Search

Our devices can use a wide range of communication technologies such as multiple cellular technologies (4G/5G), WiFi, and also Ethernet. At the same time, applications have a choice of a wide range of transport protocols such as QUIC and TCP that can be fine-tuned and optimized according to their needs. However, in spite of these advances, offering seamless multiconnectivity to applications continues to be a hard problem. The key factors that continue to be a roadblock towards achieving seamless multiconnectivity include a) applications cannot specify the communication technologies to be used by their flows, and b) the traditional definition of a connection endpoint was not designed to support mobile nodes. In this paper we discuss the key challenges that make this problem hard. We also present MULTI, a session layer approach that can be leveraged to address some of the key sub-problems of this problem. For instance, we observe that MULTI incurred a small overhead (less than 5% decrease in throughput) when using TCP compared to the native asyncio python library.

Railway passenger comfort has been considered a growingly important field to attract more passengers from other public transports such as air flights. To allow passengers and train companies to estimate the onboard passenger comfort level, we propose a phone-based hybrid machine learning (ML) model combining pre-train convolutional neural network as a feature extractor and support vector regressor as a predictor. To better demonstrate the capacity of the proposed model, two sub-models of the hybrid model and the same hybrid model but with non-pre-train feature extractor are adopted to be benchmarks. The raw field data is acquired from a corridor between the University of Birmingham station and Birmingham International station using phones, subsequently calculated to corresponding comfort level according to UIC 513. The four models are trained by the dataset in two domains – time domain and frequency domain, then optimized by random search and validated by 10-fold cross-validation. The proposed method yields the best performance with an $R^{2}$ of 0.988 ± 0.004, a root-mean-square error (RMSE) of 0.028 ± 0.015, and a mean-absolute-error (MAE) of 0.02 ± 0.005. The results of this study underpin the possibility that the railway passenger has the access to quantify the level of comfort and the real-time assistance for the train driver to calibrate the driving style from the proposed system.

Phishing attackers spread phishing links through e-mail, text messages, and social media platforms. They use social engineering skills to trick users into visiting phishing websites and entering crucial personal information. In the end, the stolen personal information is used to defraud the trust of regular websites or financial institutions to obtain illegal benefits. With the development and applications of machine learning technology, many machine learning-based solutions for detecting phishing have been proposed. Some solutions are based on the features extracted by rules, and some of the features need to rely on third-party services, which will cause instability and time-consuming issues in the prediction service. In this paper, we propose a deep learning-based framework for detecting phishing websites. We have implemented the framework as a browser plug-in capable of determining whether there is a phishing risk in real-time when the user visits a web page and gives a warning message. The real-time prediction service combines multiple strategies to improve accuracy, reduce false alarm rates, and reduce calculation time, including whitelist filtering, blacklist interception, and machine learning (ML) prediction. In the ML prediction module, we compared multiple machine learning models using several datasets. From the experimental results, the RNN-GRU model obtained the highest accuracy of 99.18%, demonstrating the feasibility of the proposed solution.

More than half of all commercial aircraft operation accidents could have been prevented by executing a go-around. Making timely decision to execute a go-around manoeuvre can potentially reduce overall aviation industry accident rate. In this paper, we describe a cockpit-deployable machine learning system to support flight crew go-around decision-making based on the prediction of a hard landing event. This work presents a hybrid approach for hard landing prediction that uses features modelling temporal dependencies of aircraft variables as inputs to a neural network. Based on a large dataset of 58177 commercial flights, the results show that our approach has 85% of average sensitivity with 74% of average specificity at the go-around point. It follows that our approach is a cockpit-deployable recommendation system that outperforms existing approaches.

Missing data are unavoidable in the real-world application of unsupervised machine learning, and their nonoptimal processing may decrease the quality of data-driven models. Imputation is a common remedy for missing values, but directly estimating expected distances have also emerged. Because treatment of missing values is rarely considered in clustering related tasks and distance metrics have a central role both in clustering and cluster validation, we developed a new toolbox that provides a wide range of algorithms for data preprocessing, distance estimation, clustering, and cluster validation in the presence of missing values. All these are core elements in any comprehensive cluster analysis methodology. We describe the methodological background of the implemented algorithms and present multiple illustrations of their use. The experiments include validating distance estimation methods against selected reference methods and demonstrating the performance of internal cluster validation indices. The experimental results demonstrate the general usability of the toolbox for the straightforward realization of alternate data processing pipelines. Source code, data sets, results, and example macros are available on GitHub. https://github.com/markoniem/nanclustering_toolbox peerReviewed

This article presents a short-term decision-making model for an electricity retailer using bilevel stochastic programing. In the proposed model, a liberalized distributed renewable energy (DRE) market in which the retailer competes with other load serving entities (LSEs) for procuring DRE is proposed. The retailer, in the upper level, decides its level of involvement in the day-ahead and real-time markets, as well as the price bids offered to DRE producers for every time period, with the goal of minimizing its expected procurement cost at a predefined risk level. On the other hand, DRE producers, in the lower level, react to the price bids offered by the retailer under study and other LSEs, to maximize their total revenues. The stochastic nature of day-ahead and real-time market prices, DRE production, electricity demand, and price bids of the retailer's rival market agents (RMAs) is taken into the formulation of the proposed model. By using the Karush-Kuhn-Tucker (KKT) optimality conditions and duality theory, the bilevel problem is transformed into its equivalent single-level mixed-integer linear programming (MILP) problem. Case studies are performed to show the effectiveness of the proposed model.

Peer-to-peer (P2P) energy trading is one of the promising approaches for implementing decentralized electricity market paradigms. In the P2P trading, each actor negotiates directly with a set of trading partners. Since the physical network or grid is used for energy transfer, power losses are inevitable, and grid-related costs always occur during the P2P trading. A proper market clearing mechanism is required for the P2P energy trading between different producers and consumers. This paper proposes a decentralized market clearing mechanism for the P2P energy trading considering the privacy of the agents, power losses as well as the utilization fees for using the third party owned network. Grid-related costs in the P2P energy trading are considered by calculating the network utilization fees using an electrical distance approach. The simulation results are presented to verify the effectiveness of the proposed decentralized approach for market clearing in P2P energy trading. Agency for Science, Technology and Research (A*STAR) Accepted version This work was supported by the Agency for Science, Technology and Research (A*STAR) through its Singapore–Germany Academic-Industry (2+2) International Collaboration under Grant A1990b0060. Paper no. TSG-01805-2019.

