Sludge pyrolysis is a complex process including complicated reaction chemistry, phase transition, and transportation phenomena. To better evaluate the use of syngas, the monitoring and prediction of a higher heating value (HHV) is necessary. This study developed an artificial neural network (ANN) model to predict the HHV of syngas, with the process variables (i.e., sludge type, catalyst type, catalyst amount, pyrolysis temperature, and moisture content) as the inputs. In the first step, through optimizing various sets of parameters, a three-layer network including 8 input neurons, 15 hidden neurons, and 1 output neuron was established. Then, in the second step, an ANN model has been successfully used to predict the HHV of syngas, with a fitting correlation coefficient of 0.97 and a root mean square error (MSE) value of 14.62. The relative influence of input variables showed that the pyrolysis temperature and moisture content were the determining factors that affected the HHV of syngas. The results of optimization experiments showed that when temperature was 895 °C and the moisture content was 45.63 wt%, the highest HHV can be obtained as 438.22 kcal/m3-N. Moreover, the ANN model showed a higher prediction accuracy than other models like multiple linear regression and principal component regression. The model developed in this work may be used to predict the HHV of syngas using conventional operational parameters measured from in situ experiments, thus further providing predictive information for the use of syngas as energy and fuel.
- Mathematical modeling
- Multiple linear regression
- Prediction of energy content
- Principal component regression
- Waste activated sludge