Everyone has a theory about that line. Nobody has ever proven one.
We are a Dutch engineering and data company from the Brainport region. We combine data analysis with the physics of machines and structures, and we grew up offshore, where nobody can just walk over and take a look.
We find the same four things in almost every plant
The line that stops
It goes down for no clear reason, usually at the worst possible moment, and the explanation changes depending on who you ask.
The reject rate after a changeover
It climbs after a product change and nobody can say precisely why, so the recipe gets adjusted by feel and the cause stays where it is.
The machine with a season
It behaves differently in summer than in winter. Everyone knows this. It has never been quantified, so it cannot be designed out.
Alarms nobody trusts
Enough false alerts that the service team has learned to ignore them, which quietly removes the value of every true one.
The gap is not knowledge. It is attention, and who sits in the middle.
Inside your organisation
- Your people have no time. The line has to run, and they are pulled back to it every day.
- The knowledge lives in people's heads rather than on paper.
- Every investigation ends with "let us keep an eye on it".
- The data is there. Nobody ever gets a clear run at it.
Out in the market
- IT firms want to build a data platform first and talk in years.
- Sensor vendors sell a box that answers only its own questions.
- Your systems integrator knows the machine but does not do data analysis.
- Almost nobody sits in between. That is the space we work in.
AI anchored in physics
Valuable machines rarely produce failure data. The events that matter most are, by definition, the rare ones, which is exactly where a purely data-driven model has least to go on.
So we let the physics supply the fundamental behaviour and ask your data only for what is specific to your situation. Predictions stay realistic outside the range you happen to have measured, results are explainable in engineering terms rather than as a black box, and you get value without first building years of history.
Physics for the fundamentals. Your data for the specifics.
The physics of a gearbox at sea is the same as the physics of a gearbox in your plant. What we learned in an environment where nobody can go and look works more easily on a machine standing on your own floor.
Vibration, thermal behaviour, loads, and what is mechanically possible at all.
Control system, historian, MES, maintenance records, order and shift data.
Which means they can act on it. A correlation without a mechanism is a guess with numbers attached.
Two ways in, both small and bounded
Each has a fixed scope and a fixed duration, and each stands on its own. Neither one is a platform you have to adopt.
Tarucca Edge
A reliability score for every sensor channel in your installed fleet, with a readable cause code behind every flag. A vibration peak can mean a real machine problem or a sensor problem, and without context the two look identical in the data. Statistical checks, physics-aware consistency and drift detection tell them apart. Software only: no new hardware and no change to your machines.
Tarucca Edge ›Plant Data Unlock
One line, one expensive question, one proven cause. We connect to your control system, historian or MES, and work until there is a technical explanation your own engineers can repeat back and therefore fix. You keep the data connection, the dashboard, the report and the scripts, whichever way the answer goes.
Plant Data Unlock ›Four specialisms
Data analysis, data engineering, dashboards, machine learning and process optimisation are the ground we share with every competent data partner. These four are where we go further.
Analysis and triggers from data
Distilling information and actionable triggers out of sensor and operational data, with data science and machine learning that respects the mechanism underneath.
Health of sensors and measurements
Monitoring whether the measurement itself is still sound, which is the layer almost everyone skips and the reason so many alarms cannot be trusted.
End-to-end data pipelines
Building and running high-volume pipelines from local sensors all the way to AI-generated signals in the cloud, with one party responsible for the whole chain.
Edge AI
Local analysis where connections are interrupted or bandwidth is scarce, with cloud compute where it makes sense, and a clean division between the two.
A short, structured route
The aim is a prioritised set of opportunities and one worked-out first project, rather than a strategy document.
Scope and data inventory
Which assets, which systems, what data actually exists and at what resolution it has been stored.
Physics and data session
Half a day with your production, maintenance and quality people. This is where the real question gets sharpened.
Feasibility and design
An honest read on whether the answer is in the data, and if so, what the first project should be.
First project and roadmap
The work itself, plus what the next two or three applications would be worth and in what order.
Honest about references
Our proven field work is in offshore wind and infrastructure, not yet in food, semicon or high-tech manufacturing. What transfers is the method: explaining vibration, thermal behaviour and loads out of measurement data is the same physics wherever the machine stands. We are deliberately looking for our first clients in this sector, which shows in the terms and in the attention you get. We would rather you judge us on two weeks of your data than on a list of names.
Tell us the problem you have not cracked in two years.
One hour is all we ask. At the end of it we decide together whether there is a concrete question worth investigating, and if there is not, we will say so. That is a good outcome too, and it costs you an hour rather than a project.