This article develops a novel nonlinear controller for the mitigation of drivetrain torsional oscillations of wind turbine generators following large disturbances. The key idea is to integrate differential geometry theory with an extended state observer. Differential geometry allows transforming the wind turbine generator nonlinear model into a simple second-order Brunovsky system, whereas the extended state observer can accurately estimate the states of the transformed Brunovsky system and compensate unknown disturbances. Simulation results show that, under various conditions in different test systems, the proposed controller achieves significant enhancements with respect to conventional approximate linearization-based controller and nonlinear sliding mode control-based method.

This letter presents a robust generalized correlation approach for monitoring of power system oscillatory modes. Due to the non-Gaussian nature of measurement noise, and presence of impulsive outliers, the wide-sense stationarity assumption of power system measurement data is violated even under ambient conditions thus limiting the effectiveness of conventional autocorrelation based approaches in mode metering. The fundamental definition of correlation function for power system measurements is extended to a generalized correlation function in order to incorporate higher order statistics into existing modal estimation technique. Simulations on synthetic and real-life test signals have been conducted to demonstrate the robustness of the proposed method.

This paper presents a grid-forming control (GFC) scheme for two-stage photovoltaic (PV) systems that maintains power reserves by operating below the maximum power point (MPP). The PV plant in GFC mode behaves like a voltage source that supports the grid during disturbances in full or limited grid-forming mode as per the reserve availability. This is a model-free method that avoids the estimation of MPP power in real-time commonly done in the literature, which makes it simpler and more reliable. The proposed control also features an enhanced current limitation scheme that guarantees containment of the current overshoots during faults, which is not trivial in voltage-sourced GFC inverters. A thorough investigation is done, exploring various generation mixtures of synchronous machines (SM), GFC and grid-following (GFL) inverters, and all common disturbances, e.g., load change, faults and irradiance transients. The results show very favorable dynamic performance by the GFC inverters, far superior to GFL inverters and directly comparable to SMs. It is found that replacing SMs with GFC inverters may improve the frequency profile and terminal voltage during disturbances, despite losing out in the mechanical inertia and the strict inverter overcurrent limits.

The increasing uncertainty level caused by growing renewable energy sources (RES) and aging transmission networks poses a great challenge in the assessment of total transfer capability (TTC) and available transfer capability (ATC). In this paper, a novel data-driven sparse polynomial chaos expansion (DDSPCE) method is proposed for estimating the probabilistic characteristics (e.g., mean, variance, probability distribution) of probabilistic TTC (PTTC). Specifically, the proposed method, requiring no pre-assumed probabilistic distributions of random inputs, exploits data sets directly in estimating the PTTC. Besides, a sparse scheme is integrated to improve the computational efficiency. Numerical studies on the modified IEEE 118-bus system demonstrate that the proposed DDSPCE method can achieve accurate estimation for the probabilistic characteristics of PTTC with a high efficiency. Moreover, numerical results reveal the great significance of incorporating discrete random inputs in PTTC and ATC assessment, which nevertheless was not given sufficient attention., Comment: Accepted by IEEE Transactions on Power Systems for future publication

Development of protection scheme for microgrid is a challenging issue due to the inclusion of distributed generations (DGs). The research work for developing the primary protection scheme for microgrid is progressing continuously. This paper proposes an integrated impedance angle (IIA) based protection scheme using wide-area positive sequence components of voltages and currents. The proposed scheme uses data retrieved from the phasor measurement units (PMUs) of IEEE C37.118.1 complied standards. Furthermore, the IIA for the line is computed considering both ends PMU information, which is the key indicator for identifying the faults in the microgrid. The proposed protection scheme is extensively tested considering the different operating conditions of microgrid, variation in DG penetration and variation in fault parameters such as fault resistance, fault types, and fault location. Furthermore, cross validation on no-fault situations including section cutoff, disconnection of high penetration DG, grid-islanding resynchronization, motor starting, capacitor switching, and load encroachment are carried out. The proposed scheme is also tested for the insensitivity towards the external faults. The test results on both 15-bus and 34-bus systems clearly indicate that the proposed IIA-based wide-area scheme can be a potential protection measure for microgrid under varied operating circumstances.

The information and communication technology (ICT) is increasingly involved in the measurement, operation, control and protection of the cyber-physical power system (CPPS). As one of the most promising communication infrastructure projects, Starlink is a satellite Internet constellation being constructed by SpaceX for providing global Internet access, which can benefit the network service supply for the communication-enabled power system equipment in locations where the network access is unreliable, expensive, or completely unavailable. In this letter, the future applications of the Starlink space network in CPPSs are explored: the communication infrastructure and transmission parameters are discussed, and the corresponding test case was emulated in real-time on the heterogeneous co-emulation platform to validate the proposed concept of space network enhanced cyber-physical power system.

With the integration of multiple types of loads and renewable generations, the number of system states significantly grows. As a result, running optimal power flow (OPF) to analyze a myriad of system states is challenging and this seriously restricts the efficiency of the state enumeration method. To address that, this paper proposes a Lagrange Multiplier based State Enumeration (LMSE) approach to accelerate the analysis without loss of accuracy. The core idea is to directly obtain the optimal load shedding of contingency states by Lagrange multiplier-based functions, rather than the time-consuming OPF algorithms. This approach can also be conveniently integrated with the impact-increment method and the clustering technique for further efficiency enhancement. Case studies are performed on the RTS-79 and IEEE 118-bus systems considering multiple types of loads, photovoltaics (PVs), and wind turbines (WTs). Results indicate that the proposed method can significantly reduce the computing time without compromising the calculation accuracy.

The rapid development of Internet in recent years has led to a proliferation of social media networks as people who can gather online to share information, knowledge, and opinions. However, the network public opinion tends to generate strongly misleading and a large number of messages can cause shocks to the public once major emergencies appear. Therefore, we need to make correct prediction regarding and timely identify a potential crisis in the early warning of network public opinion. In view of this, this study fully considers the features of development and the propagation characteristics, so as to construct a network public opinion early warning index system that includes 4 first-level indicators and 13 second-level indicators. The weight of each indicator is calculated by the “CRITIC” method, so that the comprehensive evaluation value of each time point can be obtained and the early warning level of internet public opinion can be divided. Then, the Back Propagation neural network based on Genetic Algorithm (GA-BP) is used to establish a network public opinion early warning model. Finally, the major public health emergency, COVID-19 pandemic, is taken as a case for empirical analysis. The results show that by comparing with the traditional classification methods, such as BP neural network, decision tree, random forest, support vector machine and naive Bayes, GA-BP neural network has a higher accuracy rate for early warning of network public opinion. Consequently, the index system and early warning model constructed in this study have good feasibility and can provide references for related research on internet public opinion.

