Uji Kinerja Metode Deep Convolutional Neural Networks Untuk Identifikasi Gangguan Daya Listrik
The identification model development of the power disturbance signals with the deep convolutional neural networks (CNNs) method involves a large amount of data. However, the real signal data is limited. Therefore, researchers employ synthetic signal data. These signals can be generated by the formula IEEE standardized. In these formulas, two categories have a similar formula i.e interruption and sag. The difference is only in the intensity parameter (α). This paper analyzed the model performance of identifying those disturbances where the intensity values are set differently for training and testing datasets based on the upper bound value α of sag and the lower bound value α of interruption. Several noise levels are included in the signals. So, there are several datasets with noises in this simulation. Furthermore, those datasets are trained using the model based on deep CNN. The test results show that the true positive (TP) of the model's performance in identifying the interruption signal is 93.54% and the sag signal is 78.78%. In addition, the performance of the model using a dataset without noise obtained a high percentage in accuracy, precision, and f1-score parameters with 92.4%, 97.4%, and 92,76%, respectively.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).