This paper presents a novel approach for Voltage Stability Margin (VSM) estimation that combines a Kernel Extreme Learning Machine (KELM) with a Mean-Variance Mapping Optimization (MVMO) algorithm. Since the performance of a KELM depends on a proper parameter selection, the MVMO is used to optimize such task. In the proposed MVMO-KELM model the inputs and output are the magnitudes of voltage phasors and the VSM index, respectively. A Monte Carlo simulation was implemented to build a data base for the training and validation of the model. The data base considers different operative scenarios for three type of customers (residential commercial and industrial) as well as N-1 contingencies. The proposed MVMO-KELM model was validated with the IEEE 39 bus power system comparing its performance with a support vector machine (SVM) and an Artificial Neural Network (ANN) approach. Results evidenced a better performance of the proposed MVMO-KELM model when compared to such techniques. Furthermore, the higher robustness of the MVMO-KELM was also evidenced when considering noise in the input data.

Transient stability (TS) and short-term voltage stability (STVS) assessment are of fundamental importance for the operation security of power systems. Both phenomena can be mutually influenced in weak power systems due to the proliferation of power electronic interface devices and the phase-out of conventional heavy machines (e.g., thermal power plants). There is little research on the assessment of both types of stability together, despite the fact that they develop over the same short-term period, and that they can have a major influence on the overall transient performance driven by large electrical disturbances (e.g., short circuits). This work addresses this open research challenge by proposing a methodology for the joint assessment of TS and STVS. The methodology aims at estimating the resulting short-term stability state (STSS) in stable, or unstable conditions, following critical events, such as the synchronism loss of synchronous generators (SG) or the stalling of induction motors (IM). The estimations capture the mechanisms responsible for the degradations of TS and STVS, respectively. The paper overviews the off-line design of the data-driven STSS classification methodology, which supports the design and training of a hybrid deep neural network RCNN (recurrent convolutional neural network). The RCNN can automatically capture spatial and temporal features from the power system through a time series of selected physical variables, which results in a high estimation degree for STSS in real-time applications. The methodology is tested on the New England 39-bus system, where the results demonstrate the superiority of the proposed methodology over other traditional and deep learning-based methodologies. For reference purposes, the numerical tests also illustrate the classification performance in special situations, when the training is performed by exclusively using measurements from generation and motor load buses, which constitute locations where the investigated stability can be observed.

This paper investigates the impact of increased penetration of large-scale photovoltaic (PV) generation on the short-term stability of the IEEE New England 39-bus test system, considering the stochastic nature of power systems. Based on the probabilistic models of input parameters, such as load variation and the occurrence of contingencies, Monte Carlo-type simulations are performed to iteratively evaluate the system time domain response. Also, different strategies for voltage regulation capabilities for PV systems were analyzed. The simulation results identified both, improvement and detrimental effects on short-term stability with different penetration levels of PV generation. The beneficial or detrimental effects depend on the PV systems operating mode, whether or not including voltage regulation. Greater instability conditions are reported if Full Replacing (FR) of Conventional Generating Units (CGUs) is considered. However, fewer instability conditions are obtained if Partial Replacing (PR) is considered in lieu of FR since CGU keep excitation control systems (AVR, exciter) and turbine governor controls activated.

This paper presents a novel approach for determining post-contingency dynamic vulnerability regions (DVRs), oriented to assess vulnerability in real time as part of Smart Grid applications. Based on the probabilistic models of input parameters, such as load variation and the occurrence of contingencies, Monte Carlo-type simulation is performed to iteratively evaluate the system time domain responses. The dynamic probabilistic attributes are then analyzed using time series data mining techniques, namely Multichannel Singular Spectrum Analysis (MSSA), and Principal Component Analysis (PCA), in order to recognize the system DVRs based on the patterns associated to three different short-term stability phenomena. The vulnerability criterion consists in the possibility of some N-1 contingencies driving the system to further undesirable events (i.e. N-2 contingencies), which could be considered as the beginning of a cascading event. The proposal is tested on the IEEE New England 39-bus test system. Results show the feasibility of the methodology in finding hidden patterns in dynamic electric signals as well as in numerically mapping power system DVRs due to its ability to consider relevant operating statistics, including the most probably severe events that could lead the system to potential insecure conditions and subsequent blackouts.

This paper presents a new approach to predict the time evolution of rotor angles of synchronous machines based on a modification of the Taylor series expansion in order to assess the transient stability of power systems in real time. To demonstrate the benefits of the proposed approach, a comparative analysis is made with other approaches which have been used for predicting the rotor angles, namely regression and interpolation algorithms and an approach based on original Taylor series expansion. The different prediction approaches are applied to the New England benchmark power system to predict the time evolution of the rotor angles in terms of stability and instability of first and second swing. The obtained results highlight the prediction accuracy of the proposed approach and the fulfilment of computational time requirements concerning real time applications.