Towards on-line optimal reactive power scheduling using ANN memory model based method

In an attempt towards on-line implementation of optimal reactive power scheduling (ORPS), a fast method using a mix of ANN based bi-directional associative memory (BAM) model and corrective control algorithm (CCA) is proposed. Sufficient number of previously encountered system states and their optimal control settings obtained through a conventional method are stored in the BAM's weight matrix in the encoded form. Then, the BAM is put on-line. Given any arbitrary system state, the BAM recalls the closest previously encountered system state and its corresponding optimal settings. The optimal settings are corrected for any operational constraint violation using the CCA and are implemented. The method was tested on a modified IEEE 30 bus system. Twelve sets of system states and their optimal settings were stored. Given any arbitrary system state, the BAM successfully recalled the closest stored system state and its optimal settings which were corrected for violations using the CCA. The time taken was approximately one tenth of that of the conventional method demonstrating the usefulness of the proposed method.