Fig. 1. Variation of V_p, V_s , V_p/V_s and Poisson ratio with degree of oil saturation at different temperature regimes of unconsolidated sample measured in a laboratory environment. (a) Variation of V_p, with oil saturation, (b) variation of V_s with oil saturation, (c) variation of Poisson ratio with oil saturation, and (d) variation of V_p/V_s with oil saturation. (*—temperature=24 °C, +—temperature=66 °C, o—temperature=121 °C).

Fig. 2. Schematic diagram of the MLP arranged in layers.

Fig. 3. Comparison of predicted degree of oil saturation using the trained ANN weights and the values from the simulated data. This comparison is used to test the capability of the ANN.

Fig. 4. Comparison of neural-network predicted degree of oil saturations and that from laboratory measurements. Input attributes to the net are V_p, V_s and V_p/V_s.

Fig. 5. Comparison of neural-network predicted degree of oil saturations and that from laboratory measurements. Input attributes to the net are V_p, V_s and Poisson ratio.

Fig. 6. Comparison of neural-network predicted degree of oil saturations and that from laboratory measurements. The only input variable to the net is V_p.

Fig. 7. Comparison of neural-network predicted degree of oil saturations and that from laboratory measurements. The only input variable to the net is V_s.

Table 1. Physical properties of solid components used in the model

Table 2. Petrophysical parameters describing a fluid-saturated medium and the seismic responses