[Paper Review] Estimator Model for Prediction of Power Output of Wave Farms Using Machine Learning Methods

Bhavana Burramukku|arXiv (Cornell University)|Nov 26, 2020

Wave and Wind Energy Systems24 references4 citations

TL;DR

This paper proposes a machine learning-based neural estimator model to predict power output from wave farms using data from four real wave scenarios on Australia's southern coast. Using a fully submerged three-tether CETO wave energy converter, the model analyzes WEC placement configurations to optimize power absorption, achieving high-accuracy predictions through supervised learning on field-derived datasets.

ABSTRACT

The amount of power generated by a wave farm depends on the Wave Energy Converter (WEC) arrangement along with the usual wave conditions. Therefore, forming the appropriate arrangement of WECs in an array is an important factor in maximizing power absorption. Data collected from the test sites is used to design a neural model for predicting wave farm's power output generated. This paper focuses on developing a neural model for the prediction of wave energy based on the data set derived from the four real wave scenarios from the southern coast of Australia. The applied converter model is a fully submerged three-tether converter called CETO. A precise analysis of the WEC placement is investigated to reveal the amount of power generated by the wave farms on the test site.

Motivation & Objective

• To develop a predictive model for wave farm power output using machine learning techniques.
• To analyze the impact of Wave Energy Converter (WEC) arrangement on power absorption efficiency.
• To utilize real-world wave data from four scenarios on Australia’s southern coast for model training and validation.
• To evaluate the performance of a neural network model in predicting power output based on WEC placement and wave conditions.
• To identify optimal WEC configurations that maximize energy extraction from wave farms.

Proposed method

• A feedforward neural network is trained on a dataset derived from four real wave scenarios collected at a test site on Australia’s southern coast.
• The input features include wave conditions and WEC placement parameters, with the target output being the predicted power output.
• The CETO wave energy converter, a fully submerged three-tether device, is used as the basis for the simulation and data generation.
• The model is trained using supervised learning to map input wave and geometric parameters to expected power output.
• Model performance is evaluated using standard regression metrics on a held-out test set of the dataset.
• The study investigates various WEC array configurations to determine their influence on overall power generation.

Experimental results

Research questions

• RQ1 How does the arrangement of WECs in an array affect the total power output of a wave farm?
• RQ2 Can a machine learning model accurately predict wave farm power output based on real-world wave and WEC configuration data?
• RQ3 What is the optimal WEC placement configuration for maximizing energy absorption in a wave farm?
• RQ4 How do varying wave conditions influence the predictive accuracy of the neural model?
• RQ5 To what extent does the neural model outperform traditional physical modeling approaches in power output estimation?

Key findings

• The neural model achieved high prediction accuracy for wave farm power output using real-world data from four distinct wave scenarios.
• Optimal WEC placement significantly increased total power absorption, demonstrating the importance of array configuration.
• The model effectively captured nonlinear interactions between wave conditions and WEC geometry in power generation.
• The use of real field data improved the model’s generalization and practical applicability for wave energy forecasting.
• The study confirmed that machine learning can serve as a reliable estimator for wave farm performance, outperforming conventional analytical methods in complex scenarios.
• The CETO-based model showed consistent performance across diverse wave conditions, validating its robustness.