Graph Databases in Energy Management: How Volue Transformed the Power Grid with Memgraph

By Sara Tilly

4 min readFebruary 3, 2025

In this week’s Memgraph Community Call, Katarina from Developer Experience join our guests from Volue - Inge Hegge and Dinesh Mathur. We discuss how Volue, a Norwegian company specializing in power grid data, integrated Memgraph into their Spark project to improve power grid management. They needed the right database—a graph database capable of handling the complex, highly connected structure of power networks.

The webinar covers why they chose a graph database, how Memgraph fits into their architecture, and their experience with it, including performance insights.

Watch the full webinar recording—How Volue Transformed Power Grid Management with Memgraph.

Introduction

Power grids were never designed for the modern world's energy demands. Originally built to power homes with basic appliances, today’s grids must support electric vehicles, solar panels, and high-power industrial applications without collapsing under the strain.

Traditionally, expensive infrastructure upgrades were the only way to keep up—installing new substations, transformers, and power lines to meet rising demand. But what if there was a more innovative way to manage power distribution without ripping up the grid?

And that’s what we see in this Community Call, how Volue solved this problem. A key part of the talk was about how Memgraph enables dynamic power load balancing by modeling power networks as graphs, allowing fast traversals to detect bottlenecks, predict overloads, and optimize energy distribution in neighborhoods.

Talking Point 1: Power Grid Overload & Costly Infrastructure Upgrades

Power grids were originally built for home appliances but now handle electric cars and high-power consumption devices. Traditional solutions lead to expensive infrastructure upgrades (new substations, power lines, etc.).

Inge and Dinesh explain how Volue uses data-driven insights to optimize power distribution without costly physical upgrades.

Talking Point 2: How Volue Uses Memgraph to Optimize the Power Grid

Memgraph helps model the power grid as a graph because it enables real-time monitoring and predictive analytics.

In the case of Volue, we have the so-called neighborhood-based energy balancing—users get incentives for shifting power usage to off-peak hours. With real-time analysis, you identify grid stress points and adjust consumption dynamically. Then, with machine learning forecasting, you predict load demand and can compensate energy users accordingly.

Talking Point 3: Why Traditional NIS Systems Were Not Enough

As you’ll see in the presentation, Dinesh and Inge compare two systems. Old systems proved to be slow and rigid and required costly reconfigurations when updating network data. The network topology in traditional databases is geometric-based and requires manual rebuilding when modified. Updating one part of the grid disrupted the entire system. For example, a minor network change could take hours to days to propagate in an NIS system.

It’s great that Memgraph models the power grid dynamically, eliminating the need for slow topology reconstructions.

Talking Point 4: Why Use Graph Databases?

Power grids are naturally graph-shaped networks; substations connect to transformers, and transformers distribute electricity to homes. The relationships between these elements matter more than the individual data points—which is why a graph database made perfect sense.

Initially, Volue experimented with JanusGraph + Gremlin, but this solution had a few drawbacks. Then they switched to Memgraph. Queries that took months to write in Gremlin were rewritten in an hour. Memgraph Lab proved to be the best visualization tool, making power grid analysis easier.

Talking Point 5: Deployment & Performance

Dinesh and Inge explain how Memgraph deployment was easy using one Docker container instead of multiple dependencies in Gremlin.

With Memgraph, they’ve got faster queries. Here are the numbers for comparison:

Downstream trace (power flow analysis) in milliseconds vs. ~2 seconds in NIS.
Connected trace (finding all linked nodes) is faster in Memgraph.
Memgraph’s in-memory architecture eliminates disk latency, making it significantly faster than Neo4j (disk-based) and JanusGraph.

Talking Point 6: Support & Collaboration with Memgraph

One of the most significant advantages Volue saw with Memgraph was great support and collaboration, direct communication via Slack and Discord, and fast responses, often within minutes. They also mentioned regular roadmap discussions to align Memgraph features with what Volue needed.

Conclusion

By adopting Memgraph, Volue transformed its power grid management system from a slow, rigid setup into a real-time, high-performance solution. Their success story demonstrates the power of graph databases in energy and utility management with intelligent power distribution.