Gas Pressure Prediction and Application with Missing Data Imputation Techniques for Gas Regulator Data

There are 34 general city gas suppliers operating gas regulator facilities in South Korea. Ensuring a consistent and stable gas supply to general consumers is of utmost importance. However, the aging gas regulators and associated infrastructure have increasingly become a risk. To mitigate such risks, city gas supply companies are generally focusing on developing measures for swift response and prevention of gas supply interruptions and accidents through integrated safety management systems and remote monitoring devices. Nevertheless, the technical challenges of accident prevention persist. The goal of this research is to propose a preemptive system to provide proactive solutions, emphasizing the prediction of gas regulator pressure as a foundation for risk prevention. In contrast, real-time data is vulnerable to missing data and outliers caused by communication breakdowns and gas regulator malfunctions. It is critically hindering prediction accuracy. Therefore, this paper employs and compares various imputation techniques to handle missing data in gas regulator datasets. Through this process, the robustness of the accident prevention system can be improved.