How Infigen Energy improved Grid Stability and Revenue Flow with Energy Analytics for Short-Term Forecasting

Many countries and enterprises are diversifying their energy mix with renewables as a way to address climate change. However, intermittent renewable resources present challenges because variability in wind and solar generation can cause an imbalance between supply and demand in the power grid. Power forecasting tools based on advanced data analytics can help accurately predict these fluctuations to reduce any imbalances and also lower generator market penalties. Collectively, in Australia in 2019, renewable power generators paid a whopping USD182 million in total Frequency Control Ancillary Services (FCAS) penalties.

Artificial Intelligence-driven technology being deployed in Southern Australia offers an example of the successful use of these forecasting techniques. The
Lake Bonney Stages 2/3 project by Utopus Insights, completed in partnership with Infigen Energy, ARENA and Vestas, represents an industry-leading solution that is both essential and differentiating, and led to a 26% reduction in causer pays factor for Infigen. Our forecasting models performed well early into the project to show results for Infigen, and we were the first wind power self-forecasting service to qualify on their behalf. 

The solution improves the accuracy of power output forecasts in real time, making renewable energy more cost-competitive vis-a-vis conventional energy sources, and enabling customers to make strategic and more profitable decisions about the energy they generate.

Lake Bonney Stages 2/3 Project

In  March 2019, the
Australian Renewable Energy Agency (ARENA) awarded funds to develop a new short-term wind power forecasting solution for Infigen Energy’s Lake Bonney Stages 2 and 3 wind farms in Southern Australia. The funding was part of ARENA’s Advancing Renewables Program. The solution was developed by
Vestas (Australia’s largest wind energy solution provider), together with its wholly-owned energy analytics digital solution provider,
Utopus Insights.

The overarching goals for the project were to more accurately forecast energy generation from wind power plants at a 5-minute ahead dispatch interval, reduce power dispatch uncertainty for wind generators, and improve system stability by enhanced anticipation of supply from renewable sources.

Picture Courtesy – Infigen Energy: Lake Bonney Farm 1

This class of forecasting solutions can also yield bottom line effects for power plant operators by minimizing financial penalties for inaccurate forecasts and increase revenue opportunities through more intelligent and strategic data-driven decisions about selling energy into spot markets. Wind and solar farms are penalized for failing to meet required energy generation levels. Moreover, they can be required to curb generation to match overly conservative forecasts, thus impacting their revenues. Inaccurate supply forecasts or failure of generators to meet targets can result in power system instability and higher operating costs. Being able to better predict power generation is a valuable tool in calculating the real-time value of energy, and also an important segue into a world where wind and solar energy are rapidly becoming integral sources of electricity.

Our Solution – Advanced Data Science Methodologies Used

In total, Utopus Insights developed four core, diverse models that were submitting self forecasts used for farm dispatches by AEMO. All four models performed better than AWEFS in terms of Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) at Lake Bonney 2 and 3 for most of the time. This was especially demonstrated at higher power production ranges, with only a few exceptions and for limited periods of time. Our industry-differentiating, SCADA-only model outperformed AWEFS forecasts, using existing infrastructure and not requiring any additional, cumbersome hardware. Our SaaS solution was effective in delivering the needed power forecasting accuracy requirements.

The availability and redundancy of the deployed data pipeline were instrumental in ensuring data ingestion from the assets with minimum delay and rapid recovery in the event of unwarranted system issues. With a robust data pipeline and data curation strategy, it was possible to deploy an AI-based algorithm to generate, deliver and submit forecasts via an API in compliance with Australian Energy Market Operator (AEMO) requirements.

Utopus Insights Software-as-a-Service (SaaS) solution was built using state-of-the- art cloud technology and infrastructure, to ensure that data security and scalability take the utmost importance, along with additional tools that empower the customer to understand the provided forecasts better.

Picture Courtesy – Infigen Energy: Lake Bonney Farm 2

Project Success Points

In order to carry on its energy transition journey, Australia has continued to grow its renewable energy penetration year-over-year. Specifically,
ARENA’s Advancing Renewables Program, aiming to improve system stability, is acting as a catalyst. Accurate short-term forecasting helps AEMO achieve better dispatch planning, and also helps Independent Power Producers (IPPs) reduce their FCAS liability.

Infigen, Vestas, and Utopus Insights collaborated over this funded pilot with the aim to achieve both reliable and accurate short-term forecasts. The
Australian Wind Energy Forecasting System (AWEFS) is the baseline over which self-forecasts (forecasts submitted by the participant) performance gets assessed, both from a reliability and accuracy perspective.

The machine learning model performance metrics used were Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) for the evaluation of the unconstrained self-forecast. Example: the actual power any farm is capable of producing, irrespective of any curtailment.

What started as a pilot project to test short-term forecasting, has developed into a solution from Utopus Insights that can now serve power producers in the Australian NEM, and can serve customers globally in the future as market regulations evolve.

To learn more about Short-Term Power Forecasting Solution, visit the page at

Read more about how we reduced Causer Pay Charges by 26% in Australia. Download Report

How Infigen Energy improved Grid Stability and Revenue Flow with Energy Analytics for Short-Term Forecasting News

Source link

How Infigen Energy improved Grid Stability and Revenue Flow with Energy Analytics for Short-Term Forecasting Resources

Renewable Energy