Different approaches have been proposed previously for scheduling demand-side battery storage, usually with one of two objectives; alleviating the need for distribution grid reinforcement by managing bi-directional power flows or reducing electricity bills for customers. However, without careful coordination, the potentials of demand-side approaches might not be fulfilled.
Dr. Elizabeth Ratnam from the University of California Berkeley (formerly of the University of Newcastle, Australia) and Professors Steven Weller and Christopher Kellett from the University of Newcastle in Australia investigated two optimization-based algorithms to balance an increase in the operational savings that accrue to residential customers with combined photovoltaic (PV) battery storage systems against management of distribution grid power flows to alleviate voltage rise and other local conditions that necessitate grid reinforcement. The article was published in the International Journal of Electrical Power and Energy Systems.
The two optimization-based approaches used are a centralized quadratic program energy-shifting, where selected customers implement a distributor-specified day-ahead battery schedule and a second approach, referred to as local quadratic program energy-shifting, where distributor-specified weights are incorporated into a quadratic program-based algorithm implemented directly by customers to obtain an individual day-ahead battery charge and discharge schedule. The algorithms were applied to load and generation data from 145 Australian residential customers to investigate the customer-distributor benefits of coordinated residential battery scheduling.
The researchers introduced a modeling framework consisting of a dynamical model of a residential energy system and a distribution region described by a directed graph. Residential customers were identified in a specified region and considered ways to coordinate their day-ahead battery schedules under the assumption of a financial policy of net metering.
The two different optimization-based approaches require load and generation forecasts at different locations in the network. Forecasting is performed by the distributor in the centralized QP case, while in the local QP case the forecasting is done at each residence. The authors described an approach to emulate imperfect forecasts using historical data.
As a particular application of the techniques presented, the authors investigated the case where the distributor identifies its “weakest link” (via a power flow analysis), which then provides the distribution region of interest and places constraints on the optimization problems.
When assessing the benefits of residential battery scheduling with reference to a 52-week period, the baseline profile exceeded the upper limit for the edge of interest at subgraph forecast constraints of 300KW on 9 days in a year and falls below the lower limit of -150KW on 5 days in the year. It can be said that on most days, subgraph members received a reliable supply of electricity when they do not use or have a battery.
When assessing operational savings accrued to a single subgraph member (i.e., each residence) over a period of 52 weeks denoted by annual savings in $/year, it was seen that local quadratic program energy-shifting may disproportionately penalize some customers when implementing a local quadratic program-based battery schedule. In fact, using such a localized approach resulted in a few customers seeing additional annual costs.
By contrast, the authors demonstrated that, in terms of customer benefit, the centralized quadratic program-based approach was preferable in that all customers received the same annual savings, so that no customers were penalized for utilizing battery storage.
Elizabeth L. Ratnam1, Steven R. Weller2, Christopher M. Kellett 2. Central versus localized optimization-based approaches to power management in distribution networks with residential battery storage, International Journal of Electrical Power and Energy Systems 80 (2016) 396-406.Show Affiliations
- Center for Energy Research, Department of Mechanical and Aerospace Engineering, University of California, San Diego, 9500 Gilman Drive, La Jolla, CA 92093-0411, USA.
- School of Electrical Engineering and Computer Science, University of Newcastle, University Drive, Callaghan, NSW 2308, Australia.
Go To International Journal of Electrical Power & Energy Systems