All Issue

2026 Vol.27, Issue 11 Preview Page
Concordance Analysis of Burial-Environment-Weighted and AI Models for District Heating Pipe Risk Assessment

매설환경 가중 회귀모델과 AI 모델의 일치도 분석을 통한 열수송관 위험도 평가

Myeongsik Kong1
,
Jaemo Kang2*
,
Sungyeol Lee3
공 명식1
,
강 재모2*
,
이 성열3
1Research Specialist, Department of Geotechinical Engineering Research, Korea Institute of Civil Engineering and Building Technology
2Research Fellow, Department of Geotechinical Engineering Research, Korea Institute of Civil Engineering and Building Technology
3Postdoctoral Researcher, Department of Geotechinical Engineering Research, Korea Institute of Civil Engineering and Building Technology
*Corresponding Author
1 November 2026. pp. 15-22
PDF XML Citation
Abstract

Service-life-based regression models and artificial intelligence (AI) models using maintenance-history data have been developed to predict failures in district heating pipelines in Korea. However, no previous study has used actual failure data to verify how accurately the two approaches assess pipeline risk or how closely their risk estimates agree. In this study, GIS spatial analysis was conducted using asset records for 328,279 pipe segments and 2,310 failure records maintained by a Korean district heating operator, together with public burial-environment data (road, waterway, and other) provided by VWorld. The existing service-life-based regression model was then improved using the spatial-analysis results and compared with an AI prediction model based on an XGBoost-LightGBM ensemble to quantify agreement between the models. Spearman rank correlations (ρ = 0.69–0.70, p < 0.001) and Cohen’s kappa values for high-risk classification (0.39–0.60) indicated a moderate level of agreement. Percentile-based validation against actual failure data showed that the AI model outperformed the service-life-based regression model in terms of the low-risk misclassification rate for the bottom 50 % (36.5 % vs. 45.3 %) and failure capture rates within the top 20–30 % risk groups. In contrast, the improved model was more effective than the AI model at capturing the extreme high-risk group in the top 10 %. These results suggest that applying a dual-validation framework that uses both models in practice could improve the reliability of maintenance and rehabilitation prioritization for district heating pipelines across different risk levels.

Keywords
District heating pipeline
Failure probability
Burial environment
Artificial intelligence
Model agreement

국내 지역난방용 열수송관의 파손예측을 위한 방법론으로 유지관리 이력정보를 활용한 사용기간 기반 회귀분석 모델과 인공지능 기반 모델이 개발되어 있다. 다만, 실제 파손 데이터를 활용하여 두 방법론이 시설물의 위험도를 얼마나 정확하게 판단하는지, 두 모델간 위험도 분석 결과가 얼마나 유사한지에 대한 검증은 진행된 사례가 없다. 본 연구에서는 국내 지역난방사업자가 관리하는 열수송관의 설비이력 328,279건과 파손이력 2,310건, 브이월드에서 제공하는 매설환경(도로·수계·기타) 공공데이터를 활용하여 GIS 공간분석을 진행하고, 이를 활용하여 기존 사용기간 기반 회귀분석 모델을 개선한 뒤, XGBoost·LightGBM 앙상블 기반 AI 예측모델과 비교하여 각 모델간 일치도를 정량적으로 분석하였다. 순위상관분석(Spearman ρ=0.69~0.70, p<0.001) 및 고위험군 분류 일치도(Cohen's Kappa 0.39~0.60) 분석 결과, 두 모델의 일치도는 중간 수준으로 확인되었다. 또한 실제 파손 데이터를 활용한 위험도 백분위 기반 모델 검증 결과, AI 모델이 하위 50 %의 저위험군 오분류율(36.5 % vs 45.3 %)과 상위 20~30 % 위험군 포착률에서 사용기간 기반 회귀분석 모델보다 우수한 것으로 나타났다. 반면 위험도 최상위 10 %에 해당하는 극단적 고위험군 포착 시 개선모델이 AI 모델보다 우수한 것을 확인할 수 있었다. 이러한 결과를 토대로 위험수준이 서로 다른 열수송관의 위험도를 예측할 때 현장에서 두 모델을 모두 활용하는 이중검증(dual-validation) 체계를 도입할 경우 성능개선 우선순위 결정의 신뢰성을 높일 수 있을 것으로 판단된다.

