Skip to content

AI consultant in Groningen

For grids, water and buildings in the north of the Netherlands, we start with measurement data. We keep models in production with MLOps, as we do for Waternet.

We set up the MLOps workflow for Waternet and continue to maintain it

As part of Waternet's Datalab team, we set up the workflow with an MLflow model registry and CI/CD templates, deploy code and models to production, and developed a model to minimise nitrous oxide emissions at the wastewater treatment plant.

Read the case study

What can AI do now that Groningen’s electricity grid is at capacity?

In late 2023, TenneT announced grid congestion in the west of the province; applications for new high-capacity connections are now placed on a waiting list. AI cannot expand the grid, but forecasts of electricity consumption and generation help you make better use of the capacity available.

Book a meeting
Windturbines in het vlakke Groningse land

For grids, water and buildings, AI starts with measurement data

Discuss your measurement data

Measurement time series you can trust

Sensor data from stations, pumping stations or installations, checked for gaps, sudden jumps and replaced meters before a model runs on it.

Forecasts within your grid capacity

A model that forecasts consumption and generation at each connection, so your plans stay within the capacity agreed with your grid operator.

Models that stay in production

At Waternet, we set up the MLOps workflow in the Datalab team to bring code and models into production, and we maintain it.

GACS data that does double duty

The data recorded by building automation to meet GACS requirements also helps identify waste. Our Paris Proof platform shows anomalies and efficiency losses for each building.

Pilot with one pumping station, station or building, then expand to the rest

  1. Choose the decision

    For example, when to maintain a pumping station or which building to make more sustainable first. We will measure the results against that decision.

  2. Run the numbers for one asset

    Using the history of one asset, we test whether the measurement data tells us enough to improve that decision.

  3. Roll out and monitor

    If it works, we bring similar assets into the same pipeline, monitoring for drift if a pump or meter is replaced.

Could this work for our water authority, heating company or property portfolio?

Talk it through with us

A good fit if

You manage networks, pumping stations, treatment plants, a district heating network or a portfolio of buildings. You have measurement data or a history of faults, and there is a recurring decision you could improve, such as scheduling maintenance or staying within your network’s capacity.

Not a good fit if

The control system already handles it with a fixed setting, such as a switching time or alarm threshold. Or the measurement data is not coming through reliably: in that case, start with the connection and a dashboard, not a model.

What determines the scope

The number of assets or buildings, how and where the measurement data is currently stored, whether the model affects the safety of supply, and whether your team or ours handles management.

Your data is held by a supplier

Measurement data from SCADA or building management systems can usually be retrieved through a connection or export. We set up the platform on Databricks or Fabric, with data in the open Delta format, so you are not tied to a single platform.

Where can you make the first gains: with assets, building management or customer processes?

With assets

Choose this if faults at pumping stations, treatment plants or other stations are your biggest cost, and you have sensor data and fault histories. Start with predictive maintenance.

With building management

Choose this if you manage non-residential buildings subject to GACS requirements. If you’re unsure, bear in mind that you need to collect this measurement data for GACS anyway. See the Paris Proof platform.

With customer processes

Choose this if requests, fault reports or customer questions remain unresolved, for example about connections to a congested grid. In that case, process automation often helps more than a model.

Water and building energy are already part of our work

Data Science Lab has been operating since 2016, is a Databricks and Microsoft partner, and developed the Paris Proof platform for energy in buildings. We’re based in Amsterdam; we can meet at your office in Groningen, at ours or online.

Book an introductory meeting

Your first conversation is about your assets and measurement data

Stef Stoelwinder

Stef Stoelwinder

Business Development

First, you’ll speak with Stef. You’ll discuss the networks, pumping stations or buildings you manage, the measurement data already available and the decision you want to improve first.

What measurement data do we already have for an initial application?

Briefly describe the assets, buildings or processes involved. We’ll get back to you by email or arrange an initial meeting at your premises, ours or online.

Enter a valid email address.

Let us know how we can help you.

Frequently asked questions

Not on a case study we can share. Our water and energy case studies come from elsewhere, such as the Datalab team at Waternet, which supplies drinking water in and around Amsterdam. What we learn there about models in production applies beyond any one region.

Our only office is on Westerdoksdijk in Amsterdam. For an introductory meeting, we can come to you, you can visit us, or we can meet online.

Making better use of existing capacity is one of the approaches in the province of Groningen’s grid congestion programme, alongside measures such as energy hubs and smart charging. A forecast of consumption and generation for each connection shows when you will stay within your contracted grid capacity and which consumption can be shifted.

To do that, you need sufficiently detailed measurement data for each connection. If you share capacity with other companies in an energy hub, you soon need to address shared data and who manages it: data governance.

Possibly. The AI Act classifies AI systems as high-risk if they are intended to be used as safety components in the supply of water, gas, heating or electricity (Annex III, point 2). The question is whether a failure could make that supply unsafe: a model that controls a valve is more likely to raise that question than one that proposes a schedule.

According to the European Commission, following the digital omnibus, the Annex III rules will apply from 2 December 2027. Your own legal adviser or compliance department must assess how the law applies to your use case. We help prepare the inventory and risk classification and establish AI governance. You can find an overview of the legislation under AI Act.

Usually three sources: sensor time series from SCADA, the system used to monitor and control equipment remotely; fault and maintenance history from the maintenance management system; and master data for each asset, such as its type, year of construction and location. Linking them is often the hardest part: a replacement pump does not always get a new identifier in the measurement data.

Having many comparable assets under one management organisation helps, because a model can then compare patterns across assets. From Groningen, for example, Waterschap Noorderzijlvest manages locks, weirs, pumping stations, flood barriers and treatment plants. We describe the approach on our predictive maintenance page.

Since 1 January 2026, non-residential buildings with a heating or air-conditioning system rated above 290 kW have been required to have a building automation and control system (GACS). According to RVO, the threshold will fall to 70 kW in 2030. Such a system must continuously monitor and analyse energy consumption and identify losses in efficiency.

That measurement data is also useful for energy management: identifying anomalies in each building and deciding which building to make more energy-efficient first. Where the grid is congested, reducing peak demand also frees up capacity. We discussed how to move from compliance to energy management in the Meer dan GACS webinar.

European, national and regional funding, including funding from Nij Begun, will bring a Dutch AI factory to Groningen: an expertise centre and a supercomputer that the Dutch government expects to be fully available in early 2028. The government names energy networks as one of its applications. In future, the waste heat could be used by WarmteStad to heat homes in north-west Groningen.

For most applications in grid management, water and buildings, computing power is not the constraint. The challenges are reliable measurement data, a designated owner for each model and management after launch. You can read about our work in this sector on the data and AI for the energy sector page.

With MLOps: the approach you use to version, test, deploy and monitor models. A model registry and CI/CD templates give every subsequent model the same route into production. This is also how the Datalab team at Waternet has set things up.

In the energy system, data often changes because the physical world changes: a replaced sensor, a new pump or a cold winter. Drift monitoring shows when such changes cause a model’s performance to deviate and retraining is needed. We describe this approach under MLOps.

When a fixed control setting solves the problem, such as a switching time, a heating curve or an alarm threshold. Or when measurement data is not yet coming in reliably: connecting the data and putting a dashboard on top is then the first step, and often shows you where things are going wrong. If you already know what you want to build and only need someone to make sensor data from SCADA and building management systems accessible, consider a seconded data engineer through our secondment service. You can read more about models for water levels, water quality and treatment under AI in water management. If you have defined the problem and want a model built, read more about our data science consultancy for Groningen. If you have chosen an approach and want the application built, take a look at our AI consultancy for Groningen.