Your settings

You decide per category. Necessary cookies cannot be switched off, because without them nothing works. Every entry in detail

Always active · 1 entry · 12 months

Optional · 2 entries · max. 13 months

Optional · 3 entries · max. 12 months

Expertise

Telemetry analytics and data integration Four fields

Four fields, one problem: the data is there, just not together.

We work where plants in continuous operation produce readings and nobody can bring them together in a spreadsheet. The fields differ in their physics, not in the task: capture what arises, put it in order in time and in substance, and turn it into a statement that holds in operation.

The way there is the same in all four — from the sensor to the analysis, with no break in the middle.

Our fields

01 / 04

Mechanical engineering

Rotating assets

Speed, vibration, temperature, pressure: rotating machines produce the densest telemetry data in the whole portfolio, often at millisecond intervals and over years. The value does not sit in the single reading but in the curve before it — and that curve is only visible if the history is complete and clean in time.

Which unit is drifting out of the pattern — and since when?

Signals

  • Speed
  • Vibration
  • Bearing temperature
  • Oil pressure
  • Exhaust temperature

Assets

  • Engines
  • Gas turbines
  • Petrol engines
  • Generators
  • Pump systems

02 / 04

Large plants

Process and availability

A plant is not one asset but a network of thousands. We model it as what it is: connected streams whose states explain one another. Only then can a deviation be attributed to a part instead of to a shift.

Which part of the chain is limiting throughput right now?

Signals

  • Flow
  • Pressure
  • Fill level
  • Valve position
  • Analysis values

Assets

  • Oil rigs
  • Gas pipelines
  • Power plants
  • Chemical plants
  • Industrial production
  • Refineries

03 / 04

Renewables

Storage and conversion

Generation fluctuates, storage ages, electrolysis has an efficiency of its own. Compute with daily averages here and you lose exactly the resolution in which degradation and availability become visible. We compute at cell level and keep the time reference.

How much capacity is really left after three years?

Signals

  • Cell voltage
  • State of charge
  • Stack current
  • Irradiance
  • Wind speed

Assets

  • Battery storage, small and large
  • Hydrogen electrolysers
  • Wind farms
  • Solar farms

04 / 04

Fleets

Many identical assets

A fleet is a repetition, and therefore the best data situation there is: what stands out on one unit can be checked immediately on all the others. The hard part is not the model but the connection — moving assets deliver incompletely, late and in bursts.

One-off or a series fault?

Signals

  • Position
  • Load
  • Operating hours
  • Fault codes
  • Consumption

Assets

  • Aircraft
  • Mining equipment
  • Ships
  • Trains

01 / Locations

Konstanz and Berlin

This is where we build from. Konstanz is the headquarters, Berlin the second location — close enough to the plants to drive out when we have to.

02 / Markets

Germany and the whole world

At home in the DACH region, then across Europe, then wherever the plants stand. One data model — what is defined here holds on the British Isles, in Iberia and in the Atacama just the same.

03 / Assets

The data comes from everywhere

Assets like these deliver it: ships and offshore platforms at sea, mining in Australia and South America, solar and grid in Europe, storage in Mexico, rail in Great Britain, aviation at the hubs.

  • Locations2
  • Assetsover 7,000
  • Offshore & marine
  • Mining
  • Solar & grid
  • Rail
  • Battery & storage
  • Aviation

From the sensor to the analysis

  • Capture. Sensors, controllers, historian, third-party systems. We collect the data where it is produced — even when that means a protocol from the nineties or a field device with no second chance to connect.

  • Transport and storage. A stream that does not break, and a store that keeps the raw data. What has once been thrown away is no longer available to the second model — which is why the raw layer is not negotiable here.

  • Model and quality. Assets, signals, units, time reference. A data model in which a value cannot exist without context in the first place, and checks that fire before an analysis computes over gaps.

  • Analysis. Condition monitoring, anomalies, forecasting, reports. At the end there is no demo but something that runs in operation and that the Monday-morning shift trusts.

Partners & technologies

7
  • Databricks
  • Apache Kafka
  • STACKIT
  • Apache Spark
  • Kubernetes
  • Microsoft Azure
  • Hetzner

FAQ

11 questions
How do I start my project with Control-F?
The basis for exceptional projects is robust partnership. Before we start a project, it is essential that we understand your particular needs and goals. Our approach usually involves an initial workshop that serves as the precursor to the main project. That step helps us lay a solid foundation for a successful and collaborative path forward.
How do I know whether Control-F has a solution to my problem?

Working out how we can support you begins with a conversation, with no obligation attached. Getting to know each other helps us explore the ways we might work together.

A workshop tailored to your specific needs is often our first step. In it we gather requirements, run brainstorming sessions or — if your plans are further along — sketch a roadmap for starting a PoC.

How does a project typically run?
It starts with a first conversation and a thorough analysis of what is needed, which forms the basis for a kickoff workshop where the direction of the project and our joint effort are settled. That workshop usually produces a backlog listing the tasks to be implemented over the course of the project. Depending on the project's specific requirements, we either carry the development on under the lead of Control-F's specialists or give our clients' in-house teams targeted methodological and strategic support.
How can I tell whether Control-F is the right partner for me?
Common ground is best found in conversation. This site does give you a first look at our culture and our appetite for innovation, though. Working with our clients as partners matters more to us than short-term gains.
Which data sources can be connected?
PLCs and control systems via OPC UA and Modbus, historians, MES and ERP systems, cloud telemetry and file stores.
Do you also work with legacy infrastructure?
Yes – that is rather the normal case. Plants run for 20 years, software rarely does. We build the bridge instead of demanding a replacement.
How is data security ensured?
Hosting in your own data centre, at STACKIT or at Hetzner – in the EU in every case. Encryption in transit and at rest, role-based access, audit logs.
Do you also offer long-term support?
Yes. Operation and further development are part of the model – we do not hand over and disappear.
What size of company do you work for?
We work for mid-sized firms just as we do for DAX corporations and world market leaders. Since a company's success — and the amount of data it holds — no longer depends on headcount alone, we are in principle a worthwhile partner for any company.
Which industries do you address?
We focus on mechanical and plant engineering and on the energy industry. That is where we are at home. We are happy to leave our comfort zone, though.
Why are you based in Konstanz?
Konstanz combines the best of both worlds for us: proximity to innovative companies in Baden-Württemberg and Switzerland, together with the quality of life on Lake Constance that lets us attract and keep the best people. The region is home to a great many hidden champions and world market leaders in mechanical and plant engineering — exactly the people we work for. At the same time we keep an office in Berlin, and our remote-first way of working leaves us flexible enough to work with clients across the whole DACH region.

Your case

Four fields, one route — but the break sits somewhere different in every plant: a protocol nobody speaks any more, a store that threw the raw data away long ago, a model with no bearing in time. Write and tell us where yours gives way. We will tell you what can be got out of the data you already have — and what is missing for it.

Discuss your case