Artificial neural network modeling of plastic viscosity, yiled point and apparent viscosity for wheat starch solutions

Wheat starch is a carbohydrate polymer which is widely used in drilling muds in order to increase viscosity and decrease fluid loss. Rheological properties ofdrilling muds such as plastic viscosity, yield point and apparent viscosity play an essential role in selecting the most optimum composition of drilling mud underdiverse conditions. In this study, an artificial neural network system was used to predict plastic viscosity (PV), yield point (YP) and apparent viscosity (AV) ofwheat starch solutions. Multi-layer feed forward method was applied in the architecture of the artificial neural network due to its high accuracy. To predict plastic viscosity and yield point, the structure of feed-forward neural network wasdefinite 2:2:1 which refers to input layer, hidden layer and output layer. Thecoefficient of determination (R2) values obtained for training and validation data revealed how this approach is effective in estimating the output layer. The best structure of artificial neural network architecture obtained 3:3:1 for predicting apparent viscosity. R2 value (R2=0.994) of testing data obtained by artificial neural network system revealed the high accuracy of this approach in estimating apparent viscosity.

کلیدواژه ها:

Wheat starch ، Artificial neural network ، Plastic viscosity ، Yield point ، Apparent viscosity