Abstract:In this paper, an optimal nonlinear robust sliding mode control (ONRSMC) based on mixed H2/H∞ linear matrix inequalities (LMIs) is designed for the excitation system in a “one machine-infinite bus system” (OMIBS) to enhance system stability. Initially, the direct feedback linearization method is used to establish a mathematical model of the OMIBS incorporating uncertainties. ONRSMC is then designed for this model, employing the mixed H2/H∞ LMIs. The chaos mapping-based adaptive salp swarm algorithm (CASSA) is introduced to fully optimize the parameters of the sliding mode control, ensuring optimal performance under a specified condition. CASSA demonstrates rapid convergence and reduced likelihood of falling into local optima during optimization. Finally, ONRSMC is obtained through inverse transformation, exhibiting the advantages of simple structure, high reliability, and independence from the accuracy of system models. Four simulation scenarios are employed to validate the effectiveness and robustness of ONRSMC, including mechanical power variation, generator three-phase short circuit, transmission line short circuit, and generator parameter uncertainty. The results indicate that ONRSMC achieves optimal dynamic performance in various operating conditions, facilitating the stable operation of power systems following faults.

Abstract:Relay coordination is crucial in electrical power systems to protect against malfunctions and damage caused by unexpected events like short circuits. To address the challenge associated with the reverse direction of fault current, dual-setting (DS) directional over-current relays have evolved but failed to provide proper coordination during changes of load, generation, and network. In the meantime, with the increasing number of DS relays, the total relay operating time tends to saturate. Therefore, this paper proposes a protection scheme based on the optimal deployment of conventional and dual-setting rate of change of voltage (DS-ROCOV) relays in distribution systems. This holds true for varying network topologies and is unaffected by variations in load and generation. The objective of the proposed scheme is to ensure reliable and efficient protection against faults in distribution systems by minimizing the overall operating time with the optimal number of DS-ROCOV relays. The proposed protection scheme's performance is evaluated for different coordination time interval values as well as in different microgrid scenarios. This paper outlines the design and implementation of the proposed protection scheme which is validated on the modified IEEE 14-bus system using simulations in Matlab/Simulink.

Abstract:The increasing use of renewable energy in the power system results in strong stochastic disturbances and degrades the control performance of the distributed power grids. In this paper, a novel multi-agent collaborative reinforcement learning algorithm is proposed with automatic optimization, namely, Dyna-DQL, to quickly achieve an optimal coordination solution for the multi-area distributed power grids. The proposed Dyna framework is combined with double Q-learning to collect and store the environmental samples. This can iteratively update the agents through buffer replay and real-time data. Thus the environmental data can be fully used to enhance the learning speed of the agents. This mitigates the negative impact of heavy stochastic disturbances caused by the integration of renewable energy on the control performance. Simulations are conducted on two different models to validate the effectiveness of the proposed algorithm. The results demonstrate that the proposed Dyna-DQL algorithm exhibits superior stability and robustness compared to other reinforcement learning algorithms.

Abstract:Safety and reliability are crucial for the next-generation supercapacitors used in energy storage systems, while accurate prediction of the degradation trajectory and remaining useful life (RUL) is essential for analyzing degradation and evaluating performance in energy storage systems. This study proposes a novel data processing and improved one-dimensional convolutional neural network (1D CNN)-informer framework for robust RUL prediction. In data preprocessing, all data from two structures are adjusted to a unified format, and cross-entropy loss is used to couple the 1D CNN and informer. Then, the minimum-maximum feature scaling method is used for normalization to accelerate the training process in reaching the minimum cost function. A relative position encoding algorithm is introduced to improve the Informer model, enabling it to better learn the sequence relationships between data and effectively reduce prediction variability. Supercapacitor data in different working conditions are used to validate the proposed method. Compared with other existing methods, the maximum root mean square error is reduced by 32.71%, the mean absolute error is reduced by 28.50%, and R2 is increased by 4.79%. The strategy considers the complementarity between two single models, which can extract features and enrich local details, as well as enhance the model's global perception ability. The experimental results demonstrate that the proposed model achieves high-precision and robust RUL prediction, thereby promoting the industrial application of supercapacitors.

Abstract:Accurate topological information is crucial in supporting the coordinated operational requirements of source-load-storage in low-voltage distribution networks. Comprehensive coverage of smart meters provides a database for low-voltage topology identification (LVTI). However, because of electricity theft, power line communication crosstalk, and interruption of communication, the measurement data may be distorted. This can seriously affect the performance of LVTI methods. Thus, this paper defines hidden errors and proposes an LVTI method based on layer-by-layer stepwise regression. In the first step, a multi-linear regression model is developed for consumer-branch connectivity identification based on the energy conservation principle. In the second step, a significance factor based on the t-test is proposed to modify the identification results by considering the hidden errors. In the third step, the regression model and significance threshold parameters are iteratively updated layer by layer to improve the recall rate of the final identification results. Finally, simulations of a test system with 63 users are carried out, and the practical application results show that the proposed method can guarantee over 90% precision under the influence of hidden errors.

