by combining the power data can provide auxiliary decision-making for the hidden danger investigation of transmission facilities.

Image Rotation Angle Estimation on JPEG Compressed Images via Third-order Differential Prefiltering Operation and Cyclostationarity Analysis

SONG Zhonghao, YAO Heng

2022, 40(5): 820-827. doi:10.3969/j.issn.0255-8297.2022.05.011

Asbtract ( 1681 )

PDF (1696KB) ( 54 )

References | Related Articles | Metrics

In this paper, we propose to estimate the rotation angle of JPEG compressed image based on prefilter and cyclic spectral analysis in order to improve detection performance. We first mathematically show the cyclostationarity of differentially filtered images. And then we design a prefilter to reduce the effect of JPEG block artifacts on the image spectrum, further improving the detection ability. At last, the rotation angle is estimated based on the locations of characteristic peaks. Experimental results demonstrate that the proposed detection method outperforms some state-of-the-art methods.

AM-AdpGRU Financial Text Classification Based on Cross-Domain

WU Feng, XIE Cong, JI Shaopei

2022, 40(5): 828-837. doi:10.3969/j.issn.0255-8297.2022.05.012

Asbtract ( 1635 )

PDF (903KB) ( 51 )

References | Related Articles | Metrics

Aiming at the problem that the current financial text classification model based on deep learning heavily depends on labeled data, this paper proposes an am AM-AdpGRU financial text classification model based on cross domain migration, which migrates related domain data to the target domain data by learning the classification criteria of the data. The am AM-AdpGRU model first uses deep network adaptation to overcome the migration loss caused by the difference of data distribution between the source domain and the target domain, so that the model does not need to be reconstructed even when the data distribution changes; Then, the feature selection principle of the target domain to the source domain is established by using attention mechanism, so that the model's attention to the source domain can focus on the part with higher similarity with the target domain. Experiments are carried out on the open cross domain emotion review Amazon dataset and semeval-2017 microblog financial dataset, and the am AM-AdpGRU model is compared with other methods. Experimental results show that the average classification accuracy of am AM-AdpGRU model is significantly improved compared with other models.

Review of Neural Network Pruning Techniques

JIANG Xiaoyong, LI Zhongyi, HUANG Langyue, PENG Mengle, XU Shuyang

2022, 40(5): 838-849. doi:10.3969/j.issn.0255-8297.2022.05.013

Asbtract ( 2513 )

PDF (1569KB) ( 830 )

References | Related Articles | Metrics

This paper summaries the origin and research progress of neural network pruning technologies, divides them into two categories of unstructured pruning with sparse weight parameters and coarse-grained structured pruning, and introduces the representative methods of the two categories in recent years. Because pruning reduces model parameters and compresses the model size, depth models can be applied to embedded devices, showing the importance of pruning in the field of deep learning model compression. In view of the existing pruning technologies, this paper expounds the problems existing in practical applications and measurement standards, and prospects the research and development tendency in the future.

LC Inverter Based on Novel Prescribed Performance Control

SHI Jianqiang, LI Shuang

2022, 40(5): 850-864. doi:10.3969/j.issn.0255-8297.2022.05.014

Asbtract ( 1586 )

PDF (1054KB) ( 83 )

References | Related Articles | Metrics

In order to improve the performance of LC inverters, a novel prescribed performance voltage controller is presented in this paper. Firstly, a mathematical model is established based on circuit theory and topology. Secondly, a finite-time prescribed performance function is employed to overcome the disadvantage of slow convergence in exponential function. Thirdly, with the purpose of making the voltage tracking error converge along the desired trajectory, a voltage controller is designed. Finally, a disturbance observer is designed to reduce the effects of load or parameters changes and other disturbance factors on output voltage. Simulation results obtained with MATLAB/Simulink platform verify the feasibility and effectiveness of the proposed novel prescribed performance control strategy.

Fatigue Failure Model of IGBT Chip Based on Threshold Voltage

LI You, CAO Jiwei, HAO Guangyao, YAN Ge, LIU Hongxiao

2022, 40(5): 865-875. doi:10.3969/j.issn.0255-8297.2022.05.015

Asbtract ( 1645 )

PDF (1175KB) ( 248 )

References | Related Articles | Metrics

In order to effectively evaluate the health status of IGBT during its whole life cycle, the fatigue failure mechanism of IGBT chip was studied based on the theory of semiconductor physics, and the effect of charge density at gate interface on threshold voltage was analyzed. Taking the threshold voltage as the failure characteristic quantity of IGBT, the fatigue failure model of IGBT chip was established on the basis of studying the change rule of threshold voltage with fatigue failure time. An IGBT threshold voltage test platform was built, and IGBT aging experiments were performed to verify that the model proposed in this paper can accurately characterize and estimate the aging degree of IGBT chips, and the correctness and rationality of the failure model were verified.

Graphene Preparation Control System Based on Adaptive Fuzzy PID

JI Changpeng, WANG Zirui

2022, 40(5): 876-886. doi:10.3969/j.issn.0255-8297.2022.05.016

Asbtract ( 1604 )

PDF (705KB) ( 46 )

References | Related Articles | Metrics

The electrolytic process of graphene preparation is nonlinear and uncertain. Using the traditional proportional integration differentiation (PID) control algorithm is difficult to accurately control the electrolytic current and electrolyte ion concentration in the preparation process, resulting in the problems of uneven agglomeration and stripping in the process of graphene preparation by electrolysis, Accordingly, an adaptive fuzzy PID algorithm is designed and the corresponding mathematical model is established. The process accuracy of graphene preparation can be better controlled by using the model. The comparative simulation of adaptive fuzzy PID algorithm and traditional PID algorithm is carried out with MATLAB, and the simulation results are analyzed and discussed. The adaptive fuzzy PID algorithm is better in the performs of response time, anti-interference ability and stability, and more in line with the requirements of graphene preparation control system than the traditional PID algorithm.

Table of Content