Data-Driven Prediction and Reduction of Excavation Damages
Léon olde Scholtenhuis
|| Organisation: Kadaster + Agentschap Telecom + KPN
P.D.Eng Jiarong Li
Update: July 2022
Workflow of the generation of KLIC Schaderapport
This project aims at providing strategies to reduce damages on underground pipelines and cables caused by excavation works. The objective is to develop a data-driven model or platform to predict the damage caused by construction work. This project takes the hypothesis that the (spatial) data and historical damage data might be able to explain partially the causes of damage to utilities
In the problem investigation step, we gain an understanding of the phenomenon of excavation process and underground utility, and explore causal factors of utility damages, through a literature review, interviews and workshops with experts.
I next steps, we will collect datasets of excavation damage, underground network and other geo-data from Kadaster KLIC and other organizations. Statistica analysis reveals the patterns in damage data can be investigated, and we consecutively build a model to predict excavation damages using data mining techniques like machine learning. From a practical viewpoint this (may) lead(s) to identification of sensitive factors found can contribute to providing strategies to reduce damages.
Approach: DM-Cycle