Model Peramalan Inflow Waduk Plta Koto Panjang Menggunakan Pendekatan Adaptive Neuro Fuzzy Inference System

Abstract

Many studies has shown that the ANFIS  model  has good prediction
performance results for forecasting hydrological phenomena such as reservoir
inflow. The existence of the success, it is necessary to research the  reliability of the
ANFIS models when applied to existing reservoirs in  Riau Province, namely  Koto
Panjang Hydropower Reservoir.
The results of  the ANFIS  models  in the  development  phase  of this  study
indicate that the scheme 4 (4 inputs) with the ROI value of 0.02 is the best scheme so
that the scheme is used to build the model of ANFIS. The cross validation on ANFIS
models  provide  the best  performance  prediction  with  correlation coefficient (R)
between the prediction values and the observational value is 0.90.
Keywords:  Inflow, Koto Panjang Hydropower Reservoir, ANFIS, Cross Validation .