頁籤選單縮合
題 名 | Practical Identification Analysis of Haldane Kinetic Parameters Describing Phenol Biodegradation in Batch Operations |
---|---|
作 者 | Liu, Wen-Der; Lee, Chi-Yuan; | 書刊名 | Journal of Environmental Engineering and Management |
卷 期 | 17:1 2007.01[民96.01] |
頁 次 | 頁71-80 |
分類號 | 445.9 |
關鍵詞 | Phenol; Haldane model; Parameter estimation; Parameter identifiability; Correlation coefficient; |
語 文 | 英文(English) |
英文摘要 | The capability of three techniques, “initial specific growth rates (IGRA)”, “substrate depletion data from a single batch experiment (SDES)”, and “growth rates during incubation in batch experiment (GRIB)”, were evaluated with respect to accurately determining Haldane model parameters with a few batch studies by analyzing the identifiably of Haldane kinetic constants estimated by each of these three methods. Both IGRA and SEDS are conventional techniques; they use the initial specific growth rate from batch degradation assays, and substrate depletion measurements in a single batch degradation experiment, respectively, as the basis for estimating parameters. The GRIB technique, first proposed in this study, uses specific growth rates during incubation. Six batch experiments were conducted, in which both residual phenol and biomass during the period of incubation were measured to generate data to be used as a basis for parameter estimation. Haldane parameters estimated by IGRA method appear to have a high average correlation coefficient (0.950) and significant slenderness of joint confidence regions, indicating that the derived constants are not practically identifiable. The parameters estimated by the SDES method are not identifiable either, because the average parameter correlation coefficients reach 0.994-0.999. However, the GRIB method, when two batch experiments with initial phenol concentrations of 50 and 400 mg□ were used, could give highly identifiable parameters. The best parameters obtained by the GRIB method were maximum specific growth rate □=0.34□, half-velocity concentration □=1.8mg□ and inhibition constant □=340 mg□. The corresponding average correlation coefficient was as low as 0.693 and the slenderness of joint confidence regions were considerably reduced. The findings of this study suggest that GRIB is superior to IGRA and SDES techniques because it is associated with the ability to give parameters having high identifiability. |
本系統中英文摘要資訊取自各篇刊載內容。