Pembaharuan konsep prediksi debit andalan untuk operasi dan pemeliharaan irigasi modern

Abstract

The management of conventional irrigation system, which is still being applied in the recent time is probabilistic, especially in analyzing discharge for planning of irrigation operation and maintenance. It can be seen from the process of exerted data analysis, which is two-week or ten-days empirical data analysis. Therefore, to change the management of conventional irrigation system into modern requires flexibly and real-time based due to climate change in Indonesia. The available discharge prediction analysis is done using mathematic model analysis as the replacement of probabilistic model and the use of real time observation data by utilizing automatic weather observation technology. This new concept has been attempted in Irrigation Area of Wadaslintang and Banyumas Regency, showing that automatic weather observation worked successfully and can be used as the data within mathematic model analysis. The result indicates that telemetry instruments work well as expected. The difference between the use of mathematic method of Artificial Neural Network (ANN) with probabilistic method of P80 shows that ANN method is closer to real compared to the probabilistic P80. It is shown with the validation result measured from January to August 2015. Overall, errors between water surcharge prediction with ANN and realisation is 77%.  According to the results, it is suggested that dynamic mathematic measurement method is needed, due to dynamic condition of climate in spite of not neglecting probabilistic method as comparison.