Posted on Sunday, April 3rd, 2016 in Use cases
It is early in the morning when an excavator hits a water pipe in a residential area in Apeldoorn. In the following hours, thousands liters of water flush away the soil supporting a nearby gas pipe and it subsequently breaks. Sequentially the gas pipe gets flooded with water and dirt and over a total length of 6km of gas pipes are polluted. It took a week to clean out all the pipes, causing 1317 households to be cut off from gas. In hindsight, the initial damage of restoring a broken water pipe is €1000,- to €5000,- depending on the type of pavement, extra costs for excavators, etc. These costs are disproportionate to the sequential repair costs of more than €500k for the Dutch gas network operator, Liander.
This case is not unique. Low pressure gas pipes in the proximity of water mains are often at a high risk of being damaged. Scanning the networks of both water network operators and gas network operators for locations with conditions similar to historical damage cases can deliver the rationale for preventive measurements. Reinforcing gas-pipes on X locations at highest risk may prove to be a fraction of the costs of just one of them breaking.
3D Simulation of a water main breaking
For the simulation, the 3D network data of the network operator is complemented with, amongst others, Geotop data. When in depth information regarding, for instance a water pipe or one of the other network elements is not available, we will use AHN data, preferably AHN3 as of the high-quality classifications. For optimum performance we store the data in their own data model. This is much faster to query than standard coverage services.
While loading the networks into the simulation platform, we prefer to obtain the data via distributed data models as much as possible. This allows you to work with current data, directly retracted from the network administrator in question. To pair with other relevant data sources and to achieve optimum results across multiple domains, we use the IMKL data model, complemented by network specific elements such as coats, cores, pressure and flow. The network is stored as a geometric network provided with geocoder classes. Using the application, the user can access all network data, record incidents, or run scenarios. In this case, the rupture of a water main.
Data on the incident, including 3D geo-coordinates and pipe data, together with (3D intersect) data regarding the surrounding ground is sent to the physical model for scour holes and sinkholes. Based upon actual water pressure and flow, the de-sanding is calculated. In essence, how fast will how much soil wash away? The computations are used to create a 3D dynamic feature (3D buffer) of the event in the GIS. Using the 3D buffer, we can analyse and assess the impact of the incident on other network structures or networks elements. “Is there a gas pipe inside the 3D buffer? Will it resist the pressure or will it pop out of its sleeve?”. By performing an intersect, we can see whether or not water flows into the gas pipe and query which network elements will be affected. Based on this data, the network operator can take adequate action.
Simulation scenarios, visualizations and applications are modelled to real world situations and practices. SIM-CI has launched pilot projects with several network operators. They provide historical incident information and current network data and hands-on experiences of people in the field. By doing this, academic models can be translated into valuable tools for managing and securing critical infrastructures.