Renewable power producer cutting data latency from 12 hours to 15 minutes

EDP Renewables

EDP Renewables is transforming how operational data is accessed and used across its global wind and solar portfolio. By leveraging CONNECT and Databricks to create a unified data environment, the company reduced data latency from 12 hours to 15 minutes, enabling faster access to operational data and supporting advanced analytics across its renewable energy operations.

EDP Renewables

Challenges


Legacy architecture relied on a complex web of custom APIs to push regional telemetry to the cloud

Data update delays of 12 hours forced EDPR teams to make business and operational decisions using outdated operational information

Adding new operational data streams required manual IT configurations that could stretch across weeks

Results


Gauge

Cut data update latencies from 12 hours to 15 minutes, granting teams near-real-time operational visibility, and enabling faster responses to anomalies

Windmill

Accelerated wind-turbine life-extension analysis, enabling work that previously took months to be completed in approximately one week

Enabled teams

Enabled teams to minimize market deviation and optimize pricing by adjusting energy sales based on live plant performance

Data Transfer

Shifted responsibility for CONNECT-based data transfer, synchronization, and recovery to AVEVA, reducing the infrastructure EDPR must maintain internally


In renewable energy generation, a lot can happen over the course of 12 hours. A turbine might start showing signs of a mechanical problem. A change in weather can alter a plant's expected output. Energy prices can move and rewrite the calculus that underpins how and when to sell power. If performance data takes 12 hours to arrive, any decision you make might be based on an outdated picture of your operations.

This is the challenge EDP Renewables (EDPR) faced, compounded by the sheer scale of its operations. EDPR manages more than 23 gigawatts of installed power across nearly 400 wind power plants and 100 solar plants. At each site, thousands of assets generate massive volumes of critical data. Yet much of that data reached EDPR’s analytics teams only twice a day. To unlock the full potential of its global assets, EDPR partnered with AVEVA and Databricks to build a cloud-native data pipeline that continuously transfers operational data through CONNECT and refreshes the Databricks environment every 15 minutes.

Solution

Implemented CONNECT to streamline data flows from regional AVEVA™ PI System™ to Unity Catalog in Databricks through a simplified, cloud-native architecture that continuously transfers data and delivers updates every 15 minutes.

EDPR's legacy architecture EDPR's legacy architecture relied on a complex web of custom APIs, integration services, and servers to move regional telemetry into Azure.

The latency trap

The root of EDPR's data latency troubles lay in its existing data architecture. To move telemetry from Europe and North America to an Azure Data Lake, EDPR relied on numerous connections and custom APIs.

“This architecture is complex,” said Sergio Valencia Galán, Technical Service ROPI Manager at EDPR, who presented the project at AVEVA World 2026. “We have a lot of servers, a lot of points of failure, and the important thing here is our responsibility,” Valencia said. “The responsibility of getting the data to the cloud is ours.”

Although local channel streaming captured continuous, real-time data on-site, getting that data into the cloud remained slow. Data flowing into EDPR’s analytics environment was updated only twice a day.

 

The lag created bottlenecks across the business. Energy teams had less timely information for adjusting pricing and sales decisions as market conditions changed, while maintenance teams had less immediate visibility into anomalies and less current data to inform maintenance planning and asset-life decisions. And on the IT side, adding new data streams or configuring custom API endpoints required manual setup that could stretch across weeks.

In effect, the 12-hour lag left data scientists and energy teams navigating today's changing needs with yesterday's map.


Cloud-based simplification

EDPR set out to create a simpler, faster path from its operational data to the analytics teams that could put it to work. Working with AVEVA and Databricks, it built a new architecture that connects its regional AVEVA™ PI Server™ deployments with the Databricks Data Intelligence Platform through CONNECT.

AVEVA PI System remains the source of EDPR’s industrial data, while CONNECT provides a cloud-based layer for moving that data out of its regional environments and making it available for broader use. From there, Databricks can access the data alongside other information, giving EDPR’s data teams a common environment for analytics, data science, and AI.

EDPR’s modernized architecture EDPR’s modernized architecture uses PI to CONNECT Agent and Delta Sharing to stream live data directly into Databricks every 15 minutes, shifting data transfer management to AVEVA.

The architecture uses PI to CONNECT Agent to transfer data from EDPR’s AVEVA PI Server deployments into CONNECT. Within CONNECT, EDPR can define the data it needs through the data views feature and expose it through virtual tables. Databricks then accesses those tables through Delta Sharing, allowing the data to flow into its analytics environment without requiring EDPR to build a new custom connection for every use case.

The result is a much simpler path for getting industrial data where it needs to go: AVEVA PI Server to CONNECT, and from CONNECT to Databricks. Instead of relying on a web of connections and manual data-transfer processes, EDPR has a more scalable way to make its operational data available to the people and platforms that need it.

“The same process I showed you before, with all this complexity, we reduce to this. And most importantly, all of this,” said Valencia, gesturing at a diagram on screen of the new, cloud-native architecture, “is not my responsibility. It’s AVEVA’s responsibility. I have to configure the data, yes, but the responsibility to move the data—to Databricks in this case—is AVEVA’s.”


“Most importantly, all of this is not my responsibility. It’s AVEVA’s responsibility. I have to configure the data, yes, but the responsibility to move the data—to Databricks in this case—is AVEVA’s.”

—Sergio Valencia Galán, Technical Service ROPI Manager, EDPR


From 12 hours to 15 minutes

The new architecture changes the pace at which EDPR can work with its operational data. Instead of data being updated twice a day, EDPR’s analytics environment is now updated every 15 minutes, which gives teams a near-real-time view of what is happening across the fleet. This enables faster responses when anomalies emerge and more timely information for adjusting energy prices and reducing deviations from market expectations.

CONNECT also makes it much easier for EDPR to securely share data with internal users and third parties, without building a custom connection for every new need. And faster access to data can also accelerate the analysis built on top of it. Wind-turbine life-extension analysis is one example. The work requires EDPR to analyze decades of operational data to understand the loads turbines experience and determine whether assets should be extended, maintained, or replaced. “We can do in one week something that, before, maybe months,” said Valencia.


“We can do in one week something that, before, maybe months.”

—Sergio Valencia Galán, Technical Service ROPI Manager, EDPR


Faster data, new possibilities

The faster data can move, and the easier it is to access, the more valuable it becomes. Near-real-time access opens the door to new applications, including machine-learning models for predictive maintenance and battery energy storage strategies designed to improve performance and extend asset life.

For EDPR, the new architecture is a foundation for using the enormous volume of operational data generated across its global fleet in ways that simply weren't practical when that data arrived 12 hours late.

Product highlights


AVEVA™ PI System™

Collect, aggregate, and enrich real-time operations data for immediate problem-solving and easily deliver formatted data to enterprise applications and advanced analytics.

CONNECT

Our industrial intelligence platform: securely access the broadest and deepest industrial software-as-a-service (SaaS) portfolio enhanced by the power of industrial Artificial Intelligence (AI).

Related success stories

Stay in the know: Keep up to date on the latest happenings around the industry.

 

Contact AVEVA
Live Chat
Schedule Demo