ANN-Based sensitivity analysis for the prediction of energy consumption and head rice yield in a pneumatic paddy dryer

Information from the abstract

In this study, the performance of a pneumatic paddy dryer was investigated through a sensitivity analysis with an artificial neural network, in order to identify the relative importance of the process variables affecting the energy consumption and head rice yield (HRY). Experiments were conducted using Thai Jasmine 105 paddy rice, which was dried from 24% to 14% wet basis at drying air temperatures of 60°C, 80°C, and 100°C. Increasing the drying air temperature increased the drying rate from 0.46 to 1.11 kgwater/h. Drying at 60°C resulted in the lowest specific energy consumption (SEC) of 3.17 MJ/kgwater and the highest HRY (51.91%), but approximately 6 h was needed to reach the target moisture content for a batch of 20 kg. Higher drying temperatures substantially shortened the drying time but increased the SEC and reduced the HRY to 39.01–41.15%, suggesting a trade-off between drying rate, energy use, and milling quality. The model accurately predicted the moisture content, electrical energy consumption, and HRY, with coefficients of determination (R²) ≥ 0.997 for both the training and testing datasets. Through a sensitivity analysis, the paddy grain temperature (Tpaddy) was identified as the most informative predictor, accounting for approximately 38% of the total relative importance, followed by drying time (32%) and drying air temperature (21%). These findings indicate that Tpaddy provides valuable information on the transient thermal state of the grain, and can serve as a key process variable for forecasting drying behavior. Rather than representing a standalone commercial drying solution, the framework presented here provides a basis for developing multi-stage drying strategies in which rapid pneumatic pre-drying can be followed by tempering or drying at lower temperatures to strike a balance between drying efficiency and HRY.

Why this record is monitored

This record has an Impact Signal of 72/100 based on recency, source, collaboration, and bibliographic signals. It prioritizes monitoring and is not a judgment of research quality.

Related topics: Food Drying and Modeling · Agricultural Engineering and Mechanization · Energy, Environment, Agriculture Analysis