Passing clouds and wind gusts can create unacceptable rapid voltage/power variations in power networks. Simulation results using a real Australian distribution feeder with real load demand and PV output profile show that with a high penetration of PV the voltage variations can increase beyond the allowable limit given by the standards. Similarly, wind gusts can create wind power output variations. This paper addresses the rapid voltage/power variations caused by solar or wind power outputs and presents a control strategy using the energy buffer in energy storage for their impact mitigation. By controlling the discharging/charging operation of the energy storage based on the available energy buffer in the storage unit, not only the rate of power output variations can be maintained at the desired level, but also the voltage variations can be controlled within acceptable limits. An experimental setup, consisting of a PV emulator, inverter, and grid simulator, demonstrates the potential threat of unacceptable voltage variations. The proposed mitigation strategy using energy buffer in the storage for smoothing the power outputs of solar/wind is tested and validated through simulation. Simulation results show that with the proposed control, the fast variations of the voltage/power variations instigated by solar/wind can be maintained within the acceptable limits.

Efficient, robust, and accurate early flood warning is a pivotal decision support tool that can help save lives and protect the infrastructure in natural disasters. This research builds a hybrid deep learning (ConvLSTM) algorithm integrating the predictive merits of Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) Network to design and evaluate a flood forecasting model to forecast the future occurrence of flood events. Derived from precipitation dataset, the work adopts a Flood Index ( $I_{F}$ ), in form of a mathematical representation, to capture the gradual depletion of water resources over time, employed in a flood monitoring system to determine the duration, severity, and intensity of any flood situation. The newly designed predictive model utilizes statistically significant lagged $I_{F}$ , improved by antecedent and real-time rainfall data to forecast the next daily $I_{F}$ value. The performance of the proposed ConvLSTM model is validated against 9 different rainfall datasets in flood prone regions in Fiji which faces flood-driven devastations almost annually. The results illustrate the superiority of ConvLSTM-based flood model over the benchmark methods, all of which were tested at the 1-day, 3-day, 7-day, and the 14-day forecast horizon. For instance, the Root Mean Squared Error (RMSE) for the study sites were 0.101, 0.150, 0.211 and 0.279 for the four forecasted periods, respectively, using ConvLSTM model. For the next best model, the RMSE values were 0.105, 0.154, 0.213 and 0.282 in that same order for the four forecast horizons. In terms of the difference in model performance for individual stations, the Legate-McCabe Efficiency Index (LME) were 0.939, 0.898, 0.832 and 0.726 for the four forecast horizons, respectively. The results demonstrated practical utility of ConvLSTM in accurately forecasting $I_{F}$ and its potential use in disaster management and risk mitigation in the current phase of extreme weather events.

This paper proposes a co-design optimization procedure of a high-power off-board charger for electric vehicle (EV) applications. The primary purpose is to design a 175 kW SiC DC-charging system with high power density to achieve high efficiency at a wide operating range. For the active part of the DC off-board charger, a three-phase active front end (AFE) rectifier topology is considered in the design optimization and the modelling. The design methodology focuses on the optimal design of the passive filters, accurate electro-thermal modelling of the converter, inductor design, capacitor selection, loss and geometric modelling of the passive filters and control system design. The design optimization of the high-power charging system is performed in MATLAB Simulink using a closed-loop dynamic electro-thermal simulation of the off-board charger. The switching frequency, loss and temperature-dependent efficiency of the charger is investigated in parallel. Through this proposed technique, efficiency greater than 96% is achieved at a switching frequency of 40 kHz, along with a smaller size and lower weight of the system. Moreover, it operates with a current total harmonic distortion (THDi) below 3% and a power factor (PF) above 99% at rated power condition.

This paper presents a fully optimized fuzzy proportional-integral-derivative with filter (FPIDF) load frequency controller (LFC) for enhancing the performance of a hybrid microgrid system. The Marine Predator Algorithm (MPA), a recent optimization algorithm, is used to optimize the gains as well as the input scaling factors and membership functions of the proposed fuzzy PIDF controller. The controller performance is tested on a two-area hybrid microgrid system containing various renewable energy sources and energy storage devices. The effectiveness of the MPA based FPIDF controller is compared with conventional PIDF and FPIDF controllers based on other heuristic techniques presented in literatures. Moreover, different scenarios are implemented in this study to verify the robustness and sensitivity of the proposed controller to different step load perturbations, system parameters variations, wind speed fluctuation and solar irradiance variation. The dynamic response of the system is compared using different controllers in terms of settling time, maximum overshoot and undershoot values. The results are presented in the form of time domain simulations conducted via MATLAB/SIMULINK.

Small cells are low-power and low-range radio access nodes that can be used to increase network capacity in reduced areas with high traffic demand and to provide coverage to small isolated zones. As their number is expected to be very high, connecting them to the aggregation point in a cost-efficient manner is a key challenge, and there is consensus that none of the current technologies is valid for all scenarios. Since lamp-posts are one of the most convenient locations for outdoor small cells, this work analyzes the use of power line communications (PLC) over outdoor public lighting networks (OPLN) as a backhaul technology for outdoor small cell deployments. To this end, the characteristics of the PLC channels established in OPLN are firstly assessed. The analysis combines noise measurements performed in existing OPLN and channel responses obtained from a multiconductor transmission line (MTL) model. The attenuation, delay spread and spatial correlation of the multiple-input multiple-output (MIMO) channels are investigated and their influence in the physical layer parameters of PLC systems is discussed. Afterward, the performance achieved by state-of-the-art PLC systems is assessed. To this end, estimations obtained by means of simulations and measurements taken in actual networks are included. Data throughput achieved by PLC systems based on the ITU-T G.hn standard, as well as the expected improvements obtained by $3\times 3$ MIMO systems, are given. Results indicate that PLC can be an interesting technology for coverage-driven small cell deployments with backhaul length shorter than 150 m, offering throughput values similar to existing Sub-6 GHz wireless solutions in non-line-of-sight (NLOS) conditions and wired ones like G.fast.

