Probabilistic multi-objective optimization of wood torrefaction conditions using a validated mechanistic model
•Beech wood torrefaction at different conditions using comprehensive instrumentation.•Heat treatment heterogeneity can be assessed by water vapour sorption measurements.•The data-set allowed a complete computational code to be validated.•Torrefaction schedule should be adapted to the particle size a...
Saved in:
| Published in: | Fuel (Guildford) 2023-03, Vol.335, p.126932, Article 126932 |
|---|---|
| Main Authors: | , , , |
| Format: | Article |
| Language: | eng |
| Subjects: | |
| Online Access: | Get full text |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | •Beech wood torrefaction at different conditions using comprehensive instrumentation.•Heat treatment heterogeneity can be assessed by water vapour sorption measurements.•The data-set allowed a complete computational code to be validated.•Torrefaction schedule should be adapted to the particle size and its initial MC.•Treatment conditions meeting users’ expectations found by probabilistic optimization.
This paper uses a comprehensive computational model to propose optimal wood torrefaction conditions by probabilistic optimization. Its main outcome is to propose tailor-made heat treatment conditions (temperature levels-duration of mild pyrolysis at temperature levels ranging from 200 to 300 °C) to meet users’ expectations in terms of overall mass loss, duration and homogeneity of treatment. To this purpose, beech wood boards were torrefied with a usual 3-steps treatment schedule (drying, heating and cooling) under contrasting configurations in a well-instrumented device. The heterogeneity of the treatment within the wood sample was assessed through X-ray attenuation profile and water vapour sorption isotherm. These results allowed the model to be validated. In particular, it predicts the evolution of the mass loss and internal temperatures with good accuracy, including the temperature overshot. The results highlight the need to adjust the heat treatment schedule to each input parameter such as the wood piece dimensions and its initial moisture content or density, in order to limit the effect of exothermic reactions. The torrefaction model was then embedded in a probabilistic optimization process. A case study demonstrates the ability of the model to propose an alternative 3-steps treatment schedule able to reach the target mass loss while controlling the temperature overshot within the wood piece. |
|---|---|
| ISSN: | 0016-2361 1873-7153 |