Case studies · Energy and power grids

A single view of the grid, in near real time

How Rebtech built a modern data platform that gave a mid-sized grid operator better network planning and control.

When the new grid meters began delivering more frequent readings and more types of measurement, the company had no way to follow the network in a single view. Rebtech evaluated the platforms and, in six months, built a solution that made consumption, production and capacity visible in near real time.

Client: A mid-sized Swedish power grid operator
Industry: Energy and power grids
Engagement: Evaluation and implementation of a new data platform
Duration: Roughly six months
Roles: Data architect, senior data engineer
Tech stack: Microsoft Fabric, dbt Cloud, Power BI
Geography: Sweden
Method: Traditional development

The challenge

Follow-up happened system by system

As part of its digitalisation journey, the company identified data and analytics as an area with clear gaps in both skills and tooling. At the same time, every grid meter had been replaced with new meters that allowed more frequent measurement and more types of readings.

Follow-up on the network, however, still happened system by system, or was produced manually by analysts. There was no platform giving a single view, which made network planning and bottleneck control unnecessarily hard.

The solution

Three platforms compared, one chosen

The company engaged Rebtech to evaluate platforms and tools and to implement a new data platform. The team consisted of a data architect and a senior data engineer.

Azure, Fabric and Databricks were compared first. The choice fell on a solution built on Microsoft Fabric, dbt Cloud and Power BI, a combination that matched the company's need to handle large volumes of event data.

Delivery

Six months from evaluation to production

Over a six-month period, Rebtech built a new data platform where readings from customer and grid meters could be followed in near real time. In Microsoft Fabric, the build used among other things:

Lakehouse and warehouse for data storage
Eventstream and KQL for streaming measurement data
Data pipelines for processing
Efficient Python code in notebooks for larger volumes of event data

Results

A single view of the grid, for the first time

For the first time, the company had a single view of its power grid in near real time. That makes it possible to follow capacity, consumption and production continuously.

In brief