Dela med andra

Real-time characterization of material flows for optimal operation of combined heat and powerplants, waste management facilities and sustainable use of resources within circular economy

Publicerad 2020-11-05

Om uppdragsgivaren
Mälardalen University MDH is with 16 000 students one of Sweden’s large Higher Education Institutions. MDH is characterized by close partnerships with industry and the public sector. Engineering research at MDH is particularly strong.

Beskrivning av examensarbetet
Characterization of material flows within energy and environmental systems presents an important challenge. Biomass and waste processing industries as well as wastewater treatment plants show large interest in improving process control and performance optimization to move towards a more environmentally sustainable production.
Today´s world is characterized by ever-growing energy and material demands resulting in production of large amounts of wastes in various forms. A sustainable waste management approach is required to address this environmental threat. According to the European waste management hierarchy waste re-use, recycle and energy recovery is strongly preferred over waste disposal. Components of the waste such as plastics can be sorted out and reused or recycled to increase their useful lifetime. To assist waste recycling and energy recovery, technologies that classify waste components for automated on-line sorting are required. Moreover, biomass originating from forest residues as well as from other sources is material with highly variable properties that makes its utilization in energy conversion processes complex as it creates undesired process instabilities. Therefore, there is a need for sensors that can measure the properties of interest or classify materials in real-time to optimize process operational and regulatory measures.

Proposed tasks
The main objective is to demonstrate the potential of implementing optical spectroscopy sensors for real-time qualitative and quantitative characterization of properties of interest in important material flows for optimal operation of CHPs and waste management facilities and optimal use of resources within circular economy.
Thorough identification of the properties of interest in various material flows i.e. waste, biomass, sludge etc. should be performed based on coproduction with the industrial partners and literature review. To enable optical sensor to measure identified properties of interest or classify materials we need to select and collect samples in a representative manner. Thereafter the spectral data acquisition from samples should be performed with the most suitable spectroscopy technique or combination of techniques. Special emphasis needs to be dedicated to spectral acquisition parameters such as positioning of the sensor, exposure time, acquisition frequency and movement of the scanning table to obtain images with correct aspect ratio. Furthermore, reference values representing property of interest or class need to be obtained by relevant analytical methods. Spectral data need to be then pre-processed to increase signal-to-noise ratio to achieve better results in development of real-time measurements. There are various effective pre-processing algorithms available and their application suitability depends on character of the samples and how the spectral data are acquired. Multivariate regression and classification models are then developed employing machine learning and artificial intelligence techniques by correlating spectral data to the property of interest, which can be quantifiable property for regression model e.g. elemental composition, heating value, content of contaminants etc. or assignment to the class for classification model e.g. identification of plastic component, discrimination if acceptable or not acceptable material etc.. Models are then validated using cross-validation and other independent validation schemes requiring acquisition of new data. The model’s performance validation characteristics then indicate which combination of the modeling technique and spectral pre-processing gives the best result.

Further information:
The proposed thesis is based on activities planned in ongoing research project funded by VEMM companies: Real-time characterization of material flows for optimal operation of combined heat and powerplants, wastewater treatment plants and waste management facilities and sustainable use of resources within circular economy RENAISSANCE

Jan Skvaril
Högskoleplan 1, 722 20 Västerås

Mälardalens University


Sista ansökningdag

2020, 2021

Project description

Bookmark and Share