Predicting biodegradable volatile solids degradation profiles in the composting process.

This paper presents a new method for the prediction of the pattern of biodegradable volatile solids (BVS) degradation in the composting process. The procedure is based on a re-arrangement of the heat balance around a composting system to numerically solve for the rate of BVS carbon (BVS-C) disappearance. Input data for the model was obtained from composting experiments conducted in a laboratory-scale, constant temperature difference (CTD) reactor simulating a section of an aerated static pile, and using a simulated feedstock comprising ostrich feed, shredded paper, finished compost and woodchips. These experiments also provided validation data in the form of exit gas CO(2) carbon (CO(2)-C) profiles. The model successfully predicted the generic shape of experimental substrate degradation profiles obtained from CO(2) measurements, but under the conditions and assumptions of the experiment, the profiles were quantitatively different, giving an over-estimate of BVS-C. Both measured CO(2)-C and predicted BVS-C profiles were moderately to well fitted by a single exponential function, with replicated rate coefficient values of 0.08 and 0.09 d(-1), and 0.06 and 0.07 d(-1), respectively. In order to explore the underlying shape of the profiles, measured and predicted data at varying temperature were corrected to a constant temperature of 40 degrees C, using the temperature correction function of Rosso et al. [Rosso, L., Lobry, J.R., and Flandrois, J.P., 1993. An unexpected correlation between cardinal temperatures of microbial growth highlighted by a new model. Journal of Theoretical Biology, 162, 447-463], with cardinal temperatures of 5, 59 and 85 degrees C. Multi-phase profiles were generated for both the measured CO(2)-C and the predicted BVS-C data in this case. However, when alternative cardinal temperatures of 5, 55 and 80 degrees C, or 5, 50 and 80 degrees C, were used, the predicted profiles assumed an exponential shape, and excellent fits were obtained using a double exponential function. These findings support the argument that a substrate degradation curve generated under laboratory conditions at 40 degrees C, would, given correct cardinal temperatures, generate a correct substrate degradation profile under varying temperature conditions and that this in turn would enable an accurate and precise prediction of the temperature profile, using a heat and mass balance approach. In order to realise this prospect, it is proposed that further work to obtain experimental data under completely mixed conditions, more accurately estimate the overall heat transfer coefficient and obtain correct values for the cardinal temperatures used in the temperature correction function, is required.

An evaluation of substrate degradation patterns in the composting process. Part 1: profiles at constant temperature.

This paper examines the patterns of 32 constant temperature substrate degradation profiles obtained from the composting literature, and evaluates the use of a single exponential model, a double exponential model and a non-logarithmic Gompertz model in describing their behaviour. Profiles were found to be predominantly either sigmoidal in shape, or to exhibit multi-phase behaviour, with a relatively small proportion of convex curves. Of the constant temperature profiles, 26 were either not well modelled by any of the above functions, or of such differing profiles that none of the above functions was applicable. Goodness of fit was measured using a normalised error function, and rated using a five-category descriptive scale, ranging from excellent to poor. No fits rated as excellent were observed. Fits rated as good were obtained for three data sets when using a single exponential function, for two data sets when using a double exponential function, and for one data set when using the non-logarithmic Gompertz function. The remainder of the fits were rated as moderate to poor. It is concluded that the evidence supporting the use of the single exponential model, the double exponential model or the non-logarithmic Gompertz model to describe substrate degradation profiles generated at constant temperature is limited. Further work is suggested in order to establish standard procedures and a standard simulated composting mixture for substrate degradation studies and to build a more comprehensive set of long-term substrate degradation profile data at constant temperature, and under non-limiting moisture and oxygen concentration conditions.

An evaluation of substrate degradation patterns in the composting process. Part 2: temperature-corrected profiles.

In this paper, the patterns of 44 substrate degradation profiles obtained from the composting literature are examined following their correction to a constant temperature of 40 degrees C, using a new procedure presented in this work. The applicability of a single exponential model, a double exponential model and a non-logarithmic Gompertz model in describing their behaviour is then evaluated. Multi-phase profiles were most commonly seen, with convex shapes observed in only a relatively small proportion of the profiles. Convex shapes were also embedded within other profiles, either preceeded by a lag phase, or followed by non-convex behaviour. Sigmoidal patterns were relatively rare. Of the temperature-corrected data sets examined, 33 were found to be either not well modelled by, or inappropriate for, any of the above models. Two fits rated as good were obtained when using the single exponential model, and one fit rated as excellent, plus one fit rated as good, were obtained when using the double exponential model. A single fit rated as excellent was found when using the non-logarithmic Gompertz model. The lag phase, which was observed in many data sets, was successfully modelled using the non-logarithmic Gompertz function where excellent and good fits were obtained, but as expected this phase of the profile could not be modelled by either the single or double exponential functions. When the lag phase or post-convex curve data was removed from 20 data sets, use of the single exponential function resulted in three fits rated as excellent and two rated as good. When a double exponential model was applied to these data sets, three fits rated as good were obtained, whilst application of the modified Gompertz model gave one fit rated as good. The remainder of the fits were rated as moderate to fair. It is concluded that the evidence supporting the use of the single exponential model, the double exponential model or the non-logarithmic Gompertz model to describe full substrate degradation profiles in composting following their adjustment for temperature effects is limited. Further work is suggested in order to investigate the nature of those profiles which were not well modelled, to more precisely ascertain the cardinal temperatures for composting used in the function of Rosso et al. (1993) [Rosso, L., Lobry, J.R., Flandrois, J.P., 1993. An unexpected correlation between cardinal temperatures of microbial growth highlighted by a new model. J. Theor. Biol 162, 447-463.], which was employed in the present temperature correction procedure, and to incorporate correction for varying moisture and oxygen concentrations.

