Papers in international journals

The LASCA Strategy

Vladimiro Miranda, Vera Palma Ferreira and Joana Hora Martins, “Optimizing large scale problems in a reduced space mapped by autoencoders”, in the process of submission to IEEE Transactions on Power Systems, 2013. This paper explores a technique to solve large scale optimization problems by reducing the search space dimensionality with the application of autoencoders. The technique applies autoencoders as a reversible mapping between the original problem dimension and a reduced space, which allows an evolutionary metaheuristic to evolve in a reduced space, having its objective function assessed in the original space. The technique is illustrated with an application of an EPSO (Evolutionary Particle Swarm Optimization) algorithm to a Hydro-Wind coordination problem and four benchmark optimization functions. The results obtained suggest that the new technique allows an improvement in the quality of solutions attained.

Topology discovery - comparison of approaches

Cátia S. P. Silva, Jakov Krstulovic, Joana H. Martins, Vera Palma, Vladimiro Miranda and José C. Príncipe, “Extracting topology information from electric measurements: a model comparison”, submitted to IEEE Transactions on Power Systems, 2013. This paper confirms that the network topology information lies hidden in the manifold supporting the solutions of the power flow equations and that suitable methods may make it explicit without direct information on the breaker status. A set of methods is applied to the identification of the unknown status of a switch, by dealing only with local electric information, and their performance efficiencies are compared. One of the methods uses optimal subspace projections using metric learning with an entropy functional that preserves classification accuracy. The results have direct influence on the way one may build local topology estimators either in distributed or centralized state estimation.

Auto-associative topology estimator

Jakov Krstulovic, Vladimiro Miranda, Antônio Simões Costa, Jorge Correia Pereira, Towards an auto-associative topology state estimator, IEEE Transactions on Power Systems, vol.28, no.3, pp.3311-3318, Agosto, 2013. This paper presents a model for breaker status identification and power system topology estimation based on amosaic of local auto-associative neural networks. The approach extracts information from values of the analog electric variables and allows the recovery of missing sensor signals or the correction of erroneous data about breaker status. The results are confirmed by extensive tests conducted on an IEEE benchmark network.

Auto-associative fault diagnoser

Vladimiro Miranda, Adriana Castro, Diagnosing faults in power transformers with autoassociative neural networks and mean shift., IEEE Transactions on Power Delivery, vol.27, no.3, pp.1350-1357, Julho, 2012. This paper presents a new approach to incipient fault diagnosis in power transformers, based on the results of dissolved gas analysis. A set of autoassociative neural networks or autoencoders are trained, so that each becomes tuned with a particular fault mode or no fault condition. The scarce data available forms clusters that are densified using an Information Theoretic Mean Shift algorithm, allowing all real data to be used in the validation process. Then, a parallel model is built where the autoencoders compete with one another when a new input vector is entered and the closest recognition is taken as the diagnosis sought. A remarkable accuracy of 100% is achieved with this architecture, in a validation data set using all real information available.

Autoencoders and missing data

Vladimiro Miranda, Jakov Krstulovic, Hrvoje Keko, Cristiano Moreira, Jorge Pereira, Reconstructing missing data in State Estimation with autoencoders, IEEE Transactions on Power Systems, vol.27, no.2, pp.604-611, Maio, 2012. This paper presents the proof of concept for a new solution to the problem of recomposing missing information at the SCADA of energy/distribution management systems (EMS/DMS), through the use of offline trained autoencoders. These are neural networks with a special architecture, which allows them to store knowledge about a system in a nonlinear manifold characterized by their weights. Suitable algorithms may then recompose missing inputs (measurements). The paper shows that, trained with adequate information, autoencoders perform well in recomposing missing voltage and power values, and focuses on the particularly important application of inferring the topology of the network when information about switch status is absent. Examples with the IEEE RTS 24-bus network are presented to illustrate the concept and technique.

LASCA applied to wind-hydro coordination

V. Miranda, J. da Hora Martins and V. Palma, "Optimizing large scale problems with metaheuristics in a reduced space mapped by autoencoders – application to the wind-hydro coordination", IEEE Transactions on Power Systems, accepted for publication on 14/4/2014, Copyright IEEE. This paper explores a technique denoted LASCA to solve large scale optimization problems with metaheuristics by reducing the search space dimension with autoassociative neural networks. The technique applies autoencoders as a reversible mapping between the original problem space and a reduced space. A metaheuristic then evolves in the latter, having its objective function assessed in the original space. The technique is illustrated with an application of an EPSO (Evolutionary Particle Swarm Optimization) algorithm to four benchmarking unconstrained optimization functions and to a wind-hydro constrained coordination problem. The new technique allows an improvement in the quality of the solutions attained.