Prediksi Mutu Internal Tandan Buah Segar Kelapa Sawit Secara Non-destruktif Berbasis Kapasitansi Listrik.

Abstract

Mutu tandan buah segar (TBS) kelapa sawit merupakan salah satu faktor yang menentukan kualitas dan rendemen minyak yang dihasilkan. Penentuan mutu TBS di lapangan umumnya masih dilakukan secara visual sehingga bersifat subjektif dan belum mampu menggambarkan kondisi internal buah secara akurat. Di sisi lain, analisis laboratorium untuk menentukan kadar air, kadar minyak, dan asam lemak bebas (ALB) masih bersifat destruktif, memerlukan waktu, serta kurang efisien untuk pengendalian mutu secara cepat. Oleh karena itu, diperlukan pendekatan nondestruktif yang mampu mengevaluasi mutu internal TBS secara cepat dan objektif. Salah satu pendekatan yang berpotensi digunakan adalah pengukuran kapasitansi listrik multifrekuensi karena respons listrik bahan dipengaruhi oleh perubahan komposisi internal selama proses pematangan. Penelitian ini bertujuan untuk menentukan karakteristik kapasitansi listrik multifrekuensi pada berbagai tingkat kematangan TBS kelapa sawit, menganalisis keterkaitannya dengan kadar air, kadar minyak, dan asam lemak bebas, serta mengembangkan model prediksi mutu internal berbasis kapasitansi listrik secara nondestruktif. Penelitian menggunakan 50 TBS kelapa sawit varietas Tenera yang terdiri atas lima tingkat kematangan. Pengukuran kapasitansi listrik dilakukan menggunakan Inductance, Capacitance, and Resistance (LCR) meter pada 14 frekuensi dalam rentang 50 Hz hingga 100 kHz dengan elektroda pelat tembaga yang disesuaikan dengan dimensi TBS. Parameter mutu internal yang dianalisis meliputi kadar air, kadar minyak, dan ALB. Data mutu internal dianalisis secara deskriptif dan dengan analisis ragam (ANOVA). Keterkaitan antara data kapasitansi listrik multifrekuensi dan parameter mutu internal dianalisis menggunakan Multiple Linear Regression (MLR), sedangkan model prediksi dikembangkan menggunakan Partial Least Squares Regression (PLSR) dengan beberapa perlakuan prapengolahan data. Hasil penelitian menunjukkan bahwa respons kapasitansi listrik berubah seiring dengan tingkat kematangan TBS, yang mencerminkan perubahan kondisi internal buah selama proses pematangan. Analisis menggunakan MLR menunjukkan bahwa data kapasitansi listrik multifrekuensi memiliki keterkaitan yang cukup baik dengan kadar air dan kadar minyak, dengan nilai koefisien determinasi (R²) validasi masing-masing sebesar 0,78 dan 0,80 sedangkan keterkaitannya dengan asam lemak bebas relatif lebih rendah (R² = 0,22). Pengembangan model menggunakan PLSR menunjukkan bahwa prapengolahan Multiplicative Scatter Correction (MSC) menghasilkan kinerja model prediksi terbaik untuk kadar air dan kadar minyak, sedangkan model prediksi asam lemak bebas masih menunjukkan kemampuan prediksi yang relatif rendah. Penelitian ini menunjukkan bahwa kapasitansi listrik multifrekuensi berpotensi digunakan sebagai metode nondestruktif untuk pendugaan mutu internal TBS kelapa sawit, khususnya kadar air dan kadar minyak. Sementara itu, model prediksi ALB masih memerlukan pengembangan lebih lanjut untuk meningkatkan akurasi dan keandalannya sebelum diterapkan pada skala operasional.

Oil palm fresh fruit bunch (FFB) quality is key to crude palm oil quality and oil extraction yield. In practice, FFB quality is assessed by visual inspection, which is subjective and does not accurately reflect internal quality. Laboratory analyses of moisture, oil, and free fatty acid (FFA) content are destructive, time-consuming, and less efficient for rapid evaluation. Therefore, a non-destructive approach is required to rapidly and objectively evaluate internal FFB quality. One promising approach is to measure multi-frequency electrical capacitance, as the electrical response of biological materials changes with their internal composition during fruit maturation. This study aimed to analyze the characteristics of multi-frequency electrical capacitance at different maturity stages of oil palm FFB, examine its relationship with moisture content, oil content, and FFA content, and develop a non-destructive prediction model for internal quality based on electrical capacitance. A total of 50 Tenera oil palm fresh fruit bunches representing five maturity stages were used. Electrical capacitance was measured using an Inductance, Capacitance, and Resistance (LCR) meter at 14 frequencies ranging from 50 Hz to 100 kHz with copper plate electrodes designed to match the dimensions of the bunches. The internal quality parameters analyzed included moisture, oil, and FFA content. Internal quality data were analyzed descriptively and by analysis of variance (ANOVA). The relationship between multi-frequency electrical capacitance and internal quality parameters was analyzed using Multiple Linear Regression (MLR), while prediction models were developed using Partial Least Squares Regression (PLSR) with several signal preprocessing techniques. The results showed that the electrical capacitance response changed with the maturity stage of oil palm fresh fruit bunches (FFB), reflecting changes in the internal composition of the fruit during the ripening process. Analysis using MLR indicated that multi-frequency electrical capacitance was reasonably well related to moisture content and oil content, with validation coefficients of determination (R²) of 0.78 and 0.80, respectively, whereas its relationship with free fatty acid (FFA) content was relatively weaker (R² = 0,22). Furthermore, prediction models developed using PLSR showed that Multiplicative Scatter Correction (MSC) preprocessing provided the best prediction performance for moisture and oil contents, whereas the prediction performance for FFA remained relatively low. This study demonstrates that multi-frequency electrical capacitance has considerable potential as a non-destructive approach for predicting the internal quality of oil palm FFB, particularly moisture and oil contents. However, the FFA prediction model still requires further improvement to achieve higher accuracy and robustness before practical implementation.