The integration of photovoltaic and battery energy storage systems into utility grids is favorable for electricity customers, especially for high consumption load patterns due to the high electricity bill. To increase the annual bill savings and decrease the dependency on the utility grid, a procedure of optimally sizing the PV battery system is presented in this paper. A MATLAB-based code of the genetic algorithm is used to maximize the system self-sufficiency and minimize the discounted payback period while guaranteeing the system profitability. The optimization technique is introduced under the time-of-use tariff structure for residential prosumers in two scenarios. The first introduces a sellback energy price equal to the off-peak energy price, and the second introduces a sellback energy price equal to the on-peak energy price. The economic and technical effects of this optimization are described and analyzed. The effect of PV system capital cost, battery capital cost, and discount rate variations is evaluated by sensitivity analysis. The results indicate that for a sellback price equal to the off-peak price, a minimum discounted payback period can be achieved by installing the maximum PV system size without batteries, and it is reduced to half of its value for a sellback price equal to the on-peak price at the same PV system capacity. For a sellback price equal to the off-peak price, the minimum DPBP of 5.8 years can be achieved by installing a 20 kWp PV system without batteries with 47% self-sufficiency and 2.4 years for a sellback price equal to the on-peak price by installing the same system capacity. The proposed approach highlights that increasing the battery size leads to high self-sufficiency without a considerable decrease in the annual bill savings that encourage customers to invest in the PV battery system.

Flaring has always been an inseparable part of oil production and exploration. Previously, waste gas collected from different parts of facilities was released for safety or operational reasons and combusted on top of a flare stack since there was not the possibility to treat or use this type of gas. Concerns about global warming led to several initiatives for reducing flaring or even eliminating combustion. Treating flare gas was made possible by the introduction of flare gas recovery systems that have become increasingly obligatory. Most solutions add a flare gas recovery system to an existing flare system. In a typical scenario, after analyzing the existing facility and collecting the necessary data, alternative designs are proposed and criteria are determined to make a choice between the proposed alternatives. In this paper two designs of a gas control system are proposed, and reliability was chosen as the deciding factor. Using repairable dynamic fault trees, the failure models of the two designs have been implemented. Afterwards, a novel hybrid technique, the Stochastic Hybrid Fault Tree Automaton, is used to model the working conditions in which the system operates, with the aim to achieve a more realistic assessment and evaluate the disaster likelihood associated to these failures. It is shown that the latter enables a richer analysis where the effects of failure can be better assessed. This is important for correct choice between design alternatives because, as shown in the case study, the results of the two analyses can lead to contrasting conclusions of the solution to adopt. Further investigations have been carried out focusing on the safety sub-systems and on the basic events in each design. The Importance Measure analysis revealed that some of the components were responsible for most of the critical failures, thus locating some areas of possible design improvement.

A single-pump multi-actuator hydraulic system has an amount of coupling loss because the output pressure of the pump adapts the actuator with the highest load and the other actuators with lower loads have to consume the pressure difference in throttling valves. In this paper, a multi-pump multi-actuator system with on/off valve matrix is developed to reduce the coupling throttling loss. The total flow rate of all pumps matches the requirement of all actuators so as to avoid unnecessary increase of installation power. With the help of the on/off valve matrix, on the one hand, the multiple actuators can be decoupled into several single-pump single actuator systems separately in low flow demand. On the other hand, the multiple pumps can interflow to satisfy large flow demand. A two-level control strategy is proposed to achieve energy saving and good precision tracking simultaneously. The first level is to control the on/off switches and actuator modes based on the optimal energy consumption. The second level is to control the pumps and the throttling valves aiming at desired tracking precision. In addition, a compensation strategy is proposed to reduce the flow shock in on/off switches by transiting the control signals smoothly. Finally, the effectiveness of the proposed hydraulic system and control strategy are verified by experimental results.

Reliability is a crucial consideration in the expansion of generation or transmission in a bulk power system, especially in a power grid with a high penetration of renewables. Reliability indices, such as loss of load probability (LOLP), are generally evaluated to determine the adequacy of a bulk power system in the future. When a Monte-Carlo simulation is conducted to evaluate the LOLP, the computational time is long because chronological time-series data are involved. This work proposes a novel scenario-based method for studying the LOLP in a bulk power system with a high penetration of renewables. Scenarios are generated by aggregating Markov states of hourly loads, photovoltaic power generations and wind power generations. The power flow result in each scenario is examined to ensure power balance among demand, supply and losses, so the LOLP can be obtained. This novel scenario-based method is more efficient than the traditional chronological time-series approach because the number of considered scenarios is much smaller than the number of considered time-series cases. A set of realistic data regarding Taiwan power system, consisting of 2078 buses associated with a peak load of 39.178 GW, wind power of 6.938GW and photovoltaic power of 20 GW in 2025 is used to validate the proposed method.

For a pulsed laser transmitted in the raised dust from battlefield vehicle movement, the premise of obtaining its transmission properties and anti-jamming method is to acquire the spatial distribution rules of dust mass concentration. In this study, the emission mechanism for vehicle-raised dust in the battlefield environment is analyzed, as well as various influencing factors of the dust emission velocity. A finite volume method (FVM)-based model for simulating the spatial dispersion of dust raised by a certain moving tank is built, and the effects of mass flow rate, traveling speed, wind speed and wind direction on the spatial distribution of dust mass concentration are explored in the battlefield environment. The results show that the raised dust is plume-shaped within the computational domain, and the spatial mass concentration of plume is positively correlated with the mass flow rate. Besides, it is necessary to determine the impact of tank travelling speed on the spatial mass concentration of dust based on the relationship between the travelling speed and the dust mass flow rate. Although front wind reduces the spatial mass concentration of dust plumes, the reduction rate decelerates when the wind speed exceeds 4 m/s. Crosswind leads to the deflection and decreased mass concentration of dust plumes. Under varying wind directions, the variation of crosswind component exerts a more pronounced influence on the spatial mass concentration of dust than that of the front wind component. This study provides a technical support for the further study of near-ground laser weapon system against vehicle dust interference in the battle field.

In this paper, a two-stage approach is proposed on a joint dispatch of thermal power generation and variable resources including a storage system. Although, the dispatch of alternate energy along with conventional resources has become increasingly important in the new utility environment. However, recent studies based on the uncertainty and worst-case scenario-oriented robust optimization methodology reveal the perplexities associated with renewable energy sources (RES). First, the load demand is predicted through a convolutional neural network (CNN) by taking the ISO-NECA hourly real-time data. Then, the joint dispatch of energy and spinning reserve capacity is performed with the integration of RES and battery storage system (BSS) to satisfy the predicted load demand. In addition, the generation system is penalized with a cost factor against load not served for the amount of energy demand which is not fulfilled due to generation constraints. Meanwhile, due to ramping of thermal units, the available surplus power will be stored in the backup energy storage system considering the state of charge of the storage system. The proposed method is applied on the IEEE-standard 6-Bus system and particle swarm optimization (PSO) algorithm is used to solve the cost minimization objective function. Finally, the proposed system performance has been verified along with the reliability during two worst-case scenarios, i.e., sudden drop in power demand and a short-fall at the generation end.

