BMP estimation of landfilled municipal solid waste by multivariate statistical methods using specific waste parameters: case study of a sanitary landfill in Turkey

Creative Commons License

SEL I., ÇAKMAKCI M., Ozkaya B., GURELI F.

JOURNAL OF MATERIAL CYCLES AND WASTE MANAGEMENT, cilt.19, sa.4, ss.1479-1487, 2017 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 19 Sayı: 4
  • Basım Tarihi: 2017
  • Doi Numarası: 10.1007/s10163-016-0543-7
  • Dergi Adı: JOURNAL OF MATERIAL CYCLES AND WASTE MANAGEMENT
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Sayfa Sayıları: ss.1479-1487
  • Anahtar Kelimeler: Biochemical methane potential, Principal component analysis, Multiple linear regression, Waste characterization, Landfilled municipal solid waste, ANAEROBIC-DIGESTION, BY-PRODUCTS, METHANE, GENERATION, PREDICTION, QUALITY, MODEL
  • Yıldız Teknik Üniversitesi Adresli: Evet

Özet

The main objective of this study was to determine whether methane potential of waste could be estimated more easily by a limited number of waste characterization variables. 36 samples were collected from 12 locations and 3 waste depths in order to represent almost all waste ages at the landfill. Actual remaining methane potential of all samples was determined by the biochemical methane potential (BMP) tests. The cumulative methane production of closed landfill (cLF) samples reached 75-125 mL at the end of experiment duration, while the samples from active landfill (aLF) produced in average 216-266 mL methane. The average experimental k and L-0 values of cLF and aLF were determined by non-linear regression using BMP data with first-order kinetic equation as 0.0269 day(-1)-30.38 mL/g dry MSW and 0.0125 day(-1)-102.1 mL/g dry MSW, respectively. The principal component analysis (PCA) was applied to analyze the results for cLF and aLF along with BMP results. Three PCs for the data set were extracted explaining 72.34 % variability. The best MLR model for BMP prediction was determined for seven variables (pH-Cl-TKN-NH4-TOC-LOI-Ca). R-2 and Adj. R-2 values of this best model were determined as 80.4 and 75.3 %, respectively.