Distributed Smart Grid Asset Control Strategies for Providing Ancillary Services

Abstract

Printed in the United States of America Executive summaryWith large-scale plans to integrate renewable generation driven mainly by state-level renewable portfolio requirements, more resources will be needed to compensate for the uncertainty and variability associated with intermittent generation resources. Distributed assets can be used to mitigate the concerns associated with renewable energy resources and to keep costs down. Under such conditions, performing primary frequency control using only supply-side resources becomes not only prohibitively expensive but also technically difficult. It is therefore important to explore how a sufficient proportion of the loads could assume a routine role in primary frequency control to maintain the stability of the system at an acceptable cost.The main objective of this project is to develop a novel hierarchical distributed framework for frequency based load control. The framework involves two decision layers. The top decision layer determines the optimal gain for aggregated loads for each load bus. The gains are computed using decentralized robust control methods, and will be broadcast to the corresponding participating loads every control period. The second layer consists of a large number of heterogeneous devices, which switch probabilistically during contingencies so that aggregated power change matches the desired amount according to the most recently received gains. The simulation results show great potential to enable systematic design of demand-side primary frequency control with stability guarantees on the overall power system. The proposed design systematically accounts for the interactions between the total load response and bulk power system frequency dynamics. It also guarantees frequency stability under a wide range of time varying operating conditions. The local device-level load response rules fully respect the device constraints (such as temperature setpoint, compressor time delays of HVACs, or arrival and departure of the deferrable loads), which are crucial for implementing real load control programs.

Loads as a Resource: Frequency Responsive Demand

“…In FY 13, a systematic approach was proposed in [2] for a demand-side primary frequency strategy to regulate the frequency of the system to its nominal value and restore power balance for a multi-machine power system model. The load control law was designed in a hierarchically decentralized manner consisting of two interactive decision layers.…”

“…Another objective in FY13, as detailed in [2], was to perform extensive simulation studies to investigate the impact of a population of autonomous Grid Friendly™ Appliances (GFAs) on the bulk power system frequency stability. A preliminary investigation of the grid impacts of the GFA design indicated that the location of GFAs affected system stability performance.…”

“…It is clear that the network topology has been preserved by the above model, which can more accurately represent the transient dynamics of the power system than the multimachine power system model considered in FY13 detailed in [2] and [15].…”

“…The power modulation i u is designed as a feedback control input characterized by a control gain i k . As opposed to the design proposed in FY13 ( [2] and [15]), it is now assumed that the frequency at each load bus i is the only available state. Therefore, the feedback power modulation can be defined as follows, 0 ()…”

“…The controllers and their appliances responded reliably to each shallow under-frequency event, which was an average of one event per day and shed their loads for the durations of these events. The report presented in [10] studied the dynamic effects of populations of Grid Friendly™ water heater controllers on the power grid during under-voltage and under-frequency events. In particular, models for water heater with GFA controllers were developed using a scripting programming language and simulated in GE Positive Sequence Load Flow (PSLF), a software suite designed to study power systems in both steadystate and dynamic environments.…”