<![CDATA[Repeated disturbances in power distribution networks can reduce system reliability and increase maintenance costs. However, maintenance activities are often performed reactively because recurring disturbance patterns have not been systematically identified from historical operational data. This study aims to identify recurring disturbance patterns in the distribution network of PT PLN (Persero) by applying the Apriori algorithm within the CRISP-DM framework. The dataset consisted of 29,446 historical disturbance records containing disturbance type, cause, and location. After the data preparation stage, which included data cleaning, standardization, and transformation, 10,274 valid transactions were obtained for association rule mining. The Apriori algorithm was implemented using a minimum support threshold of 0.5%, a minimum confidence threshold of 70%, and lift values greater than 1 as the evaluation criterion. The analysis produced 57 frequent 1-itemsets, 180 frequent 2-itemsets, 35 frequent 3-itemsets, and 64 association rules. The results indicate that Core FO Problem, FOC – Putus Core, and Putus Kabel FO are the most frequently occurring disturbance patterns. The strongest association rule, Banten and FOT – Software → Konfigurasi Di Router Problem, achieved a support of 0.993%, a confidence of 81.6%, and the highest lift value of 30.936, indicating a strong positive relationship between the antecedent and consequent. These findings demonstrate that the Apriori algorithm is effective in discovering recurring disturbance patterns from historical operational data and can provide maintenance-oriented knowledge to support inspection prioritization and preventive maintenance planning in power distribution networks.]]>