키워드
열수송관
파손확률 예측
매설환경
인공지능
일치도 분석
References
1

Chen, T. and Guestrin, C. (2016), XGBoost: A Scalable Tree Boosting System, Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pp. 785~794.

10.1145/2939672.2939785
2

Ke, G., Meng, Q., Finley, T., Wang, T., Chen, W., Ma, W., Ye, Q. and Liu, T.Y. (2017), LightGBM: A Highly Efficient Gradient Boosting Decision Tree, Advances in Neural Information Processing Systems, Vol. 30, pp. 3146~3154.

3

Kong, M.S. and Kang, J.M. (2021), Development of a Methodology for Estimating the Failure Probability of District Heating Pipes, Journal of the Korean Geo-Environmental Society, Vol. 22, No. 11, pp. 15~21 (in Korean).

4

Kong, M.S., Kang, J.M., Kim, J.Y. and Lee, S.Y. (2024), Development of Burial Environment Data and Analysis of Failure Trends for District Heating Pipes, Journal of the Korean Geo-Environmental Society, Vol. 25, No. 12, pp. 15~20 (in Korean).

5

Landis, J.R. and Koch, G.G. (1977), The Measurement of Observer Agreement for Categorical Data, Biometrics, Vol. 33, No. 1, pp. 159~174.

10.2307/2529310
6

Lee, S.Y., Kang, J.M. and Kim, J.Y. (2023), Prediction Modeling of Ground Subsidence Risk Based on Machine Learning Using the Attribute Information of Underground Utilities in Urban Areas in Korea, Applied Sciences, Vol. 13, No. 9, Article No. 5566.

10.3390/app13095566
7

Lee, S.Y., Kang, J.M., Kim, J.Y. and Kong, M.S. (2025), Prediction of High-Risk Failures in Urban District Heating Pipelines Using KNN-Based Relabeling and AI Models, Applied Sciences, Vol. 15, No. 20, Article No. 11104.

10.3390/app152011104
8

Lundberg, S.M. and Lee, S.I. (2017), A Unified Approach to Interpreting Model Predictions, Advances in Neural Information Processing Systems, Vol. 30, pp. 4765~4774.

9

Ministry of Government Legislation (2020), Integrated Energy Supply Act, Republic of Korea (in Korean).

10

Ministry of Trade, Industry and Energy (2019), Comprehensive Measures to Strengthen the Safety Diagnosis of District Heating Pipelines, Ministry of Trade, Industry and Energy, Republic of Korea (in Korean).

11

Rafati, A. and Shaker, H.R. (2024), Predictive Maintenance of District Heating Networks: A Comprehensive Review of Methods and Challenges, Thermal Science and Engineering Progress, Vol. 53, Article No. 102712.

10.1016/j.tsep.2024.102722
12

Sernhed, K. and Jönsson, A. (2017), Risk Management for Maintenance of District Heating Networks, Energy Procedia, Vol. 116, pp. 381~386.

10.1016/j.egypro.2017.05.085
13

Snyder, J.P. (1987), Map Projections: A Working Manual, U.S. Geological Survey Professional Paper 1395, United States Government Printing Office, Washington, D.C., USA.

10.3133/pp1395
14

Vega, A. (2020), Assessment of the Lifetime of District Heating Pipes, Doctoral Thesis, Chalmers University of Technology, Gothenburg, Sweden.

Information
  • Publisher :Korean Geo-Environmental Society
  • Publisher(Ko) :한국지반환경공학회
  • Journal Title :Journal of the Korean Geo-Environmental Society
  • Journal Title(Ko) :한국지반환경공학회 논문집
  • Volume : 27
  • No :11
  • Pages :15-22
  • Received Date : 2026-09-23
  • Revised Date : 2026-09-29
  • Accepted Date : 2026-10-15
  • DOI :https://doi.org/10.14481/jkges.2026.27.11.15
Journal Informaiton Journal of the Korean Geo-Environmental Society Journal of the Korean Geo-Environmental Society
Online Submission
  • NRF
  • KOFST
  • KISTI Current Status
  • KISTI Cited-by
  • orcid
Journal Informaiton Journal Informaiton - close