Reliable prediction of biodrying efficiency using interactive regression models

SAYIN F. E., AKMAN G., ÖZBAY B., ÇALLI B., GÖKTAŞ R. K., ÖZBAY İ.

Waste Management, cilt.213, 2026 (SCI-Expanded, Scopus) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 213
  • Basım Tarihi: 2026
  • Doi Numarası: 10.1016/j.wasman.2026.115373
  • Dergi Adı: Waste Management
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, BIOSIS, Compendex, EMBASE, Environment Index, Geobase, INSPEC, MEDLINE, Public Affairs Index
  • Anahtar Kelimeler: Biodegradability, Biodrying, Calorific value, Multivariate, Prediction, Ridge regression
  • Kocaeli Üniversitesi Adresli: Evet

Özet

The inability of municipal solid waste (MSW) to meet incineration standards often undermines the sustainability and economic feasibility of waste-to-energy applications. Biodrying offers a promising, eco-friendly pretreatment to enhance the calorific value of MSW. This study evaluated the performance of biodrying based on the final calorific value (FCV) using simple and interactive regression models. Both conventional parameters; moisture content (MC), bulk density (BD), airflow rate (AFR), and initial calorific value (ICV) and unconventional indicators; the Temperature Index (TI), Biodrying Index (BI), and oxygen consumption (L) as a measure of biodegradability were used as predictors. Besides conventional regression models (OLS), to minimize multicollinearity of the dataset with Variance Inflation Factor (VIF) of higher than 10 Ridge regression (RR) analyses were also applied. AFR was the strongest positive variable in all the tested models and achieved maximum impact in RR3 Model with value of 2189.47 at significance level of p < 0.01. In the same model, triple impact of AFR*TI*MC was strong and negative (−819.60 at p < 0.05). In both regression approaches, interactive models provided better prediction efficiencies considering higher R2 and reduced error metrics. Professionals in this sector may consider the use of RR in FCV predictions to be both an innovative and practical approach.