Recurrent network based power flow solution for voltage stability assessment and improvement with distributed energy sources

The increasing penetration of alternative energy sources into the integrated energy systems often influences the voltage stability (VS) of the entire system. However, developing a technique that comprehensively analyze the energy flow in this environment is a major challenge. This paper presents a novel heuristic-based recurrent type Hopfield Neural Network (h-HNN) planning tool for VS assessment of power system; towards reducing the computational cost of conventional power flow (PF) method. The proposed approach is a Jacobian-less, energy function-based approach, which was formulated using power residuals of the system. The dynamics of neural networks were governed by the differential equations of energy function, which would be minimized by the heuristic particle swarm optimization-gravitational search algorithm to deduce the unknown parameters of voltage magnitude and phase angle. The proposed technique was coded in MATLAB and its effectiveness was tested on IEEE 14-, 30-, and 57- buses, as well as a 1354-bus test system. The obtained results were compared with well-known PF techniques, and the robustness was demonstrated for ill-conditioned network. A composite severity index was proposed to rank the critical contingency of the energy network. Then, the VS assessment was performed in IEEE 14-bus system under severe contingency conditions and improvement of VS is observed under the penetration of distributed energy sources (DES). During the case of DES placement, (i) the voltage profile of the system is maintained within the acceptable range of 0.95 to 1.05 pu and (ii) the VS of the system evaluated using stability indices are enhanced by an amount of 16.54 % to 88.16 %. The application results indicate that the proposed method is useful for electric energy utilities to assess the state of the system under monitoring process.