IMLANNs for congestion management in power system

In this paper, Integrated Multi-Layer Artificial Neural Networks (IMLANNs) model has been developed for congested line prediction in a power system. The master characteristic of an ANN is the superiority to achieve complicated input-output mappings through a learning procedure, without exhaustive programming efforts. The IMLANNs model was developed to predict the congested lines in a power system. Before the IMLANNs model is developed, a case study was selected to receive an early result in power system load current during normal condition and contingency based on heavily loaded term. In order to optimize the architecture of the neural network and minimize the computational effort, but those state variables with major impact on the power system are selected as inputs. A pre-developed index, namely Fast Voltage Stability Index (FVSI) is employed as a benchmark to identify the locations declared as congested lines. This indicator was produced which aims for an analytic thinking, sustainable power system when an excessive load was imposed on the power system network. In addition, voltage collapse can be identified when the index is approaching 1.000 or unity. The value of FVSI is chosen as the targeted output in the IMLANNs model. The strength of the proposed IMLANNs model has been validated on the IEEE 30-Bus RTS. Results obtained from the study demonstrated that the proposed IMLANNs is feasible for congested line prediction, which in turns beneficial to power system operators in the planning unit of a utility.