Energy structures from non-conventional energy source has become highly demanded nowadays. In this way, the maximum power extraction from photovoltaic (PV) systems has attracted the attention, therefore an optimization technique is necessary to improve the performance of solar systems. This article proposes the use of ABC (artificial bee colony) algorithm for the maximum power point tracking (MPPT) of a PV system using a DC-DC converter. The procedure of the ABC MPPT algorithm is using data values from PV module, the P-V characteristic is identified and the optimal voltage is selected. Then, the MPPT strategy is applied to obtain the voltage reference for the outer PI control loop, which in turn provides the current reference to the predictive digital current programmed control. A real-time and high-speed simulator (PLECS RT Box 1) and a digital signal controller (DSC) are used to implement the hardware-in-the-loop system to obtain the results. The general system does not have a high computational cost and can be implemented in a commercial low-cost DSC (TI 28069M). The proposed MPPT strategy is compared to the conventional perturb and observe method, results show the proposed method archives a much superior performance.

The submodule capacitor voltage fluctuation of the modular multilevel converter (MMC) leads to the increase of the size of the submodule capacitor, the increase of system cost, and the decrease of system reliability and stability. In order to solve the problem, the mathematical model of circulating current voltage drop is derived, and the mechanism of circulating current on the submodule capacitor voltage fluctuation is studied through the submodule switching function. A control strategy for suppressing the submodule capacitor voltage fluctuation is proposed, which is based on circulating current voltage drop balance (CCVDB). The simulation model is built to compare three control strategies for suppressing submodule capacitor voltage fluctuation under the step changes of DC bus voltage and active power, namely, the circulating current is suppressed to zero, the second-order component of the submodule capacitor voltage is suppressed to zero, and the CCVDB strategy. The research results show that the CCVDB strategy can effectively suppress the submodule capacitor voltage fluctuation, reduce the capacitance demand, and reduce the total harmonic distortion (THD) of the submodule capacitor voltage, which verifies the feasibility and effectiveness of the CCVDB strategy.

The structural mass of a building provides inherent thermal storage capability. Through sector coupling, the building mass can provide additional flexibility to the electric power system, using, for instance, combined heat and power plants or power-to-heat. In this work, a mathematical model of building inertia thermal energy storage (BITES) for integration into optimized smart grid control is introduced. It is shown how necessary model parameters can be obtained using generalized additive modeling (GAM) based on measurable building data. For this purpose, it is demonstrated that the ceiling surface temperature can serve as a proxy for the current state of energy. This allows for real-world implementation using only temperature sensors as additionally required hardware. Compared with linear modeling, GAM enable improved modeling of the nonlinear characteristics and interactions of external factors influencing the storage operation. Two case studies demonstrate the potential of using building storage as part of a virtual power plant (VPP) for optimized smart grid control. In the first case study, BITES is compared with conventionally used hot water tanks, revealing economic benefits for both the VPP and building operator. The second case study investigates the potential for savings in CO2 emission and grid connection capacity. It shows similar benefits when using BITES compared to using battery storage, without the need for hardware investment. Given the ubiquity of buildings and the recent advances in building control systems, BITES offers great potential as an additional source of flexibility to the low-carbon energy systems of the future.

Frequency response analysis is a powerful tool for mechanical fault diagnostics in power transformers. However, interpretation schemes still today depend on expert analyses, mainly because of the complex structure of power transformers. One of the fundamental shortcomings of experimental investigations is that mechanical deformations cannot be managed on real transformers to obtain data for different scenarios because they are too destructive. To address this issue in a systematic way, the current research used a specially designed laboratory transformer model that allows mechanical defects to be introduced so its frequency response can be evaluated under different conditions. The key feature of this model is the non-destructive interchangeability of its winding sections, allowing reproducibility and repeatability of frequency response measurements. Numerical indices were compared over key performance indicators (linearity, sensitivity and monotonicity). The analysis indicated that comparative standard deviation offered promising results for evaluation of mechanical deformations on the laboratory winding model given its monotonic behaviour, sensitivity and linear increase with fault severity. Additionally, support vector machine learning, radial basis function neural network and the statistical k-nearest neighbour method were used for fault classification with different strategies and configurations. While limited data from different transformers are used in the available literature, the approach discussed here considers 371 measurements from the same transformer model. The test results are supportive and demonstrate great accuracy when machine learning is used for winding fault classification.

A hypergraph is a generalization of a graph in that the restriction of pairwise affinity scores is lifted in favor of affinity scores that can be evaluated between an arbitrary number of inputs. Hypergraphs clustering is the process of finding groups in which members of a given hypergraph exhibit a high similarity and dissimilarity with members outside their group. In this paper, we generalize the well-known MapEquation, an optimization equation used in the clustering of nonhypergraphs, for hypergraphs. We develop an agglomerative algorithm, Hypergraph Random Walks (HRW), to find an approximate solution to the generalized MapEquation. Our algorithm requires neither hyperparameter setting nor any restriction on the underlying hypergraph. We show that our algorithm has a strong theoretical performance on the newly defined ring of hyper cliques and demonstrate that our algorithm scales to hypergraphs with large edge sets.

In the modern world of technology, the cascaded DC-DC converters with multiple output configurations are contributing a dominant part in the DC distribution systems and DC micro-grids. An individual DC-DC converter of any configuration exhibits complex non-linear dynamic behavior resulting in instability. This paper presents a cascaded system with one source boost converter and three load converters including buck, Cuk, and Single-Ended Primary Inductance Converter (SEPIC) that are analyzed for the complex non-linear bifurcation phenomena. An outer voltage feedback loop along with an inner current feedback loop control strategy is used for all the sub-converters in the cascaded system. To explain the complex non-linear dynamic behavior, a discrete mapping model is developed for the proposed cascaded system and the Jacobian matrix's eigenvalues are evaluated. For the simplification of the analysis, every load converter is regarded as a fixed power load (FPL) under reasonable assumptions such as fixed frequency and input voltage. The eigenvalues of period-1 and period-2 reveal that the source boost converter undergoes period-2 orbit and chaos whereas all the load converters operate in a stable period-1 orbit. The proposed configuration eliminates the period-2 and chaotic behavior from all the load converters and is also validated using simulation in MATLAB/Simulink and experimental results.