Mathematical modelling of the composting process: a review.

In this paper mathematical models of the composting process are examined and their performance evaluated. Mathematical models of the composting process have been derived from both energy and mass balance considerations, with solutions typically derived in time, and in some cases, spatially. Both lumped and distributed parameter models have been reported, with lumped parameter models presently predominating in the literature. Biological energy production functions within the models included first-order, Monod-type or empirical expressions, and these have predicted volatile solids degradation, oxygen consumption or carbon dioxide production, with heat generation derived using heat quotient factors. Rate coefficient correction functions for temperature, moisture, oxygen and/or free air space have been incorporated in a number of the first-order and Monod-type expressions. The most successful models in predicting temperature profiles were those which incorporated either empirical kinetic expressions for volatile solids degradation or CO2 production, or which utilised a first-order model for volatile solids degradation, with empirical corrections for temperature and moisture variations. Models incorporating Monod-type kinetic expressions were less successful. No models were able to predict maximum, average and peak temperatures to within criteria of 5, 2 and 2 degrees C, respectively, or to predict the times to reach peak temperatures to within 8 h. Limitations included the modelling of forced aeration systems only and the generation of temperature validation data for relatively short time periods in relation to those used in full-scale composting practice. Moisture and solids profiles were well predicted by two models, but oxygen and carbon dioxide profiles were generally poorly modelled. Further research to obtain more extensive substrate degradation data, develop improved first-order biological heat production models, investigate mechanistically-based moisture correction factors, explore the role of moisture tension, investigate model performance over thermophilic composting time periods, provide more information on model sensitivity and incorporate natural ventilation aeration expressions into composting process models, is suggested.

Physical modelling of the composting environment: a review. Part 1: Reactor systems.

In this paper, laboratory- and pilot-scale reactors used for investigation of the composting process are described and their characteristics and application reviewed. Reactor types were categorised by the present authors as fixed-temperature, self-heating, controlled temperature difference and controlled heat flux, depending upon the means of management of heat flux through vessel walls. The review indicated that fixed-temperature reactors have significant applications in studying reaction rates and other phenomena, but may self-heat to higher temperatures during the process. Self-heating laboratory-scale reactors, although inexpensive and uncomplicated, were shown to typically suffer from disproportionately large losses through the walls, even with substantial insulation present. At pilot scale, however, even moderately insulated self-heating reactors are able to reproduce wall losses similar to those reported for full-scale systems, and a simple technique for estimation of insulation requirements for self-heating reactors is presented. In contrast, controlled temperature difference and controlled heat flux laboratory reactors can provide spatial temperature differentials similar to those in full-scale systems, and can simulate full-scale wall losses. Surface area to volume ratios, a significant factor in terms of heat loss through vessel walls, were estimated by the present authors at 5.0-88.0m(2)/m(3) for experimental composting reactors and 0.4-3.8m(2)/m(3) for full-scale systems. Non-thermodynamic factors such as compression, sidewall airflow effects, channelling and mixing may affect simulation performance and are discussed. Further work to investigate wall effects in composting reactors, to obtain more data on horizontal temperature profiles and rates of biological heat production, to incorporate compressive effects into experimental reactors and to investigate experimental systems employing natural ventilation is suggested.

Physical modelling of the composting environment: a review. Part 2: Simulation performance.

This paper reviews previously published heat balance data for experimental and full-scale composting reactors, and then presents an evaluation of the simulation performance of laboratory and pilot-scale reactors, using both quantitative and qualitative temperature profile characteristics. The review indicates that laboratory-scale reactors have typically demonstrated markedly different heat balance behaviour in comparison to full-scale systems, with ventilative heat losses of 36-67%, and 70-95% of the total flux, respectively. Similarly, conductive/convective/radiative (CCR) heat losses from laboratory reactors have been reported at 33-62% of the total flux, whereas CCR losses from full-scale composting systems have ranged from 3% to 15% of the total. Full-scale windrow temperature-time profiles from the literature were characterised by the present authors. Areas bounded by the curve and a 40 degrees C baseline (A(40)) exceeded 624 degrees C. days, areas bounded by the curve and a 55 degrees C baseline (A(55)) exceeded 60 degrees C days, and times at 40 and 55 degrees C were >46 days and >24 days, respectively, over periods of 50-74 days. For forced aeration systems at full scale, values of A(40) exceeded 224 degrees C days, values of A(55) exceeded 26 degrees C days, and times at 40 and 55 degrees C were >14 days and >10 days, respectively, over periods of 15-35 days. Values of these four parameters for laboratory-scale reactors were typically considerably lower than for the full-scale systems, although temperature shape characteristics were often similar to those in full-scale profiles. Evaluation of laboratory-, pilot- and full-scale profiles from systems treating the same substrate showed that a laboratory-scale reactor and two pilot-scale reactors operated at comparatively high aeration rates poorly simulated full-scale temperature profiles. However, the curves from two moderately insulated, self-heating, pilot-scale reactors operated at relatively low aeration rates appeared to closely replicate full-scale temperature profiles. The importance of controlling aeration rates and CCR losses is discussed and further work suggested in order to investigate the links between simulation of the composting environment and process performance.