Smart combustion technology for biomass cogeneration plants
Pre-warning system for slag-free plant operation
In the Fuelband2 project, an early warning system for biomass cogeneration plants is being developed that can detect the formation of deposits live and then propose countermeasures to the operator. For the first time, a »machine learning« approach will be applied to biomass cogeneration plants and tested in real operation. The project is funded by the Federal Ministry of Economics and Energy and runs until mid 2021.
The elimination of the EEG remuneration has drastically changed the economic framework for many power plant operators. In order for plants to continue to be used economically, low-cost but also problematic residual material fractions of different quality must be used. These involve a high risk of slagging, which in turn impairs the availability and overall cost-effectiveness of a plant.
Machine Learning in a biomass power plant
The aim of the project is to develop strategies for reducing the tendency to slagging, especially with problematic fuels. In a project part, an early warning system for slagging will be developed based on a »machine learning« approach. The self-learning system is based on a simulation environment for slagging prediction developed in the preliminary project and should be able to detect deposits live. Then it suggests countermeasures to the plant operator, e.g. an adjustment of the firing parameters. For the first time, "machine learning" and thus an approach of experience-based control strategies will be applied to biomass cogeneration plants.
In addition, various approaches to fuel pre-treatment will be demonstrated and evaluated in the project and possibilities for operating optimization will be demonstrated.
Less slagging with problem fuels
Martin Meiller, Group Leader Energy from Biomass and Waste at the Fraunhofer Institute says: »Our goal is to make plant operation more economical, especially for problematic fuels. The combination of fuel pre-treatment, optimized plant operation and the early warning system seems promising«.
The system integration of the early warning system (software and hardware) is developed by aixprocess GmbH and tested for the first time in real operation in a combined heat and power plant of the Tauberfranken municipal utility. The validation in advance will take place in the Fraunhofer UMSICHT energy and biomass pilot plant in Sulzbach-Rosenberg. The project is coordinated by the Chair of Energy Process Engineering (EVT) of the Friedrich-Alexander-University Erlangen.
Picture project partner
The project consortium in the wood chip warehouse of the natural heat power plant / Tauberfranken public utility in Bad Mergentheim:
Dr. Martin Wenig (Managing Director aixprocess GmbH), Paul Gehrig (Managing Director Stadtwerk Tauberfranken), Dr. Martin Haberbehl (aixprocess GmbH), Prof. Dr.-Ing. Jürgen Karl (FAU Chair of Energy Process Engineering), Martin Meiller (Fraunhofer UMSICHT), Gerhard Hirschlein (Tauberfranken), Thomas Plankenbühler (FAU EVT), Dr.-Ing. Dominik Müller (FAU-EVT).