Modelling Of Biogas Production From Banana Stem Waste With Neural Networks Learning Strategies To Optimse The Production
Biogas production from waste is a valuable renewable energy and with better process design, maximum biogas yield can be obtained from the same amount of waste. Modelling and optimisation are widely used in biological and chemical process domain to increase and to improve the efficiency of this proce...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
JATIT
2017
|
| Subjects: | |
| Online Access: | http://umpir.ump.edu.my/id/eprint/17428/ http://umpir.ump.edu.my/id/eprint/17428/ http://umpir.ump.edu.my/id/eprint/17428/1/Modelling%20of%20Biogas%20Production%20from%20Banana%20Stem.pdf |
| Summary: | Biogas production from waste is a valuable renewable energy and with better process design, maximum biogas yield can be obtained from the same amount of waste. Modelling and optimisation are widely used in biological and chemical process domain to increase and to improve the efficiency of this process. In recent years, intelligence computation is applied to design a better process model and optimised biogas yield. This paper presents a comparative study of several neural networks learning (back-propagation, resilient propagation, Lavenberg-Marquardt and particle swarm optimisation) algorithms for process modelling and optimisation and its relation with the optimisation result. The result shows an improvement of around 10% of biogas production and 8% more from the engineering mathematical optimisation. Two main complications were identified, first one is the high accuracy modelling is not a guarantee for optimised production and the second is a false solution with high optimised production may happen. To clarify this situation, a solution is suggested using factor deviation percentage. |
|---|