| Title |
Development of a Model Composting Process for Food Waste in an Island Community and Use of Machine Learning Models to Predict its Performance |
| ID_Doc |
17975 |
| Authors |
Lytras, C; Lyberatos, V; Lytras, G; Papadopoulou, K; Vlysidis, A; Lyberatos, G |
| Title |
Development of a Model Composting Process for Food Waste in an Island Community and Use of Machine Learning Models to Predict its Performance |
| Year |
2024 |
| Published |
|
| DOI |
10.1007/s12649-024-02697-9 |
| Abstract |
PurposeA novel composting process suitable for handling food waste in an island community is developed. Food waste collection exhibits substantial variation in quantities over the year and is based on the separate disposal of food waste by residents and shops at the source.MethodsThe food waste is properly mixed with recycled compost and bulking material, consisting of a mixture of prunings, leaves and sawdust, and placed in one of 24 1 m3 closed containers. After approximately one month, it is transferred to a second 1m3 container and after one more month to a final 1 m3 container. It is then sieved using an automated sieve and separated into two fractions. The net product is bagged and returned to the citizens, while the "reject" fraction is mixed with the feed. Approximately 15 tons of food waste were processed in 44 batches, producing approximately 3.4 tons of compost product. During the composting batches, the composting mixture temperature and volume were monitored, the mixture was stirred 2-3 times weekly, and water was added as needed to maintain a good level of moisture. For each of the 44 batches, the mean and maximum temperature reached, the mean ambient temperature, the process duration and the amounts of crude and net compost obtained after sieving are presented.ResultsFive machine learning models (Linear Regression, Decision tree regressor, K-Neighbors Regressor, Support Vector Regression, XGBoost Regression) were developed to predict these outputs, using ambient temperature, mixture amount and mixture composition as inputs, with excellent results.ConclusionThe developed machine learning models are effective for predicting the outcome of composting food waste and could be used to optimise the design and operation of composting plants. |
| Author Keywords |
Food waste; Composting; Artificial intelligence; Machine learning |
| Index Keywords |
Index Keywords |
| Document Type |
Other |
| Open Access |
Open Access |
| Source |
Science Citation Index Expanded (SCI-EXPANDED) |
| EID |
WOS:001291942900001 |
| WoS Category |
Environmental Sciences |
| Research Area |
Environmental Sciences & Ecology |
| PDF |
|