Pemodelan Beban Puncak Konsumsi Listrik Di Wilayah Kupang Menggunakan Bayesian Mixture Normal Autoregressive

Pallo, Marchy (2016) Pemodelan Beban Puncak Konsumsi Listrik Di Wilayah Kupang Menggunakan Bayesian Mixture Normal Autoregressive. Masters thesis, Institut Teknologi Sepuluh Nopember.

[img]
Preview
Text
1314201024-Master_Thesis.pdf - Accepted Version

Download (3MB) | Preview

Abstract

Di Indonesia, persentase pertumbuhan pemakaian listrik naik setiap tahun. Adanya peningkatan tersebut menyebabkan Perusahaan Listrik Negara (PLN) harus menyediakan daya yang cukup untuk memenuhi permintaan pelanggan. Ketidakcukupan daya berdampak pada berbagai sektor rumah tangga, industri dan komersial. Kupang adalah salah satu wilayah di Indonesia yang sering mengalami kekurangan daya sehingga perlu direncanakan persediaan listrik. Perencanaan persediaan listrik membutuhkan manajemen yang efisien terhadap sistem daya yang ada dan optimasi keputusan yang berkaitan dengan tambahan kapasitas berdasarkan tingginya beban puncak. Beban puncak pemakaian listrik memberikan efek drop tegangan yang bervariasi bagi pengguna listrik. Data beban puncak berindikasi memiliki bentuk mixture dan bersifat lepto-kurtik, uni-modal dan fat-tails sehingga cocok untuk dimodelkan menggunakan metode Mixture Normal Autoregressive (MNAR) dengan pendekatan Bayesian yang diimplementasikan dalam doodle WinBUGS. Untuk mendapatkan model terbaik dari 2 dan 3 komponen mixture dilihat dari nilai Deviance Information Criterion (DIC) terkecil. Hasil yang diperoleh yaitu model untuk 2 komponen mixture atau MNAR(2; [1,2,4,6],6]) memiliki nilai DIC yang minimum sebesar 16467,6 dibanding ketiga model MNAR lainnya sehingga model ini baik untuk dilakukan peramalan jangka pendek. ====================================================================================================== In Indonesia, the percentage of growth in electricity consumption rises every year. To fulfill the increment of consumptions in the State Electricity Company should provide enough power to satisfy customer demand. Insufficiency of power has an impact on various sectors of household, industrial, and commercial. Kupang is one of the areas in Indonesia, which is often lack the power that needs to be planned electricity supplies. Electric supply planning requires efficient management of the existing power systems and optimization decisions related to the additional capacity due to high peak loads. Electricity peak load effects varying voltage drop for power users. Data indicate as a mixture of leptokurtic, uni-modal, and fat-tails making it suitable to be modeled using Mixture Normal Autoregressive (MNAR) with a Bayesian approach implemented in WinBUGS. The best models of two and three component mixture can be seen from the smallest value of the Deviance Information Criterion (DIC). The results shows that the models with two component mixture or MNAR (2; [1,2,4,6], 6) has a minimum value of DIC (16467.6) compared to the three other models, so the best selected model is better to do short-term forecasting

Item Type: Thesis (Masters)
Additional Information: RTSt 519.542 Pal p
Uncontrolled Keywords: Bayesian, beban puncak, DIC, MNAR, WinBUGS
Subjects: Q Science > QA Mathematics > QA279.5 Bayesian statistical decision theory.
Divisions: Faculty of Mathematics and Science > Statistics > 49101-(S2) Master Thesis
Depositing User: Yeni Anita Gonti
Date Deposited: 04 May 2020 02:25
Last Modified: 04 May 2020 02:25
URI: https://repository.its.ac.id/id/eprint/75929

Actions (login required)

View Item View Item