Download Bukti Potong
| Nama |
|---|
| NIP |
|---|
| NPWP |
|---|
| Alamat |
|---|
| Tahun |
|---|
The load coefficient methodh as been used PLN was still gives the prediction error is quite largeon average ranged between 4% -5%. Giving rise to substantial power loss to PLN foreach unit of time. It can be concluded that PLN needs is how to predict short-termelectric load more accurately. For the considered approach to soft computing.Therefore, this study will use the algorithm Backpropagation Neural Network with Gradient Descent method as an approach to predict the short-term electric load. Will implement the algorithm using MatLab 2009b.The algorithms have been developed in this study will be applied to load data via a simulation model. This study used data load, 1 January to 31 May 2011 to load in the morning, afternoon and evening. Evaluation carried out by observing short-term load prediction results from the application of Back Propagation Algorithm with Gradient Descent Method. Performance measurement is done by calculating the average magnitude of error that occurred through the Mean Square Error (MSE). The smaller value of MSE stated the closer the predicted value with actual value. Thus it can be seen the accuracy of Back Propagation algorithm with the Gradient Descent method.
Key Word : Algoritma Backpropagation, Metode Gradien Discent, Beban Listrik