Accuracy of solar irradiance forecasting is imperative for the effective utilization and integration of solar energy into the power system. To forecast global horizontal irradiance based on a multivariate meteorological data; this study first evaluates five standalone models, including recurrent deterministic policy gradient (RDPG), long short term memory (LSTM) neural network, extreme gradient boosting (XGB), Gaussian process regression (GPR), and support vector regression (SVR). The RDPG model outperforms the standalone counterparts by demonstrating 2.485 W/m2, 20.591 W/m2, 18.316 W/m2, and 23.176 W/m2 reductions in RMSE compared to the LSTM, XGB, GPR, and SVR models. Afterward, the performance of the RDPG model is further enhanced, by developing two-stage hybrid models, including improved complete ensemble empirical mode decomposition with additive noise-RDPG (ICEEMDAN-RDPG) and variational mode decomposition-RDPG (VMD-RDPG). The subsequent construction of the hybrid model ICEEMDAN-SE-VMD-RDPG (ISVR) results in the further improvement of the two-stage hybrid models. The ISVR surpasses all the established models including VMD-RDPG, ICEEMDAN-RDPG, RDPG, LSTM, XGB, GPR, and SVR by displaying 18.401 W/m2, 13.908 W/m2, 33.223 W/m2, 53.111 W/m2, 67.704 W/m2, 67.502 W/m2 and 69.943 W/m2, respectively, decrease in RMSE. The overall results are enlightening and reveal the extension feasibility of the IVSR model for the other forecasting tasks, e.g., renewable energy integration and electrical load forecasting.

Detecting electricity price spikes is crucial as it helps market participants to gain confidence and formulate appropriate strategy to maximize their benefits. In this paper, a multi-feature based approach with the incorporation of variable thresholds is developed to detect electricity price spikes in the national electricity market of Australia. The variable thresholds, which are determined using a weighted sliding window average and an adjusted standard deviation, help to segregate spikes from normal price variations. Also, significant features are extracted from the market after thoroughly analyzing the underlying causes resulting into the price spikes. These features are employed as inputs to a support vector machine to classify electricity prices as spikes or non-spikes. A case study is conducted using a dataset acquired from the state of New South Wales, Australia. The results show that the proposed method can successfully detect the price spikes with high accuracy and confidence.

Networked microgrids (MGs) have several advantages over individual MGs such as reliability improvement and cost reduction. To promote the mutual connection of individual MGs, a rational and predictable profit-sharing rule is required. This study investigates a rule for the fair distribution of profit in networked MGs according to their contributions that come from connecting between them. Cooperative game theory defines profit-sharing problems such as the Nash bargaining solution (NBS) and Shapley value. However, as the two solution concepts are used assuming that the network is complete, they do not account for the positional contribution of each MG in a given network. We propose a variation of the Shapley value designed for an incomplete network, the Myerson value. We investigate how Myerson value-based profit-sharing rule can account for both the role and positional contributions of each MG. Using Korean data, we compare the profit distribution results for the three sharing rules (the NBS, Shapley value, and Myerson value). The result confirms that the proposed rule fairly distributes the profit according to one’s contribution, even when MGs are incompletely connected.

The wireless channel-based Secret Key Generation (SKG) algorithms aim at securing the wireless link against unauthorized eavesdropping by exploiting the channel’s randomness for generating matching secret keys at the legitimate nodes for message encryption/decryption. To counter differences in hardware and noise conditions at the legitimate nodes, which can lead to key mismatch, the SKG algorithms typically include the intermediate steps of sampling, quantization, information reconciliation, and privacy amplification. These steps collectively aim to improve the performance trade-offs between Key Generation Rate (KGR), Key Agreement Probability (KAP), and Secret Key Randomness (SKR) properties. This paper derives a closed-form expression for the Average Contiguous Duration (ACD) of Generalized Gamma (GG) fading wireless channels. The ACD is a recently introduced novel quantifier for characterizing the second-order statistics of fading channels, which includes Average Fade Duration (AFD) as its special case. The proposed GG fading ACD expression is shown to include, as its special cases, the ACD for commonly observed fading distributions such as Gamma, Nakagami- $m$ , and Rayleigh. By exploiting the derived GG ACD expression, a multi-level quantization scheme for SKG is proposed that determines suitable quantization intervals for identical likelihood of an equal number of consecutive channel samples falling in each quantization interval. A comprehensive comparative analysis of the proposed ACD-based quantization for SKG is conducted in relation to conventional Uniform Quantization (UQ) and Cumulative Distribution Function (CDF)-based Non-Uniform Quantization (NUQ) schemes. The presented numerical results confirm the superior performance trade-off between KGR and KAP offered by the proposed ACD-based quantization in relation to that offered by UQ and CDF-based NUQ.

According to the principles of bionics and predatory behavior of dolphins, Leader Dolphin Herd Algorithm (LDHA) was designed, whose performance is also tested on the basis of heuristic algorithm, showing that LDHA is obviously superior to the other algorithms in calculation accuracy and efficiency. The minimum safety factor of sliding surface was taken as the objective function, the recurrence formula of safety factor based on simplified Bishop Method was regarded as constraint condition, a non-linear multi-objective optimization mathematical model used for analyzing slope stability was established. the process of searching slope sliding surface and solving safety factor were designed on account of LDHA. By taking homogeneous slope and layered slope as calculation examples compared to other intelligent algorithms, the results show that the calculation result of LDHA is close to that of Genetic Algorithm and Simulated Annealing Algorithm and improved Genetic Algorithm. Moreover, the LDHA has less iterations and faster convergence rate, and the result is closer to the optimal solution, which indicates the feasibility and practicability of LDHA in solving the minimum stability safety factor.

