Grid planning: Agent based approach for early notification of air conditioning loads to smart grid

Abstract

Electricity needs are increasing with more ownership of air conditioning systems in houses. Air conditioning loads are a significant portion of the overall peak load on grids. In order to match generation and load in a more optimal way and reduce the strain on the grids, we need load information of houses in advance. The main problem is that the majority of houses are still not smart enough to notify the grids about future loads. Even the smart houses with such capability, the users need to keep track of it, which is a cumbersome manual task. Furthermore, the houses do not have the ability to sense the occupant and air conditioning requirements if the occupant is on the road. In this paper we have proposed a system to tackle these problems. The proposed system is a smart agent developed by combining existing technologies such as smart phones, GPS sensors, Internet over GSM networks, cloud services, and a novel application of image processing to combine interfaces of different components and technologies. Development of this smart agent requires low cost, less resources, and no modification to the existing systems. Lab results show that the prototype agent can easily sense and notify the upcoming air conditioning loads to the grid operator. A conservative estimate of the proposed systems' effect on the grid shows that in peak hours by just keeping 1000 ACs away from the grid for a single minute may result in a significant reduction of grid load by 1 MW.