Performance Analysis of Fuzzy Logic and Artificial Neural Network for Classifying Hydropower Plant Remotization Technology Readiness

Hydropower plant remotization can improve operational efficiency. Despite this advantage, less than 42 % of PLN Nusantara Power's hydropower plants have implemented this technology due to varying maturity levels at different sites. Assessing the Technology Readiness Level (TRL) of hydropower plants is crucial for identifying specific plants that require upgrades in remotization technology. Currently, the TRL classification method process is still manual, based on expert opinions, which has the potential to cause various limitations, especially related to subjectivity. To address this problem, this study proposes fuzzy logic through the Fuzzy Inference System-Takagi-Sugeno-Kang (FIS-TSK) for classifying hydropower plant remotization technology readiness. This method creates fuzzy rules in order to capture expert opinions in TRL classification. The classification considers the following six factors: Information and Communication Technologies (ICT) infrastructure, Sequence (SQC) start-stop control, Governor (GOV) turbine control, Automatic Voltage Regulator (AVR) control, Supervisory Control and Data Acquisition (SCADA), and Substation Control and Protection (SCP). The results showed that FIS-TSK outperforms the Artificial Neural Network (ANN), achieving a Mean Absolute Percentage Error (MAPE) of 0.06 in assessing the Remote Operation Readiness Level (RORL) for 23 hydropower plants and higher accuracy, precision, and recall of up to 100 % in TRL classification. These findings provide valuable insights for developing web-based classification applications in future work and helping decisionmakers evaluate the readiness for hydropower plant remotization.