A severe energy crisis has plagued Yemen for decades, and most of the population lack access to electricity. This has harmed the country’s economic, social, and industrial growth. Yemen generates electricity mainly from fossil fuels, despite having a high potential for renewable energy. Unfortunately, the situation has recently been compounded by the country’s continuing war, which has been ongoing since early 2015. It has impacted the country’s energy infrastructure negatively, resulting in power outages. Therefore, this paper aims to provide an updated perspective on Yemen’s current energy crisis and explain its key issues and potential solutions. Besides, it examines the potential, development, and current state of renewable energy sources, such as solar, wind, geothermal, and biomass. Based on the findings, Yemen is one of the world’s wealthiest countries in terms of sunlight and wind speed, and these two resources are abundant in all regions of the country. In addition, this paper sheds light on the solar energy revolution that has arisen since the war started due to the complete outage of the national electricity. Within a few years, solar energy in Yemen has increased its capacity by 50 times and has recently become the primary source of electricity for most Yemenis. Furthermore, the paper discusses the difficulties and challenges that face the implementation of renewable energy investment projects. Numerous recommendations for potential improvements in Yemen’s widespread use of renewable energy are also provided in this paper. All of the ideas presented in this paper are hoped to increase the efforts to grow renewable energy production in Yemen, thereby solving the issues of energy poverty and reducing environmental effects. The presented analysis can be used as a scientific reference for researchers and industrial companies looking for suitable solutions to advance Yemen’s renewable energy.

The modular and voltage scalable features of modular structure based power converter has attracted for wide power range applications. However, balancing the capacitor voltage becomes complex when the number of sub modules is increased. In a conventional system, for balancing a capacitor voltage, separate monitoring and control are required for each sub module, which further increases the cost and burden to the converter. This paper proposes a simplified logical method (SLM) for balancing the capacitor voltage of each sub module without any additional sensors and controllers. The capacitor voltage of N sub modules is balanced based on the simple logical function, which eventually balances the capacitor voltage and thus reduces the voltage ripple of each capacitor. In addition, this method suppresses the circulating current and common mode voltage, which increases the efficiency and cost effective of the converter. The level shifted carrier technique is used for providing desired pulse signals for the switching devices in the converter. The effective performances of the proposed method are validated using simulation and experimental setup.

DC subgrids consisting of modern active loads (ALs) and local dc distributed generation (DG) units are normally interfaced with the main ac grid by utilizing bidirectional voltage source converters (VSCs). Under the very weak grid (VWG) conditions, the integration of voltage-oriented controlled (VOC) VSCs in the inversion mode becomes very challenging and therefore undamped oscillations in the power and angle responses are yielded. Most of the existing works address this issue for VSCs in the inversion mode of operation. However, integration of VSCs in the rectification mode with the consideration of the outer loop controllers into the VWGs has not been reported. To fill up this gap, a state-space model of the bidirectional VSC-to-weak grid (VSC-WG) system is developed in this work with an emphasis on the rectification mode of operation. A modal-sensitivity analysis is then utilized to evaluate small-signal stability of the system, identify the dominant modes, and investigate the system states that have a major influence on these modes. The results reveal two pairs of unstable complex modes that are correlated with the dynamic interaction between the VOC-based VSCs and the VWG impedance. It is also shown that the stability margin of VSCs in the rectification mode is less than that of the inversion mode under the same VWG conditions. To enhance the integration of the VSCs in the rectification mode, a dual-active compensation (DAC) scheme is proposed to mitigate the instabilities under VWG conditions. Several time-domain simulation results are presented to verify the validity of the small-signal model and demonstrate the effectiveness of the DAC scheme under the VWG conditions. Finally, hardware-in-the-loop (HIL) real-time experimental results are presented to validate the simulation results.

Farmers around the world face the challenge of growing more food for the increasing world population. On top of that, external threats such as pests (weeds and insects) pose a threat to crop production and it is necessary to take early steps to reduce the risk. This paper presents semantic segmentation of canola field images collected under natural conditions. The dataset contains four unbalanced classes; background, crop, weeds, and damages in the crop. The damages to the crop leaves are small round shaped and share the same texture and colour as whitish stones from the background. We propose, DeepVeg, a deep learning segmentation model that focuses on the smallest (damage) class without affecting other classes to solve the class imbalance issue. Early stage canola field image dataset is utilized for training and testing the proposed model. Evaluation results show that the proposed method outperforms the benchmark deep learning models and effectively addresses the weed and damaged canola plants segmentation problem. The DeepVeg model demonstrates a superior mean intersection over union score greater than 0.76 and $accuracy$ above 0.97 for four class segmentation. The model also shows robustness in detecting unlabelled, newly grown weeds and canola and is also able to distinguish the similar rounded structured canola plant and weed with small amounts of data for model training, which is suitable for early stage damage and weed segmentation.

Model predictive control (MPC), manly based on a direct use of an explicit and identifiable model, has been widely used in controller design in different applications both by academia and industry. The reason for such popularity is due to its strong ability for providing high performance electric drive systems, as highly recognized as the most reliable control approach compared with field-oriented control (FOC) and direct torque control (DTC). In general, the MPC has numerous features and advantages, such as direct switching states to the converter without any modulation, online optimization with multivariable control, low current total harmonic distortion, low switching loss, etc. The aim of this paper is to provide a comprehensive review for major development and achievements of the recent progress on the MPC for electrical machines and drives. This review begins with the innovative topologies and operating principles of fundamental MPC, and ends to summary on different advanced MPC algorithms. Typical MPC techniques have been fully adopted to enhance the drive performance of the electrical drives, mainly including finite-set model predictive control (FS-MPC) based on tuning weighting factors, without weighting factors, maximum torque per ampere, low number of switching vectors, and multi voltage vectors in one sample period. Finally, great attention has been paid to the discussion of the new trends and future research topics.

The installed capacity of renewable energy has continued to increase in recent years, causing excess power to be transmitted to neighboring provinces by DC transmission with wind and thermal power bundling. However, with increasing coal cost and decreasing investment benefit of thermal power, the economic benefit of thermal power investment has decreased. Therefore, this paper proposes a method that replaces the thermal power unit with renewable energy to assume the peak regulating function of the power system. First, a novel heat storage strategy for concentrating solar power (CSP) stations is established. On this basis, the CSP station is included in the wind-thermal bundling model to replace the peak regulating function of some thermal power units, and the planning method of wind-thermal-CSP bundling with an adjustable heat storage strategy is developed. By adjusting the heat storage strategy and comparing the economic parameters under the different strategies, the optimal ratio of the installation of different powers is obtained.

