Alvaro Torres

Power Systems Security Evaluation Using Petri Nets

Misoperation and hidden failures of protections in electric power systems have been recognized as a contributing factor to power system cascading outages and catastrophic events. For that reason, it is important to develop new tools to study the sequence of the operation of protections given a contingency in the system. This paper proposes a methodology for modeling the operating sequences of protections in the IEEE 493 system and a meshed transmission system using generalized stochastic petri nets (GSPNs), contemplating main and back-up protections, and specially how they interact.

Spectral Graph Theory and Network Dependability

The paper introduces methods of graph theory for ranking substations of an electric power grid. In particular, spectral graph theory is used and several ranking algorithms are described. The procedure is illustrated on a practical numerical example

Contingency Analysis and Risk Assessment of Small Signal Instability

Power system operating planning must determine adverse conditions that could affect the normal performance of the system. Load variations and contingencies could provoke these adverse conditions, which produce angular instability problems or voltage stability problems, among others. In addition, the short-term operating planning must evaluate risk of instability conditions for angular and voltage instability. This paper proposes a methodology to assess the angular instability risk due to load uncertainty based on modal analysis of the small signal analysis for short-term planning purposes. In addition, the risk assessment of angular instability is done for selected contingencies based on an eigenvalue variation contingency ranking proposed in this paper. The methodology is tested in the New England 39 nodes system.

Angular Instability “Day Ahead” Risk Forecasting ‑ Probabilistic Dependency on Load

This paper proposes a methodology in order to evaluate the probabilistic impact of load variations on angular stability of power systems, around 24 hours day ahead planned operation conditions. For each operation condition, the load is forecasted; however, at the operation moment the real load is different. So, a probabilistic small signal analysis around the operational condition allows the establishment of a probabilistic density function (pdf) related to the behaviour of critical eigenvalues of the dynamic lineal system. The hourly pdf is computed based on an hourly Monte Carlo simulation and, once the main statistical moments are computed, the angular instability is determined from the pdf. The method is tested using the New England 39-node system.

Analysis of Electrical Industrial Systems using Probabilistic Networks

This paper presents a methodology in order to evaluate the security of electrical industrial systems considering security, quality, reliability and availability factors and the system response to sudden disturbances such as short circuit and harmonics.The models proposed using the probabilistic networks theory: Bayesian Networks and Generalized Stochastic Petri Nets. The obtained results allow the identification of hidden failures and erroneus coordination of protection. The models are tested on the IEEE 493 system proposed by the IEEE Gold Book.

Bayesian Networks for Power Quality Analysis in the Industrial Sector

This article presents a model to evaluate the power quality phenomenon denominated voltage sags by means of Bayesian networks. This model considers the main events that can be generated by the presence of voltage sags in the power industrial systems, especially on electronic and sensitive equipment. The relations of causality established between the main causes and the final effects consider both the internal and the external electrical conditions of the industrial power system, such as environmental effects. The probabilities of the causality relations are assigned from statistical data available in literature, from Monte Carlo simulations, sensitivity analysis and expert opinion. The model is very versatile, since it analyzes the sag phenomenon for any type of industry and voltage level and its impact on possible actions or corrective measures in aspects such as: maintenance, selection and coordination of protections, etc. in order to correct or to reduce the voltage sag phenomena and improve the service quality to the client and to other clients connected to the same distribution circuit

Analysis of the prospects for distributed generation (DG) for Colombian electric power sector

The main aim of this paper is to present the results of an analysis of the prospective for the distributed generation (DG) in the Colombian electric power sector applying the Delphi technique. This study helped us understand the experts opinion directly and systematically on the introduction of the DG in fields like: technology, product innovations, materialization process, effect on the countrys political and social situation, economic, technologic and commercial restrictions. The results of this work are intended to provide support for the electric and productive sectors to analyze the feasibility of using the DG in the medium and long term in Colombia, since its application would have a very strong effect on the companys planning and operation activities regarding electric issues.

Bayesian Networks and Monte Carlo Simulations in the Evaluation of the Risk of Terrorism for the Colombian Electrical Infrastructure

The aim of this paper is to propose a probabilistic method to assess the level of exposure and consequences due to terrorist attacks on electrical substations or other electrical infrastructure assets. The proposed method of inference and simulation allow us to establish a risk-ranking list for the electrical infrastructure assets. The proposed method is based on Bayesian networks, using causality relationship and conditional probabilities to determine the probability of the attacks associated to each one of the substations and to quantify the damage or loss associated to those possible attacks

Soft computing techniques for analysis of kidnapping crimes in Colombia

Techniques associated with soft computing have been proven to be useful for solving problems where lack of information and uncertainties are present. The flaws of traditional methodologies stem from the fuzziness, partial true, lack of information, and event imprecision of problems related to crime analysis such as kidnapping. Case based reasoning and experts knowledge are the main inputs of some soft computing techniques that may be applied for analysis of kidnapping crimes. Kidnapping is a problem framed by the lack of veracity, scarceness, fuzziness of information, and tacit, qualitative and unstructured knowledge in the minds of few experts. So, kidnapping is the kind of problem that could be modeled and analyzed by using some of the techniques included in the field of soft computing. This paper describes the use of fuzzy logic, fuzzy clustering and classification to analyze and evaluate kidnapping crimes in Colombia. These techniques are intended to predict the captivity duration, captor group and liberation conditions for any victim. Also, these techniques could be used to assess the impact of government security and economic policies on the evolution of the kidnapping crime

Reliability assessment of distribution power repairable systems using NHPP

Generally, stochastic point processes are employed in reliability analysis of repairable systems. However, when failures and reparations of these systems occur in a non-stationary trend over time, for example due to aging, it is required to use non-homogeneous Poisson process. In this paper, it is presented a general methodology for reliability assessment of repairable systems using non-homogeneous Poisson process. The methodology uses sequential Monte Carlo simulation and is applied to a power distribution test system affected by adverse weather conditions and aging of its components. As a conclusion, it was found non-homogeneous Poisson process is an excellent tool for modeling the reliability of repairable systems when it is required to consider realistic conditions in the analysis.