Figure 6 Relative deviation plot for the proposed MLP-ANN.
Related Figures (7)
Figure 1. The schematic diagram of MLP-ANN algorithm. This investigation is a purely modeling study; so for the training and validation of predicting model, the needed experimental data were extracted from reliable paper. For this purpose, a total number of 78 HHV data points which contain HHV in terms of carbon, hydrogen, oxygen, sulfur, and nitrogen Figure 2. The brief summary of this investigation. the accuracy of the predicting model. Another approach for comparison of predicted and experimental HHVs is the simultaneous plot of predicted and experimental values against the index number as shown in Figure 5 which illustrates the great agreement of predicted values and experimental HHVs. Also, for better clarification of the performance of MLP-ANN, the relative deviation between MLP-ANN outputs and experimental HHVs is shown in Figure 6. The relative deviations lie close to the y = 0 straight line which expresses the accuracv of the predicting model. Table 1. Optimum weight and bias values. Table 2. Statistical analyses obtained for MLP-ANN model. Conclusions Figure 3. Levenberg Marquardt algorithm performance in different iterations. Figure 5. Simultaneous plot of predicted and experimental values against the index number
