Renewable energy management system: Optimum design and hourly dispatch

Baraa Mohandes, Maisam Wahbah, Mohamed Shawky El Moursi, Tarek H.M. El-Fouly

Research output: Contribution to journalArticlepeer-review

42 Scopus citations

Abstract

This paper introduces a new framework for optimum design and operation of hybrid renewable energy plants (HREP) augmented with battery energy storage systems (BESS). A new renewable energy management system (REMS) is developed comprising three components: 1) Enhanced joint forecasting of wind and solar outputs based on deep neural networks and also multiplicative weights update (MWU); 2) an advanced optimization model for sizing the HREP-BESS components and the policy of BESS operation; and 3) Augmenting the rolling hourly dispatch for HREP-BESS with a novel dynamic ramping limit and a criterion for reduction of deviations from the hour-ahead dispatch schedule. The proposed REMS tool enables maintaining the inter-hourly ramping of the HREP-BESS output within a threshold. In this context, a novel dynamic ramp limit is proposed to minimize the energy curtailment during operation and maximize energy sales to the power grid. The advantage of the proposed REMS tool over the classical renewable energy systems operation scheme is the mitigation of the volatility of renewable energy sources (RES) by suppressing extreme ramping events with minimum curtailment. Moreover, the costs and revenues of the HREP-BESS design and operation are assessed over a 25 years period. The design problem is solved for different scenarios, and the optimal solution always encloses a hybrid mix of renewables where the share of the PV plant can reach up to 37.1% of the total plant size. With the proposed REMS, the curtailment of RES never exceeds 12.9% even when the HREP is operated without a reserve margin. For the selected design, the optimum BESS capacity is 12.9% of the HREP capacity. The number of hours which observe a ramping violation event is 2.4% of the season's length (2184 hours). 99% of all ramping events fall within the defined ramping limits. The use of the MWU method increases the total profit by 2.53% compared with adopting the average forecast.

Original languageBritish English
Article number9351770
Pages (from-to)1615-1628
Number of pages14
JournalIEEE Transactions on Sustainable Energy
Volume12
Issue number3
DOIs
StatePublished - Jul 2021

Keywords

  • Deep learning
  • dynamic ramping limit
  • hybrid renewable energy
  • multiplicative weights update
  • proxy value of battery energy
  • rolling horizon

Fingerprint

Dive into the research topics of 'Renewable energy management system: Optimum design and hourly dispatch'. Together they form a unique fingerprint.

Cite this