Prediction of moisture content for a single maize kernel based on viscoelastic properties

Area of Science:

• Agricultural Engineering
• Food Science
• Materials Science

Background:

• Accurate moisture content detection is crucial for maize quality assurance.
• Existing rapid detection methods face challenges with cost, environmental needs, and accuracy.
• A simple, effective method for single maize kernel moisture detection using viscoelasticity is proposed.

Purpose of the Study:

• To develop and validate a novel method for rapid, accurate moisture content detection in single maize kernels.
• To explore the efficacy of viscoelastic properties in determining maize kernel moisture.
• To compare the predictive power of viscoelastic parameters versus force-time graph data.

Main Methods:

• Performed viscoelastic experiments: relaxation tests (60-100 N) and frequency-sweep tests (0.6-1 Hz).
• Extracted viscoelastic parameters using the four-element Maxwell model.
• Utilized force-time graph data and viscoelastic parameters as inputs for predictive models, exploring preprocessing techniques.

Main Results:

• Models using force-time data demonstrated higher accuracy than those using viscoelastic parameters.
• Partial least squares regression on S-G smoothed relaxation test data (100 N) yielded the best model.
• Achieved high correlation coefficients (0.954 calibration, 0.905 prediction) and low RMSE (0.021 calibration, 0.029 prediction).

Conclusions:

• Viscoelastic properties offer a viable approach for fast and accurate single maize kernel moisture detection.
• This method presents a novel technique applicable to analyzing various cereal components.
• The findings support the development of advanced, cost-effective quality control tools for grains.