Since forecasting renewable energy production depends on predicting tomorrow’s wind, sunlight, temperatures and other factors, IPPs and power utilities need a way to gather wind, sun and temperature forecast data where it is available. With digital technologies and advanced energy data analytics, weather forecast data can be used to accurately predict clean energy production.
For IPPs who sell renewables to utilities, this forecasting data is critical to their success. Until recently, utility companies purchased renewable energy when the energy was produced. Now, utilities are under pressure to provide assurances they can supply the power to customers in a timely and cost-effective manner. As a result, utilities have mandated that IPPs tell them exactly how much energy they are going to sell to them in 15-minute intervals. If IPPs produce less than reported, they are subject to penalties. If they produce too much, the utilities are not obligated to buy the excess.
To help IPPs forecast and deliver consistent power to utilities, they must also use grid-level batteries to store excess energy when it is created. With energy storage systems (ESS) such as those from Hitachi, IPPs can make up for forecasting errors, and store unsold energy to resell later at a better prices.
Hitachi invests more than $4 billion annually in research and development (R&D), focused on exploratory research (chemistry, physics, nuclear science, etc.), technology innovation (evolutionary research and product development) and Hitachi Global Center for Social Innovation (GCSI). GCSI works closely with organizations such as IPPs to solve important energy challenges that impact society.
Because renewable energy production and forecasting is so important to meeting our energy needs in the U.S., Hitachi identified energy forecasting as an essential part of a renewable energy solution. Working closely with one of our customers, we are modeling the terrain of a wind turbine farm, as well as the age of turbines, humidity, operators, efficiency, previous failures and other machine-based information.
When we combine weather data with IoT data coming from the sensors on the turbines, and continuous learning from historical data, we can create models that predict accurate power generation forecasts. By collecting and analyzing the data from a customer’s location, we can help IPPs more accurately forecast renewable energy generation.
Atria Power, one of India's largest IPPs, approached GCSI to create a wind turbine energy forecasting solution for its farms. Atria has multiple wind farms throughout India and uses a variety of wind turbines from different manufacturers. They needed a forecasting solution that would work with the turbines at multiple locations with disparate terrains and weather conditions.
A typical wind turbine farm is considered profitable if it produces 25% of its rated capacity. If a plant is designed for 32 megawatts (MW), then the breakeven point for the plant is 8 MW. To help improve its power generation capabilities when the wind wasn't blowing, Atria placed solar panels at the base of its wind turbines, creating a hybrid solar-wind farm.
Hitachi began by analyzing a constant stream of data from 100 turbines each producing 1MW to 5MW of capacity. We also analyzed data from sensors in the turbines (between 600 to 2,000 sensors per turbine), which measured the amount of power generated, speed and angle of turbine's blade, temperature of components and other detailed machine data. We needed fast, accurate and continuous data collection, so we pulled data from the turbines every 10 seconds.
Once we had data from the sensors, we began to model the entire wind turbine farm, using machine learning techniques to understand the farm's behavior over an entire year, capturing seasonal variations and creating a more accurate forecasting model.
Every 90 minutes, the Hitachi’s wind turbine solution sends Atria detailed energy generation data that forecasts how much energy will be produced over the next 24 hours. The forecast error is required to be within 15%. Hitachi has achieved this goal and we continue to improve. This helps Atria successfully forecast energy generation to utilities and store excess energy in ESS for future delivery. “Hitachi’s understanding of the nuances of forecasting has provided us with insights that we believe will create value as we test out and validate the early results,” said Sunder Raju, Managing Director, Atria Power Corporation. “Even more importantly, the handholding that Hitachi labs have done has ensured the solution provided is a robust solution to our problem statement.”
Bringing renewable energy forecasting to IPPs in the U.S. highlights the collaboration between Hitachi and our customers to co-create new green energy solutions that are benefiting society. Over the past six months, many of these IPPs have visited our Social Innovation Lab to better understand the technologies we’ve developed and how they can improve forecasting and profitability. With increasing reliance of wind energy as a renewable source and near-record growth in deployment in the U.S., the need for energy forecasting is more important than ever.1
Hitachi GCSI is committed to identifying key problems facing society in the U.S. and discovering innovative, real-world solutions to challenges such as wind turbine power forecasting. Armed with decades of experience in operational technology (OT) and IT, as well as technologies such as big data analytics, AI and ML, Hitachi energy forecasting will help improve sustainability and provide safer, cleaner and smarter energy for society.
For more information, visit Hitachi Global Center for Social Innovation
Learn how Hitachi is implementing industry-leading digital solutions to build simplified, sustainable energy management systems to help customers meet their environmental, social, and governance (ESG) goals.
1 AWEA U.S. Wind Industry First Quarter 2017 Quarterly Market Report.
Search
U.S.A.