This study aims to evaluate the innovation adoption performance of the renewable energy alternatives. Within this context, a new model is created that consists of two different stages. Firstly, the decision combinations of innovation adoption life cycle patterns are identified by considering integer code series. On the other side, in the second stage, the innovation adoption life cycle performance of the renewable energy alternatives is ranked. In this framework, interval type-2 (IT2) fuzzy technique for order preference by similarity to ideal solution (TOPSIS) methodology is taken into consideration based on alpha cut levels. Moreover, a comparative evaluation is also conducted by considering IT2 fuzzy VIseKriterijumska Optimizacija I Kompromisno Resenje (VIKOR). It is identified that the analysis results are quite coherent and reliable. The findings demonstrate that solar energy is the most appropriate renewable energy type to make innovation. Additionally, wind and geothermal energies are also other significant renewable energy types that have a great innovation potential. However, it is also concluded that biomass and hydroelectric energy have lower importance in comparison with the others. Another important point is that the ranking results are quite similar for different alpha cuts. This condition indicates that the analysis results are coherent and reliable. It is obvious that making investment to solar energy alternatives provides opportunity to adopt technologic innovation more easily. Therefore, it is recommended that investors should give priority to solar energy projects so that it can be more possible to increase the efficiency and the profitability of the investments.

In this paper, a self-learning multi-class intelligent model for wind turbine fault diagnosis is proposed by using MFQL (Modified-Fuzzy-Q-Learning) technique. The MFQL is adaptive in nature and extension of fuzzy-Q-learning method where look-up table of Q-learning is conquered by fuzzy based approximation strategy to reduce the curse of dimensionality of the Q-learning. The proposed MFQL classifier diagnoses the mechanical and imbalance faults without using mechanical sensors. Proposed methodology is addressed with relying on PMSG (Permanent Magnet Synchronous Generator) stator current signals, which is already being used by protection system of wind turbines. According to the aforementioned description, non-stationary current signals of PMSG have been pre-processed to extract the input features by empirical mode decomposition followed with J48 algorithm based most relevant input feature selection. For the one-step ahead performance demonstration of the proposed MFQL approach, results have been compared with neural network, support vector machines, fuzzy logic, and conventional Fuzzy-Q-Learning techniques. Demonstrated results outperform the capability of proposed MFQL approach. Moreover, MFQL is developed first time to implement in the area of WTGS fault diagnosis in the literature.

With the deteriorating effects resulting from global warming in many areas, geographically distributed data centers contribute greatly to carbon emissions, because the major energy supply is fossil fuels. Considering this issue, many geographically distributed data centers are attempting to use clean energy as their energy supply, such as fuel cells and renewable energy sources. However, not all workloads can be powered by a single power sources, since different workloads exhibit different characteristics. In this paper, we propose a fine-grained heterogeneous power distribution model with an objective of minimizing the total energy costs and the sum of the energy gap generated by the geographically distributed data centers powered by multiple types of energy resources. In order to achieve these two goals, we design a two-stage online algorithm to leverage the power supply of each energy source. In each time slot, we also consider a chance-constraint problem and use the Bernstein approximation to solve the problem. Finally, simulation results based on real-world traces illustrate that the proposed algorithm can achieve satisfactory performance.

Low frequency oscillation (LFO) in the electric multiple units (EMUs)-traction network cascade system (ETNCS) can lead to traction blockade and abnormal operation. Using the Chinese CRH3 EMUs as an example, a new LFO suppressing method by applying the energy storage (ES) device on EMUs is proposed in this paper and the ES is connected to the DC-link on board. The equivalent model of the ETNCS is established at first. The critical condition of LFO in the ETNCS is analyzed via impedance ratio criterion and Bode diagram. Then, the proposed on-board bidirectional DC-DC converter integrated with the ES system is controlled via a virtual DC machine (VDCM) strategy, and the VDCM small signal model derivation and the parameter design are carried out. Using the Bode diagram of the impedance ratio, the system stability with the ES device controlled by the proposed VDCM scheme is analyzed again; simulation results verify the effectiveness of the control algorithm to suppress the LFO adaptively. Compared with the traditional PI controller, the system steady-state and dynamic performance can be improved with the proposed method, including decreased overshoot, fluctuation, response time, and increased gain margin of the DC-link voltage.

The multi-region integrated energy system (MRIES) is one promising option for exploiting the complementarity potentials between power system, thermal system and loads, as well as improving flexibility and economic competitiveness. However, the optimal planning has remained challenging since it involves multiple forms of energies with the nature of coupling. A mixed integer linear programming (MILP) model is proposed for joint optimization of MRIES planning and operation, which considers the flexible load’s demand response (DR). It is realized by combining the multi-region heating network IES coupling model with the flexible cooling-heating-electric load DR model. The results show that the proposed model can assist planners to comprehensively analyze and evaluate the impacts of multiple factors on planning, and to determine the optimal planning configuration and operation strategy.

In this article, a hybrid Artificial Neural Network - Newton Raphson (ANN-NR) is introduced to mitigate the undesired lower-order harmonic content in the cascaded H-Bridge multilevel inverter for solar photovoltaic (PV). Harmonics are extracted by the excellent choice of opting switching angles by exploiting the Selective Harmonic Elimination (SHE) PWM technique accompanying a unified algorithm in order to optimize and reduce the Total Harmonic Distortion (THD). ANN is trained with optimum switching angles, and the estimates generated by the ANN are the initial guess for NR. In this study, the CHB-MLI is combined with a traditional boost converter, it boosts the PV voltage to a superior dc-link voltage Perturb and Observe (P&O) based Maximum Power Point Tracking (MPPT) algorithm is used for getting a stable output and efficient operation of solar PV. The proposed system is proved over an eleven-level H-bridge inverter, the work is carried out in MATLAB/Simulink environment, and the respective results are confirmed that the proposed technique is efficient, and offers an actual firing angles with a few iterations results in a better capability of confronting local optima values. The suggested algorithm is justified by the experimental development of eleven-level cascaded H-bridge inverter.

Electrical discharge is a sign of insulation damage of power equipment and an essential part of power inspection. For the recognition of plasma discharge patterns, the accuracy rate of traditional manual inspection relies heavily on the experience of the staff, and misjudgment is inevitable. The emerging method based on manual feature extraction of discharge signals also has its limitations. Thus, a high-accurate recognition method based on convolutional neural network (CNN) and visible images is proposed in this paper. Under the deep learning framework-TensorFlow, CNN model is established and pre-trained weights are loaded through transfer learning. The visible image dataset used for training and evaluation includes four discharge patterns: arc discharge, corona discharge, creeping discharge and plasma jet. In data processing, data amplification and data enhancement are used to expand the sample. Besides, typical network models with fully connected layers are adopted and compared. Three optimization methods of ReLU, BN and Dropout are applied to solve the overfitting problem, and the corresponding effects are discussed in the model of VGG16 and ResNet152. The test accuracy of the optimized VGG16 reaches 99.6% without overfitting, which indicates the promising potential in the recognition of plasma discharge patterns in power inspection.