Abstract:The urban power grid (UPG) combines transmission and distribution networks. Past studies on UPG congestion mitigation have primarily focused on relieving local congestion while ignoring large-scale energy transfer with safety margins and load balancing. This situation is expected to worsen with the proliferation of renewable energy and electric vehicles. In this paper, a two-layer congestion mitigation framework is proposed, one which considers the congestion of the UPG with flexible topologies. In the upper-layer, the particle swarm optimization algorithm is employed to optimize the power supply distribution (PSD) of substation transformers. This is known as the upper-layer PSD. The lower-layer model recalculates the new PSD, known as the lower-layer PSD, based on the topology candidates. A candidate topology is at an optimum when the Euclidean distance mismatch between the upper-and lower-layer PSDs is the smallest. This optimum topology is tested by standard power flow to ascertain its feasibility. The optimum transitioning sequence between the initial and optimum topologies is also determined by the two-layer framework to minimize voltage deviation and line overloading of the UPG considering dynamic thermal rating. The proposed framework is tested on a 56-node test system. Results show that the proposed framework can significantly reduce congestion, maintain safety margins, and determine the optimum transitioning sequence.

Abstract:This paper proposes a novel fault diagnosis method by fusing the information from multi-sensor signals to improve the reliability of the conventional vibration-based wind turbine drivetrain gearbox fault diagnosis methods. The method fully extracts fault features for variable speed, insufficient samples, and strong noise scenarios that may occur in the actual operation of a wind turbine planetary gearbox. First, multiple sensor signals are added to the diagnostic model, and multiple stacked denoising auto-encoders are designed and improved to extract the fault information. Then, a cycle reservoir with regular jumps is introduced to fuse multidimensional fault information and output diagnostic results in response to the insufficient ability to process fused information by the conventional Softmax classifier. In addition, the competitive swarm optimizer algorithm is introduced to address the challenge of obtaining the optimal combination of parameters in the network. Finally, the validation results show that the proposed method can increase fault diagnostic accuracy and improve robustness.

Abstract:To clarify the electromagnetic, vibration, and loss characteristics of the internal components of a converter transformer under DC bias conditions and their influencing mechanisms, a series of studies are conducted using the finite element method and model experiments. This paper quantifies the influence of different DC contents on the magnetic flux density, force, and displacement distribution characteristics of the iron core and winding and analyzes the internal relationship between various indicators. The inflection point of the DC bias coefficient on the vibration is obtained and the contribution mechanism of the different responses of the iron core and winding to this inflection point is explained. The value of the DC bias coefficient for changing the main vibration frequency is determined. When the DC bias coefficient is 1.0 and 1.5, the main frequency of vibration moves to the right to 250 Hz and 350 Hz. Based on the principle of similarity, a DC bias vibration experimental platform for converter transformers is developed, and DC bias magnetic experiments are conducted to verify the reliability of the simulation results.

Abstract:To ensure system stability, the fixed-PID (F-PID) controller with small parameters is usually adopted in hydropower stations. This involves a slow setting speed and it is difficult to realize optimal control for full working conditions. To address the problem, this paper designs a variable-PID (V-PID) controller for a hydraulic turbine regulation system (HTRS) based on the improved grey wolf optimizer (INGWO) and back propagation neural networks (BPNN). These can achieve excellent regulation under full working conditions. First, the nonlinear HTRS model containing the nonlinear hydro-turbine model is constructed and the stable domain is obtained using Hopf bifurcation theory to determine the available range of PID parameters. The optimal PID parameters in typical working conditions are then calculated by the INGWO, and the optimal PID parameters are generalized through training the V-PID neural networks which take the optimal PID parameters as sample data. The V-PID neural networks with different structures are compared to determine the optimal structure of the variable-PID controller model. The V-PID controller-based nonlinear HTRS model shows that the PID parameters can be automatically adjusted online according to the working condition changes, realizing optimal control of hydropower units in full working conditions.

Abstract:In order to better realize the energy recovery and storage of hybrid EVs (HEVs), a switched reluctance starter/generator (SRS/G) with both starting and power generation functions is investigated in this paper. First, the iron loss of SRS/G is mainly studied to reduce the motor loss and improve the power generation efficiency. Then, the energy storage of hybrid EVs can be effectively improved. Secondly, a magnetic flux density (MFD) wave-forms solution method is proposed to solve the difficulty in calculating the iron loss of the SRS/G. Compared with the commonly used finite element method, the proposed solution method has the advantages of simple, fast and small computational amount. Meanwhile, considering the different operating conditions of SRS/G, the iron loss models for both the time-domain and frequency-domain are established. In addition, the calculation formula of the variable coefficient Bertotti three-term loss separation is improved. As the hysteresis loss coefficient, the Steinmetz coefficient and the stray loss coefficient are respectively fitted by the Fourier fitting method. This method is also applied to solve the iron loss of SRS/G. Finally, through an experimental verification, it is indicated that the development of proposed method has high accuracy.

Abstract:This paper proposes an intelligent hybrid method for reducing harmonics to enhance power quality in a distribution system based on renewable energy sources. The proposed intelligent method, namely, the AOS-FAT technique, consolidates atomic orbital search (AOS) and a feedback artificial tree (FAT). The main objective of the proposed approach is to improve the quality of power by mitigating the harmonics. The AOS method is used to find the best values for basic and harmonic loop settings, like the shunt active power filter's direct current, voltage and the voltage at the terminals. Based on the change in load and PV parameters, a dataset variation is generated based on the objective function for minimum error. The optimal control signals are then generated using the FAT approach, which predicts the optimal parameters from the accomplished datasets. The proposed approach mitigates the overall harmonic distortion through the switching control pulses to enhance power quality. The control method concentrates on improving the maximum PV power when there is harmonic distortion by inserting the exact compensation current via the hybrid shunt active power filter. The proposed approach is implemented in MATLAB, and its performance is examined by comparing to existing methods. From the simulation outcome, the maximum PV power is 12 kW, and the THD is 1.1%.