Any questions?

Your partner for digitalization and efficiency improvements:

Georg Fischer

Managing Advisor


Boost Efficiency with Machine Learning

The project goal set by gas transmission network operator bayernets GmbH was to develop an optimization tool together with WECOM that enables the identification of the cost-optimal compressor (gas or electricity-powered) in day-to-day business on the basis of the latest information on market prices and transport requirements in order to minimize energy costs.

Client Profile

  • bayernets GmbH
  • gas transmission system operator
  • > 1500 km pipeline network
  • > 100 employees
  • > 10 bcm/a transport volume

bayernets connects almost fifty distribution network operators and numerous large industrial customers to the Bavarian gas network, furthermore bayernets ensures the network connection for important storage facilities in the Bavaria/Austria area and for system-relevant gas power plants and cooperates with other transmission system operators in Germany and Austria.



Market participants should be granted non-discriminatory access to the gas network infrastructure at reasonable tariffs. To this end a regulatory framework was introduced that subjects transmission system operators such as bayernets to a periodic efficiency bechmarking. This shall prevent e.g. unnecessarily high energy costs for generating pressure (through compressor units) in the gas network. In the specific case of bayernets, the cost-optimal operation depends on the decision to use a gas-powered or electricity-powered compressor unit.

Since the costs for operating a compressor unit are subject to a large number of factors such as energy consumption and the prices for energy (including market prices for electricity or gas and CO2 certificates), grid usage fees, maintenance costs and start/stop costs, the cost-optimal dispatching decision is not immediately clear.

The project faced the following challenges:

  • Development of a forecasting model for the actual energy consumption of the respective compressor unit depending on the planned mode of operation
  • Consideration of the variety of variables in the optimization question
  • Consideration of customer-specific complexity in terms of infrastructure, connection situation and grid bottlenecks
  • Holistic consideration of the affected business processes from a legal-regulatory, procedural and IT perspective



In a first step, the target values, variables and constraints of the optimization question were defined in detail and energy consumption billing plus network charges (incl. various levies) were analyzed. A machine learning model was developed to predict energy consumption for arbitrary transport conditions. At the same time, past energy requirements and market price scenarios for electricity, gas and CO2 were analyzed and visualized in order to quantify historical cost savings potentials through optimal use of compressors and to assess simplification potentials through the use of heuristics.



The result of the project is an optimization tool that enables the gas flow dispatching department in its day-to-day business to select the cost-optimal compressor unit, based on current market prices and transport requirements.

The project also created additional value through generating detailed know-how about potential cost savings, relative compressor efficiency and statistical connections between historical market prices (electricity and gas/CO2), gas transport and energy consumption.


  • > 200 k

    technical data points

  • > 25 k

    price data as input

  • > 20

    parameterised compressor diagrams

  • 3

    historical years with daily optimised operation

Back to the news overview

Further Case Studies