Frequency variation in non-destructive measurement of impedance of oil palm fruits for moisture and oil content determination

Moisture and oil content are important elements in oil palm fruits quality monitoring during ripening process. Conventionally, moisture and oil content were determined by oven dried and Soxhlet extraction methods, respectively. However, both of these methods are commonly time consuming, laborious and tedious. In this study, impedance spectroscopy technique was introduced to measure moisture and oil content of oil palm fruits. A total of 90 fruits sample from 12 weeks after anthesis (WAA), 16 WAA and 20 WAA at different maturity stages were tested. Fruits sample was divided into two sets which were 60 fruits for testing and another 30 fruits for validation. A pair of an electrocardiogram (ECG) electrodes were attached to the Kelvin clip leads (16089E, Agilent Technologies, Japan) which was connected to a LCR meter (4263B, Agilent Technologies, Japan) at four discrete frequencies of 1 kHz, 10 kHz, 20 kHz and 100 kHz to measure impedance value of each oil palm fruits. Statistical analysis showed that impedance value change with different maturity stages and it was observed that high maturity stage (20 WAA) showed the highest impedance value due to the less moisture and high oil content inside the fruits. A good correlation was found at frequency of 100 kHz which has the highest regression coefficient (R2) and lowest root mean square error (RMSE) for moisture and oil content prediction. For moisture content prediction, a linear regression equation of y100 = -17.85x + 79.08 was develop with R2 of 0.77 and RMSE of 5.85% while for oil content prediction, another linear regression equation of y100 = 16.07x + 23.65 was develop with R2 of 0.72 and RMSE of 5.71%. These results indicated that non-destructive measurement of impedance has good potential to determine moisture and oil content of oil palm fruits.

Preview Text FK 2016 68 - IR.pdf Download (1MB) | Preview

Actions (login required)

View Item