Prediksi Form Factor Kapal Dari Data Uji Model Kapal Dengan Metode Robust Least Squares

Towing tank adalah salah satu media pengujian model kapal dengan skala tertentu untuk meperoleh tahanan total model kapal, dimana nilai ini dipakai dalam menentukan resistance kapal full scale. Perhitungan extrapolasi untuk mendapatkan nilai resistance pada kapal skala penuh diperlukan nilai form factor, yang secara klasik dapat dilakukan dengan menarik model kapal dengan besaran froude number antara 0.1-0.2 yang dikenal dengan metode Prohaska. Selain metode klasik, form factor bisa didapatkan dengan metode statistik dari principal dimensions kapal. Laboratorium pengujian model kapal, Laboratorium Hidrodinamika (LH) – Badan Riset & Inovasi Nasional (BRIN) mempunyai data yang cukup untuk dipakai dalam penelitian prediksi form factor melalui statistik. Tujuan penelitian adalah mendapatkan dan menganalisa persamaan prediksi form factor dan nilai Effective Horse Power (EHP) pada kapal full scale. Principal dimensions kapal displacement terlebih dahulu dilakukan transformasi menjadi data sekunder dengan operasi matematis. Selanjutnya menggunakan metode Variable Selection Least Squares (VARSEL) untuk mendapatkan variabe bebas (variabel X) dengan nilai R2 yang optimal. Dengan metode Grubbs’ diperoleh outlier, sehingga digunakan metode Robust Least Squares (RLS), dimana diperoleh persamaan prediksi form factor adalah Log Y = 0.03569 + 0.770546.(B/L) + 0.00165. (L/T C_M ) + 0.001419. (√(W_SA ))- 0.033767. (B/T C_P )+ 0.00388. (√(W_SA )/√(3&∆))- 0.736. (T/√(W_SA )) Validasi persamaan regresi prediksi form factor dilakukan dengan tiga model kapal, yaitu mini LNG, Ferry RO-RO desain awal dan modifikasi dimana perbedaan prosentase kedua metode adalah 0.1 – 0.852273 %, sehingga persamaan regresi prediksi form factor metode RLS dapat direkomendasikan sebagai alternatif untuk mendapatkan nilai form factor. =================================================================================================================================== Towing tanks are one of the media for testing ship models with a certain scale to obtain the total resistance of a ship model, where this value is used to determine the resistance of a full scale ship. Extrapolation calculations to obtain resistance values for full-scale ships require form factor values, which classically can be done by drawing ship models with a Froude number between 0.1-0.2, known as the Prohaska method. In addition to the classical method, form factor can be obtained by statistical methods from the ship's principal dimensions. Ship model testing laboratory, Hydrodynamics Laboratory (HL) – National Research & Innovation Agency (BRIN) has sufficient data to be used in research on form factor prediction through statistics. The aim of the research is to obtain and analyze the form factor prediction equation and the Effective Horse Power (EHP) value on a full scale ship. Principal dimensions of displacement ships are first transformed into secondary data with mathematical operations. Then use the Variable Selection Least Squares (VARSEL) method to get the independent variable (variable X) with the optimal R2 value. Using the Grubbs method, outliers are obtained, so the Robust Least Squares (RLS) method is used. The form factor prediction equation is Log Y = 0.03569 + 0.770546.(B/L) + 0.00165. (L/T C_M ) + 0.001419. (√(W_SA ))- 0.033767. (B/T C_P )+ 0.00388. (√(W_SA )/√(3&∆))- 0.736. (T/√(W_SA )). The validation of the form factor prediction regression equation was carried out with three ship models, namely mini LNG, Ferry RO-RO initial design and modification which the difference in the percentage of the two methods was 0.1 – 0.852273 %, so the form factor prediction regression equation with the RLS method can be recommended as an alternative to obtain form factor values