Neural network based real-time pricing in demand side management for future smart grid

Abstract

Electricity grid is currently being transformed into smart grid. Increased number of renewables require more and more ancillary services to backup intermittent power generation. A very important topic in tomorrow's electricity grid is demand side management. This tool should be used as an alternative for traditional backup power reserves. It requires a deep understanding on how consumption depends on dynamic pricing. This paper proposes a method for modelling the electricity demand response to a real-time pricing. A virtual smart house is modelled using Gridlab-D smart grid simulator. The HVAC system is setup to respond to real-time price sent by the utility. The paper is analysing the ability of neural network to predict the exact price, which is sent to the end user in order to maintain the supply balance in the system. It should also reduce the peaks in demand and increase system resilience.