AI-Based Condensation Prevention Automation System

Overview

Utility tunnels are densely packed with critical infrastructure such as power, telecommunications, gas, and water systems. Condensation within these tunnels can cause structural corrosion and facility damage, potentially causing serious safety incidents. Previously, inspections relied primarily on manual methods, limiting real-time response and reducing efficiency.

Condensation also contributes to functional degradation of facilities and increased maintenance costs, threatening long-term infrastructure stability. Consequently, there is a demand for a system capable of proactive prevention through data-driven AI analysis, moving beyond the existing passive maintenance approach.

This project aims to develop an AI-based condensation prediction and ventilation control system. The goal is to establish an innovative condensation prevention service that integrates and manages temperature and humidity data within the utility tunnel, trains an AI model based on this data to predict condensation occurrence, and derives optimal ventilation solutions.

Project description

The project established a system that collects temperature and humidity data within utility tunnels in real time, stores it in a data lake, and utilizes AI to predict the likelihood of condensation occurrence.

External temperature and humidity data from the Korea Meteorological Administration were integrated with sensor data from inside the utility tunnel to build an advanced dataset. This data was refined and processed for AI model training, and being labeled whether condensation occurred to enhance the prediction model's accuracy.

During AI model development process, an artificial neural network (ANN) was used to analyze temperature and humidity change patterns, enabling the prediction of condensation likelihood overtime Computational fluid dynamics (CFD) analysis was also employed to stimulate airflow within the tunnels and determine optimal operating conditions for ventilation fans.

Finally, an optimized ventilation control model was implemented by combining the AI condensation prediction algorithm with CFD analysis results. This enables the system to automatically control ventilation fans and adjust the internal environment when condensation is likely to occur under specific temperature and humidity conditions.

Key technological features

• Predicts the likelihood of condensation occurring within utility tunnels and derives optimal ventilation control measures to implement a real-time automated system for condensation prevention
• Core Algorithms: Combination of Artificial Neural Network (ANN) and Computational Fluid Dynamics (CFD) Model
  • Learning temperature and humidity data to predict future conditions, and analyzing airflow through CFD analysis

Expected outcomes

Through this project, an AI-based condensation prevention system has been successfully developed and its effectiveness verified through field testing. By analyzing sensor data within the utility tunnel in real time and predicting the likelihood of condensation occurrence in advance, maintenance has become more efficient than with previous manual inspection method.

Furthermore, the introduction of an AI-based automated ventilation control system has improved the ability to maintain optimal conditions without requiring manual labor. This is expected to prevent structural damage to the utility tunnel, reduce maintenance costs, and extend the facility's lifespan.

In the future, this project plan to conduct additional data collection and analysis to further enhance the predictive accuracy of the AI model. Furthermore, this project intended to analyze condensation occurrence patterns under various environmental conditions to develop a more sophisticated model and supplement it with an automated solution capable of real-time response. To expand the service's reach, research will be conducted to enable application beyond utility tunnels to include power tunnels, underground conduits, road tunnels, and railway tunnels. This aims to enhance the maintenance efficiency of underground facilities and establish a smart safety management system for national infrastructure.

Launch year: 2024

Attribution

Case study courtesy of the National Information Society Agency, Republic of Korea, Apolitical's Content Partner. Sourced from Use Cases of Public AI Service, Vol.1,  https://eng.nia.or.kr/site/nia_eng/ex/bbs/View.do?cbIdx=31975&bcIdx=28997&parentSeq=29037