A hybrid statistical-machine learning approach for experimental analysis of biogas production in a waste to energy plant using a vacuum evaporator systems
Anaerobic digestion offers a promising avenue for sustainable energy generation, waste management while producing biogas. Nonetheless, it yields a liquid by product known as digestate, necessitating treatment to prevent environmental harm and optimize resource utilization. However, improper operatio...
Saved in:
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-09-01
|
| Series: | Results in Engineering |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2590123025025204 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Anaerobic digestion offers a promising avenue for sustainable energy generation, waste management while producing biogas. Nonetheless, it yields a liquid by product known as digestate, necessitating treatment to prevent environmental harm and optimize resource utilization. However, improper operation of these systems can lead to cavitation, which can cause significant equipment damage and lower production of biogas and bio-hydrogen. This study addresses a critical research gap by investigating the influence of key operating parameters as pressure (P), temperature (T), and flow rate (FR) on cavitation phenomena within vacuum evaporators, which significantly impact system durability during large-scale digestate treatment. Optimizing these parameters while mitigating cavitation effects improves energy efficiency, prolonged equipment lifespan, and reliable operation of vacuum evaporators in large-scale biomass digestate treatment systems. k-means machine learning clustering validated by statistical modelling combined methodology is used for this analysis. From the clustering analysis, the optimal setting conditions for cavitation free system came out to be T = 70 ⁰C, P = 85 bar and FR = 10 kg/s. Also, the percentage errors came out to be minimum for machine learning model of around 5.3 %, 5 % and 4.7 % for energy efficiency, equipment durability and process throughput respectively. The model achieved dominance over others with a desirability score of 0.96. Dataset number 8 performed the best as compared to other datasets, falling into the optimal (best) cluster, with energy efficiency of 80 %, equipment durability of 18 months, and process throughput of 200 L/hr. |
|---|---|
| ISSN: | 2590-1230 |