Forecast-Based Stochastic Optimization for a Load Powered by Wave Energy: Article No. 120330

Trent Dillon, Benjamin Maurer, Michael Lawson, Brian Polagye

Research output: Contribution to journalArticlepeer-review

Abstract

Stand-alone renewable energy systems, which harness and use electricity without connection to a distribution grid, can benefit from using resource forecasts when making real-time decisions on how to manage power. An example of this is an ocean observation system that uses a wave energy converter (WEC) and battery bank to produce and store electricity in order to meet electrical load requirements for oceanographic sensors and instrumentation. By using wave forecasts to optimize the power consumption of this system, it is possible to reduce the size of the WEC and battery bank needed, thereby reducing the overall system's cost and complexity. In this paper, we assess the benefits of forecasting for a generic wave-powered ocean observation system. We model an observation platform powered by a WEC and battery-storage system that can switch between four power states, full power: 600 W, medium power: 450 W, low power: 45 W, and survival mode: 1 W. To determine which power state to enter over time, we present a stochastic optimization method that interprets wave forecasts and system information to select a power state on an hourly basis and simulate over two months of consecutive decisions. Using this simulation framework, we compare 8 power management strategies across a range of WEC sizes (3-5 m) and battery capacities (2.5-35 kWh). We find that it is possible to maintain full (600 W) power consumption over the entire simulation window with a 5 m diameter WEC and 25 kWh battery bank, which is on the upper end of the ranges evaluated. To employ a smaller WEC or battery, load flexibility is required. We find that forecast-based methods for this power management handle tradeoffs in performance (e.g., cumulative power consumption, reducing intermittencies and meeting scheduled load targets) more effectively than power management strategies that do not use forecasting. Overall, our results indicate that forecasting for our stand-alone renewable energy system is most impactful when some loads must be met on or within a defined schedule.
Original languageAmerican English
Number of pages24
JournalRenewable Energy
Volume226
DOIs
StatePublished - 2024

NREL Publication Number

  • NREL/JA-5700-90686

Keywords

  • demand management
  • forecasting
  • Markov decision processes
  • wave energy

Fingerprint

Dive into the research topics of 'Forecast-Based Stochastic Optimization for a Load Powered by Wave Energy: Article No. 120330'. Together they form a unique fingerprint.

Cite this