A distributed, real-time and non-parametric approach to demand response in the smart grid

Abstract

This paper considers a novel approach to demand response in a smart grid, wherein a set of flexible loads with given energy requirements and deadlines for completion adjust their instantaneous power consumption in such a way as to make the aggregate total power consumption on the grid as smooth as possible. Each load periodically adjusts its own power level using only a minimal amount of feedback i.e. the aggregate power consumption observed in the previous period, and knowledge of its own energy requirement and deadline. Our approach allows for a fully distributed, real-time implementation and does not require any knowledge of arrival process statistics or forecasts of future load. We illustrate the working of our algorithm with a detailed simulation of generation and load for a fictional small city modeled after the city of Ames, IA using data from a variety of sources. In this simulation, charging electric vehicles serve as flexible loads. We also present an approximate analysis of the bahaviour of flexible loads in our scheme based on a continuous time version of the proposed algorithm to gain insight into its properties and limitations.

Serving the Grid: an Experimental Study of Server Clusters as Real-Time Demand Response Resources

“…For example, one of our assumptions in the power tracking experiment was a workload which could be parallelized without job dependencies or strong SLA restrictions. Along those lines, distributed control as in [47], [48], but incorporating the effect of SLA and workload structure (for example, MapReduce jobs) seems to be an attractive starting point for future research.…”

PAS: An Efficient Privacy-Preserving Multidimensional Aggregation Scheme for Smart Grid

“…Compared with traditional grid, smart grid supports centralized two-way transmission and efficiency-driven response. Besides, smart grid relies on cyber-physical systems or the Internet of Things to provide intelligent scheduling for power transmission and distribution [9]. For instance, smart meter equipped with network interfaces (e.g., wireless sensors) reports power consumption to the operation center via the advanced meter controlled by the regional center, as shown in Figure 1.…”

Distributed control system for demand response by servers

“…The adaptation scheme is designed for distributed implementation at the loads without any global coordination, and the only communication required is knowledge of the aggregate load in the grid in the previous time period (or more precisely the deviation of the aggregate load from the desired operating point). We presented this idea in [17] using simulated case study of electric vehicle battery chargers as deferrable loads for distributed demand response.…”

“…In the generator dispatch case, a generator jostles to generate its relative share of the load according to its marginal cost of generation. In [17], the battery chargers jostle to consume their share of the load according to their charge state and the time remaining until their deadline. In the present work, the computational jobs jostle to consume their share of the load according to the state of and constraints on their workload throughput.…”