The secure transmission in the multiple-input-single-output simultaneous wireless information and power transfer (SWIPT) system is an important issue. Considering channel uncertainties, we investigate the robust secure transmission scheme, which maximizes the worst-case secrecy rate under transmit power constraint and energy-harvesting (EH) constraint. The optimization problem is a nonconvex problem. Omitting the rank-one constraint on transmit covariance, it is transformed into a solvable semidefinite program (SDP) where the rank relaxation performance upper bound is obtained. Since the obtained rank relaxation transmit covariance may not be rank one, we propose a lower bound-based rank-one suboptimal (LB-Sub) solution by employing Charnes-Cooper transformation. We also propose the suboptimal Gaussian randomized (GR) solutions based on the rank relaxation upper bound and the lower bound, respectively. Simulation results have shown that our proposed LB-Sub solution and suboptimal GR solution based on the rank relaxation upper bound outperform the nonrobust scheme.

Considering worst-case channel uncertainties, we investigate the robust secure beamforming design problem in multiple-input-single-output (MISO) full-duplex, two-way secure communications. Our objective is to maximize worst-case sum secrecy rate under weak secrecy conditions and individual transmit power constraints. Since the objective function of the optimization problem includes both convex and concave terms, we propose to transform convex terms into linear terms. We decouple the problem into four optimization problems and employ the alternating optimization algorithm to obtain the locally optimal solution. Simulation results demonstrate that our proposed robust secure beamforming scheme outperforms the nonrobust one. It is also found that when the regions of channel uncertainties and the individual transmit power constraints are sufficiently large, because of self-interference, the proposed two-way robust secure communication is proactively degraded to one-way communication.

With the continuous development of the smart grid, details of residential users’ electricity consumption have received lots of attention. In recent years, a new concept of nonintrusive load monitoring (NILM) based on graph signal processing (NILM-GSP) has been proposed. NILM allows researchers to access the electricity consumption profile of every household appliance (AP) online, which greatly reduced the cost of grid operation. This article proposes a new NILM based on alternating optimization (NILM-AO) to solve the load disaggregation problem. Since the power consumption of residential users is basically stable in its operation cycle, a power consumption constraint is constructed to make NILM realistic. In order to improve real-time processing capability, we propose a statistical downsampling method (NILM-AODM) to find the optimal sampling rate. Simulation results show that NILM-AO is more accurate than NILM-GSP, and the sampling rate of NILM-AODM is also smaller than that of NILM-AO, which leads to less communication load and time delay.

This paper focus on the application of Physics Informed Neural Network (PINN) for extracting parameters of photovoltaic (PV), wind, and energy storage equipment models. Accurately extracting the parameters of these models is essential for effectively controlling and optimizing the overall stability of Chongqing power system (CPS). Despite numerous algorithms proposed to tackle this issue, accurately and reliably extracting the parameters of these remains a significant challenge. This paper proposed an improved PINN, named Uniform Physics Informed Neural Network (UPINN), with Proximal Policy Optimization (PPO) based reinforcement learning, for extortion of parameters of these models. The PINN difficulty is overcome in UPINN by configuring four strategies: feedback operator, GRU gating mechanisms, transfer operator with historic population, and modification factor with PPO aided reinforcement learning. UPINN models are trained iteratively to maximize parameters and reduce RMSE. UPINN accurately extracts parameters and describes the behavior of PV, wind, and energy storage equipment models as it converges towards optimal solutions through parameter adjustments and RMSE evaluations. The UPINN was implemented for real-time voltage stability monitoring of CPS. The results show that UPINN performs better than other neural network models in respect of accuracy and stability, demonstrating the effectiveness of improved strategies. Moreover, its emphasis the importance of computed and estimated indices obtained through UPINN for predicting voltage collapse occurrences within the system.

Each substation is critically essential to the overall operation of the electrical power system. Potential dangers include thermal stress, noise, slip, trip, fall hazards, animal waste, and nonionizing radiation. These are the causes of joint failures of cables and overhead lines, failure of one or more phases of circuit breakers, and melting of fuses or conductors in one or more phases. These kinds of failures bring about a decline in the substation’s level of dependability. On the consumer side, power cannot be received adequately because there are losses in the transmission line. To accomplish the objective of enhancing the voltage profile, the DG must be optimized. The substation’s transient analysis utilizing a variety of factors, a study of faults and transients that occur in the substation and their effects using ETAP, and an optimization of a range of parameters using artificial intelligence techniques are all used for this goal. This paper offers the complete simulation of a 500kv substation. The simulation uses advanced software Electrical Transient Analyzer Program (ETAP) with detailed load flow analysis and short circuit study of the 500 kV substation system using ETAP software. From the ETAP-generated load flow details and the short circuit details, which are studied by varying loads or other parameters, these whole simulations are carried out multiple times using real-time data from the past eighteen months. A simulation data set contains data on both standard and different faulty conditions. In the 1st step, the normal and faulty conditions are classified. In the 2nd step, the reasons for fault occurrence include line-to-line, line-to-ground, and double line-to-ground using the Artificial Intelligence technique. In both steps, Catboost performs well, followed by Support Vector Machine and Logistic Regression. In the first step, Catboost classifies normal and faulty conditions with an accuracy of 98%, SVM is 96%, and Logistic regression is 93%. Again, in the 2nd step to identify different faulty conditions, the accuracies of Catboost SVM and Logistic Regression are 97%, 95%, and 92%, respectively.