The effect of incomplete price signals on a price-responsive distribution feeder

Using the smart grid simulator Gridlab-D, a representative feeder model was implemented with residential price-responsive air conditioning units. A real-time price signal was constructed based on CAISO FERC reports and was fed to residential price-responsive thermostat controllers. To simulate the effects of dropped packets in the communications network, the price signal sent to particular residences was not updated on the typical hourly basis simulating dropped packets in the communication network; the dropped packet rate was varied from 0% to 90%. The average load during the peak period of noon to 7pm was examined and found to be constant versus the dropped packet rate, though thermostate controllers, in response to higher prices, were able to reduce the energy bill by shifting air conditioning usage out of high price periods. The cumulative energy bill for the entire feeder was affected by the dropped packet rate with an increasing drop rate leading to higher energy bills.

I. Introduction

ONE of the key features of the emerging smart grid is the proliferation of measurement and communication infrastructure throughout the grid and one proposed use of this new infrastructure is to enable a wider level of participation in the energy market. Traditionally, residential customers are not billed based on the time the energy is consumed, only on the total amount of consumption. Though the wholesale price of energy varies throughout the day, retail customers do not have visibility or financial incentive to this price and thus do not respond to it. With smart grid initiatives being developed, it is now possible to use this new communication infrastructure to create a true market for energy, a market where a real-time price signal will influence residential consumer behavior and effect the load on the system.