Rizky Wira Utomo, Muhammad (2021) Analisis Perbandingan Prakiraan Kecepatan Angin Menggunakan Metode RNN dan Multivariate ARIMAX. Sarjana thesis, Universitas Brawijaya.
Abstract
Besarnya daya listrik yang mampu dihasilkan oleh pembangkit listrik tenaga bayu (PLTB) bergantung pada besarnya kecepatan angin yang digunakan. Namun, kecepatan angin sendiri bersifat acak, intermiten dan cenderung tidak pasti. Kondisi seperti ini akan membawa dampak signifikan ke jaringan dan kesulitan pada pengiriman daya. Maka dari itu, prediksi dari kecepatan angin yang akurat dan presisi sangat penting untuk memastikan kestabilan operasi PLTB. Prediksi tersebut dapat dilakukan dengan menggunakan metode artificial intelligence maupun metode konvensional. Pada penelitian ini metode artificial intelligence, Recurrent Neural Network (RNN) digunakan dan dibandingkan dengan metode Auto Regressive Integrated Moving Average with Exogenous Variable (ARIMAX). Hasil penelitian menunjukan: 1) Arsitektur RNN untuk melakukan prakiraan kecepatan angin dengan optimal adalah dengan pembagian data training 75% dan data testing 25% dengan 75 Hidden Neuron 2) Hasil peramalan kecepatan angin di Malang menggunakan metode RNN memiliki nilai MAE terkecil sebesar 0,2516 m/s dan nilai RMSE sebesar 0,3232 m/s, sedangkan hasil peramalan menggunakan metode Multivariate ARIMAX memiliki nilai MAE 0,7027 m/s dan nilai RMSE sebesar 0,8864 m/s. 3) Hasil peramalan kecepatan angin di Basel menggunakan metode RNN memiliki nilai MAE terkecil sebesar 1,0052 m/s dan nilai RMSE sebesar 1,3016 m/s, sedangkan hasil peramalan menggunakan metode Multivariate ARIMAX memiliki nilai MAE 1,5527 m/s dan nilai RMSE sebesar 2,0532 m/s. Hal ini menunjukkan bahwa metode RNN lebih akurat dan optimal daripada metode Multivariate ARIMAX.
English Abstract
The amount of wind power generated by Wind Power Plant depends on the wind speed that the system use. However, the wind speed itself is random and inconsistent. This will bring significant impact and difficulties in power dispatching. Therefore, an accurate and precision wind speed prediction is required to secure the stability of the power plant. Wind speed predictions can be made by using artificial Intelligence methods such as Recurrent Neural Network (RNN) or statistical conventional methods such as Multivariate Auto Regressive Integrated Moving Average with Exogenous Variable (ARIMAX). RNN training and ARIMAX modelling are apllied using the hourly weather data during specified time interval, based on the variation of training and testing data. In RNN method, the trained network will be used to forecast the wind speed. The RNN forecast result will be compared with the actual data to get the RMSE and MAE values in order to see the accuracy level of the forecast. In Multivariate ARIMAX method, each weather data used is the input as the exogenous variable. Exogenous variables are used to define the parameter of the Multivariate ARIMAX model that will be used to perform the forecasting. The forecast result of Multivariate ARIMAX method will also be compared to the actual data to get the RMSE and MAE values. The result of the research shows that: 1) The optimal RNN architecture for wind speed forecasting in Malang was using 75% training data and 25% testing data with 75 hidden heurons. 2) Forecasting result in Malang with RNN method has MAE 0,2516 m/s and RMSE 0,3232 m/s, whereas forecasting with Multivariate ARIMAX has MAE value 0,7027 m/s and RMSE 0,8864 m/s. 3) Forecasting result in Basel with RNN has MAE 1,0052 m/s and RMSE 1,3016 m/s whereas forecasting with Multivariate ARIMAX has MAE 1,5527 m/s and RMSE 2,0532 m/s. This concludes that RNN method is more accurate and optimal than Multivariate ARIMAX method.
Other obstract
-
| Item Type: | Thesis (Sarjana) |
|---|---|
| Identification Number: | 621.381 |
| Uncontrolled Keywords: | Prakiraan, Kecepatan Angin, RNN, Multivariate ARIMAX, forecasting, wind speed. |
| Subjects: | 600 Technology (Applied sciences) > 621 Applied physics > 621.3 Electrical, magnetic, optical, communications, computer engineering; electronics, lighting > 621.38 Electronics, communications engineering > 621.381 Electronics |
| Divisions: | Fakultas Teknik > Teknik Elektro |
| Depositing User: | Unnamed user with email gaby |
| Date Deposited: | 21 Oct 2021 05:56 |
| Last Modified: | 24 Feb 2022 04:51 |
| URI: | http://repository.ub.ac.id/id/eprint/184623 |
![[thumbnail of DALAM MASA EMBARGO]](http://repository.ub.ac.id/style/images/fileicons/text.png) |
Text (DALAM MASA EMBARGO)
Muhammad Rizky Wira Utomo.pdf Restricted to Registered users only until 31 December 2023. Download (30MB) | Request a copy |
Actions (login required)
 |
View Item |