Demand response-based voltage security improvement using artificial neural networks and sensitivity analysis

As a precautionary remedy, load shedding has always been regarded as a strong choice when facing a voltage collapse. On the other hand, Demand Response (DR) is often an interactive communication which highlights customer´s participation, more often in smart grid technologies. Moreover, DR plan is introduced as an appropriate choice when system Voltage Stability is jeopardized. In this paper, a new approach for improving voltage security is brought up using DR plan, sensitivity analysis and neural network which is accentuated by its super-fast processing. Since different load patterns result in different Pmax and PV curve, a unique way of DR units participation is explored in which the optimum load decrease pattern and consequently the optimum VSM improvement are met when the least amount of DR units participation is employed, In this research, IEEE 39-BUS power grid is selected as the case study, and PV curve method is used for voltage seeurity analysis. Then MLP ANNs are used to speed up the calculations during the system operation.