Prediksi Kandungan Kimia Buah Kelapa Sawit secara Non-Destruktif Berdasarkan Sifat Konduktansi Listrik

Abstract

Kualitas minyak kelapa sawit ditentukan oleh kematangan buah. Panen pada waktu yang tepat menghasilkan kadar ALB rendah. Penentuan kematangan secara visual berdasarkan warna memiliki kekurangan karena bersifat subyektif, sedangkan dengan metode kimia sangat akurat, tetapi memerlukan waktu yang lama dan biaya mahal. Penelitian ini bertujuan mengembangkan metode prediksi kandungan kimia buah kelapa sawit secara non-destruktif berdasarkan sifat konduktansi listrik menggunakan metode Partial Least Squares (PLS). Pengukuran sifat listrik dilakukan menggunakan LCR meter, kandungan kimia (kadar air, kadar minyak, dan kadar ALB) ditentukan menggunakan metode kimia. Data konduktansi listrik dilakukan pre-treatment, kemudian dikalibrasi dengan data kimia. Prediksi kadar air terbaik diperoleh menggunakan pre-treatment MSC dengan 1 faktor PLS (r = 0,94; SEC = 8,14%; SEP = 8,80%; CV = 13,68%; RPD = 2,67). Prediksi kadar minyak terbaik diperoleh menggunakan pre-treatment Normalize dengan 1 faktor PLS (r = 0,91; SEC = 7,51%; SEP = 6,66%; CV = 29,92%; RPD = 2,27). Sedangkan prediksi kadar asam lemak bebas terbaik adalah metode konduktansi menggunakan pre-treatment SNV dengan 1 faktor PLS (r = 0,66; SEC = 1,35%; SEP = 1,74%; CV = 50,78%; RPD = 1,04). Metode konduktansi dan PLS mempunyai potensi digunakan untuk prediksi kadar air dan kadar minyak buah sawit secara non-destruktif, namun tidak dapat digunakan untuk prediksi asam lemak bebas.

The quality of palm oil is determined by the ripeness of the fruit. Harvesting at the right time results in low FFA levels. Determining ripeness visually based on color has its drawbacks because it is subjective, while chemical methods are very accurate but require a long time and are expensive. This study aims to develop a non-destructive method for predicting the chemical composition of palm oil fruits based on electrical conductance properties using the Partial Least Squares (PLS) method. Electrical properties were measured using an LCR meter, and chemical composition (moisture content, oil content, and FFA content) was determined using chemical methods. Electrical conductance data were pre-treated and then calibrated with chemical data. The best water content prediction was obtained using the MSC pre-treatment with 1 PLS factor (r = 0.94; SEC = 8.14%; SEP = 8.80%; CV = 13.68%; RPD = 2.67). The best oil content prediction was obtained using the Normalize pre-treatment with 1 PLS factor (r = 0.91; SEC = 7.51%; SEP = 6.66%; CV = 29.92%; RPD = 2.27). Meanwhile, the best prediction of free fatty acid content was obtained using the conductance method with SNV pre-treatment and 1 PLS factor (r = 0.66; SEC = 1.35%; SEP = 1.74%; CV = 50.78%; RPD = 1.04). Conductance and PLS methods has potentially used for non-destructive prediction of water and oil content in palm fruit, but cannot be used for predicting free fatty acid content.