Success Stories

From Challenges to Working Solutions

Find out how our AI and data solutions and services have helped our customers solve problems and generate value for their organizations.

Client’s needs

Automated calculation of positional configurations of indoor satellites for TRUMPF’s real-time locating systems, allowing for full coverage with the least amount of sensors.

Satellite positioning for indoor localization

#SmartFactory #Optimization #ILP #Heuristics

Our Approach
  • 3D estimation of the factory floor with all relevant obstacles
  • Utilizing a heuristic approach to calculate a close-to-optimal solution, thus reducing complexity
  • Full integration with customer’s frontend application
Our Key Results
  • Guaranteeing full coverage of the area of interest
  • Reduce computation time by 93% compared to existing solution
  • Reduced number of required satellites by up to 30% compared to the existing solution
  • More efficient and competitive product
Client’s needs

Gain operational insights regarding electricity consumption and production in the low voltage grid by automatically analyzing time series sensor data from transformer stations.

AI-based analysis of low-voltage grid data

#TimeSeries #DataAnalytics #ElectricVehicles #Photovoltaics #Python

Our Approach
  • Clean, enrich, and prepare sensor data for analysis 
  • Building and training machine learning models for detecting charging cycles of electric vehicles (EVs) and estimating the photovoltaic energy generation
Our Key Results
  • Confident detection of EV charging cycles
  • Estimation of effective power generation from PV systems
  • Detection of patterns, anomalies, and large consumers
  • Improved decision-making for operations and planning
Client’s needs

Development of an AI digital twin, that uses sensor data to optimize the operation of large heat networks and thus reduces heat losses in the systems.

District heating network optimization

#CognitiveTwin #AI #Optimization #Python #AWS

Our Approach
  • Physical modeling and simulation of the heat network based on available network and sensor data
  • Energy demand forecasting for all relevant points in the heat grid
  • Flow optimization to meet consumer demand and reduce heat losses at the same time
Our Key Results
  • AI digital twin deployed in AWS cloud
  • Simulation, forecasting and optimization are executed once per minute with streaming data
  • 15% reduction in heat losses
  • 10% lower CO2 emissions
Client’s needs

Automated approach for generating and maintaining complex country-specific product catalogs to reduce manual processes and risk of invalid product data.

Automated data generation and validation for a CAD/BIM library

#ProcessAutomation #QualityAssurance #Python #DataWrangling #CADBIMProductData

Our Approach
  • Data consolidation from different internal sources and data formats
  • Extracting product data for individual countries
  • Generate complete catalogues
Our Key Results
  • End-to-end process automation
  • Automated quality assurance
  • Ensure consistency of global product catalogs
Client’s needs

Development of a novel approach for electricity grid planning in order to create optimized solutions and reduce manual input and modeling.

AI-based grid planning

#Optimization #AI #ProcessAutomation #LowVoltageGrid #GCP

Our Approach
  • Combine heuristics and reinforcement learning 
  • Optimize positioning of transformer stations
  • Optimize the connection between energy consumers and producers
Our Key Results
  • Reduction of computational time compared to existing methods.
  • Up to 30% capex reduction compared to existing network layouts
  • Deployed as a REST-API in GCP
  • Presented in CIRED 2019 conference in Madrid