Building a unified view of facilities for a renewable energy company

Energy companies face a complex challenge: monitoring and managing facilities spread across vast distances, where a single maintenance issue can impact thousands of users.

ju:niz, a German renewable energy company, is tackling this challenge head-on, revolutionizing how distributed energy facilities are monitored and managed. Using advanced stream processing and real-time analytics, they're transforming traditional plant monitoring into an intelligent, data-driven operation that ensures reliable, sustainable energy delivery.

The opportunity

Growing pains in monitoring distributed renewable energy facilities

ju:niz is focused on sustainable real estate as well as the renewable production, storage, conversion and economical use of energy. As the company grew to 25 plants generating a total of 260 megawatt-hours of capacity, it became increasingly challenging to monitor and manage all of its facilities effectively:

• Low-resolution plant data: ju:niz initially relied on traditional plant monitoring systems that provided limited data and lacked granularity, providing only high-level status information and making it difficult to pinpoint specific issues.
• Limited data insights: ju:niz deployed InfluxDB v1 databases in each plant, however, they only captured data every 5 seconds, which is insufficient for in-depth plant and sub-systems analysis.
• Inaccessible data: The InfluxDB v1 databases were installed at the edge, making it difficult to access the data to monitor and manage the plants. ju:niz had to dispatch an engineer to check issues when they occurred.

“We have one plant with five different types of machines: inverters, climate control systems, batteries and control systems. They all have statuses, but the plant monitoring system hides those details because it can't handle them. This leads to vague alerts: You have 20 batteries and one died. Which one? We don’t know.”

‍

Ricardo, Head of IT Infrastructure and Security, joined the company in September 2022 with a vision of implementing a more data-driven future for ju:niz. He started to build a centralized monitoring system by migrating data from InfluxDB v1 edge instances to InfluxDB v2 in the cloud, and subsequently became an early adopter of InfluxDB v3 when the more performant variant was introduced. Ricardo’s migration journey came with considerable difficulty:

• Unreliable data transfer: Legacy plants had custom Python scripts (used to transfer data from edge devices to the cloud) that were unreliable and inefficient.
• InfluxDB v1 to v2 and v3 incompatibilities: Retention policies per bucket and data type variations per field prevented data migration to newer InfluxDB versions.
• Line Protocol issues: ju:niz relied on InfluxDB Line Protocol files for data migration, but manual modifications were required due to compatibility issues and data type conversions. This proved too cumbersome to be practicable. 
• InfluxDB v3 early adoption challenges: The lack of a monitoring system and rate limits (~ 2MB/s) for the v3 Cloud cluster made it difficult to optimize line protocol file ingestion without risking cluster overload.
• Data movement complexity: Moving large amounts of data between buckets, particularly for development and testing, was impossible without dedicated tools.

“Moving data around is also not easy. Now we collect up to 400 GB of compressed Influx files per day on the biggest plants. Moving that around with scripts was out of the question.”

The solution

A centralized monitoring system built with Quix and InfluxDB v3

To realize his vision, Ricardo needed a more professional approach to data migration and management. Ricardo chose Quix to integrate data from all the sources into a single InfluxDB v3 Cloud production bucket. With Quix, Ricardo built a solution that:

• Integrates data from all instances of InfluxDB v1 and v2, as well as various new plant sources using Telegraf, into InfluxDB v3.
  
• Reprocesses data in InfluxDB v3 from Dev and Test buckets into the single prod bucket to provide a single source of truth.
• Centralizes all device monitoring data into a single InfluxDB v3 bucket. This enables comprehensive monitoring and the creation of device-type specific alert rules, significantly reducing management overhead.

“I faced complicated issues until I found Quix. It just works – so much more professional than before.” 

Quix helped Ricardo solve a number of technical challenges when building this solution:

• Simplified data integration: Quix Connectors, templates and development workflows simplify data movement between complex systems.
• Easy data preparation: Quix pre-processing allowed Ricardo to integrate data from all instances and versions of InfluxDB by re-structuring the data before sinking it to v3
• Performance control: Quix enables precise control over data ingestion rates and provides backpressure support, preventing cluster overload and ensuring stable performance.
• Efficient data pipelines: Quix efficient infrastructure allows Ricardo to manage and continue collecting more data points per second. With enhanced data granularity, ju:niz gains deeper insights into plant operations. They can now monitor individual components, such as battery cells, for improved fault detection and analysis.

"Quix saves me a lot of time. I'm not a full-time programmer so I'm using Quix templates and AI to do everything - it's helping me a lot actually."

The outcome

A transformed data infrastructure delivering real business value

ju:niz has achieved a significant transformation in its data operations now that Ricardo’s vision of implementing a more data-driven future is a reality. The company moved from a limited and unreliable data infrastructure to a robust and scalable system that provides rich insights into plant operations. This data-driven approach enables ju:niz to optimize performance, enhance maintenance practices, and make more informed decisions. Specifically, the centralized plant monitoring system built with Quix and InfluxDB v3 has:

• Enhanced data granularity: increased data collection from having a total database size of 20GB in Feb 2023, to collecting 1,000GB of data every single day. This provides a wealth of detailed information for analysis.
• Cost savings: by using Quix to migrate to InfluxDB v3, ju:niz reduced data storage costs by a factor of 100, enabling them to collect 30TB of data per month for the same cost as collecting 300GB per month.
• Improved monitoring and alerting: The centralized monitoring system built with Quix enables comprehensive monitoring with granular insights and simplified alert management. For example, alerts are now specific to the exact cells in the batteries.
• Improved scalability and flexibility: Quix allows ju:niz to easily scale its data pipeline and adapt to future growth and changing requirements.
• Streamlined development: Quix connectors, templates and its user-friendly interface simplify development, even for those without extensive programming experience.

Ricardo’s success story at ju:niz demonstrates how Quix empowers energy and manufacturing companies to harness the power of their plant data for improved efficiency and better business outcomes. To learn more, contact us https://quix.io/book-a-demo.