Gas Outage Management Part 1

Blog Post created by tdewitte-esristaff Employee on Mar 22, 2019

Getting Started By Identifying Customers Impacted

By Tom Coolidge and Tom DeWitte


News of a gas outage can arrive from various sources.  It can come from a sensor indicating an abnormal condition.  Maybe it comes from a customer calling into customer service.  Or, a contractor calling operations after an excavation mishap.  Another possibility is a citizen calling in to report gas odor at a location.  Regardless of the source of the outage news, confirmation of an outage triggers one of a gas utility’s priority processes – restoring safe and reliable service to customers.


As important and critical a task as gas outage management is to a gas organization and to the community it supports, this process has changed little over the last 100 years.  For many gas organizations, it can take several hours to identify which customers have been impacted.  Once the customers are identified, getting the list of impacted customers to the field gas operations staff is still primarily a paper process. Someone literally must get into a vehicle and drive the list of customers to the location of the gas outage event.  As the field gas operations staff begins the gas relight process, they too still tend to use paper to document the status of each customer.  This means that management will always have a delayed understanding of the progress of restoring gas service.  When the mayor or governor calls asking for an update, gas executives are often get caught with little current information to pass on.


There has got to be a better way.


And, there is.  In fact, most of the gas industry already possesses the software to resolve these issues and significantly improve a gas organization’s response to a gas outage event.  The software I am referring to is the ArcGIS software currently widely used by gas organizations around the world.  This blog is the first in a series of three blogs explaining how the standard capabilities of the ArcGIS software can be deployed to address these common gas outage management challenges.  All functionality described in these blog articles are standard capabilities available today.  No customization or coding is needed. 


This first blog addresses the issue of identifying the customers impacted by a gas outage event.  This task often takes several hours when it needs to be accomplished in minutes. Additionally, the historical processes have had problems with accurately identifying the impacted customers and communicating precisely where those customer meters are located. 


The second blog will address the issue of communicating the list of impacted customers to the gas operations field staff.  The typical paper process takes too much time, causing delayed field operations and lower customer satisfaction. 

The third blog will address the gas relight process.  This process is also typically performed with paper.  The use of paper to track and communicate progress adds difficulty and inefficiency to this process.  The use of paper not only engrains a delay in relaying the update status to gas management and other interested parties, it also inputs a delay in relaying the status of individual meters between deployed field staff.


Identifying Impacted Customers

Current methods used by many gas organizations are lacking in accuracy and timeliness when identifying the customers impacted by a planned or unplanned gas outage.  One common method is to use the Customer Information System (CIS) for identifying impacted customers.  Since a CIS typically lacks an understanding of the connectivity of the pipe system, it is forced to rely on street address ranges.  The use of address ranges is inaccurate.  At every street intersection are four corner parcel lots.  Whether they are included in the address range is dependent on what street the house is listed under. This inaccuracy often requires a time-consuming manual process of having someone review the list, identify all crossing streets within the address ranges, determine the address ranges of those crossing streets, identify the corner lot addresses, then determine for each corner lot, whether it gets its gas from the impacted line, or from the gas line running down the cross street.  


Another common method is to use flow analysis systems to perform an isolation trace to identify the impacted customers.  This process is quicker, but it too is imprecise. The imprecision is due to the flow analysis software’s requirement to cluster groups of customers onto the gas pipe system at a singular location even though they each have individual service lines connecting to the gas main at discrete locations. In today’s gas pipe systems, the majority of gas mains are constructed of pinchable polyethylene plastic pipe. A gas event can be isolated or pinched at nearly any point along the plastic gas main.  The clustering of customer locations along the pipe system creates an inherent conflict between where gas operations places a clamp to pinch the pipe, and where the flow modeling engineer chose to aggregate the cluster of customers. This conflict creates an inaccuracy in the identification of impacted customers.


Accurately and quickly identifying impacted customers

The solution to addressing this problem is to use a system that understands the connectivity of the entire pipe system from its source, such as a town border station, to its end destination at the customer meter. ArcGIS provides the ability to maintain a connected representation of the entire pipe system, and the ability to perform a gas isolation trace to identify the meter or meter sets impacted by a gas outage. To perform this trace, you will require the following software:

  • ArcGIS 10.2.1 or higher, with a geometric network


  • ArcGIS Pro 2.3 or higher, ArcGIS Enterprise 10.7 or higher with a utility network


Additionally, your ArcGIS representation of the gas pipe system will need to model the following gas system assets:

  • mains
  • services
  • isolation valves
  • regulator stations (if regulator station valves are not individually mapped)
  • town border stations (if town border station valves are not individually mapped)
  • meters or meter sets


NOTE: If using meter sets you will need a link to a table identifying all meters contained within the meter set. This table is often an extraction of information from the Customer Information System


Your mains and services will at a minimum need to include the material of the pipe, so pinchable pipe can be differentiated from non-pinchable pipe.


The Gas Isolation Trace

The gas isolation trace is a more complex trace algorithm than simply identifying those pipes connected to the location of the pipe system failure, which are also between isolating valves.  With most gas pipe systems, the network is deliberately looped, to provide multiple sources of gas to any given location in the pipe system.  If this were true for every location on the pipe system, a simple connected trace defined to stop at barriers such as isolating valves or pinch points would be all that is needed.  But, there are portions of most gas systems where locations have only one source of gas.  Think of a gas pipe running along a dead-end street or a cul-de-sac.


If there is an isolating valve or pinch point at the location where the single feed pipe subsystem integrates with the larger looped pipe system, then the simple connected trace would ignore the customers on the downstream side of the barrier.  A more intelligent trace algorithm is required.  This more intelligent trace algorithm is generally referred to as the gas isolation trace.  A gas isolation trace is a multi-trace trace.  This means that the isolation trace runs a series of traces.  The first trace is the connected trace to identify the barriers (isolating valves and specified pinch locations).  Then a second round of traces is performed for each selected barrier.  This second round of traces is checking to verify that there is a source of gas feeding the barrier from the opposing side of the barrier.  This is to identify those dead-ends which do not have access to another source of gas.  Those customers downstream of the barrier on the dead-end need to be included in the list of customers impacted by the outage.


Gas Isolation Trace tools

The ArcGIS gas user community is fortunate in that there are multiple options for tools which can perform this industry specific type of trace.


One option is to download the free Gas utility editing tools provided by Esri. This ArcMap Add-In is available from the following Esri web site:

Another option is to leverage ArcMap Add-In tools from one of our business partners, such as Schneider Electric or Magnolia River.


For the ArcGIS Pro environment leveraging the utility network, this trace is a base capability as of the ArcGIS 10.7 release.


Identifying Impacted Customers

Operating the gas isolation trace tool is not complicated.  Simply identify the estimated location of the pipe system failure on the map.  In GIS speak this is called placing the flag to identify the start location of the trace.


When the isolation trace is run it will select all customers within the impacted area.  In my screen shot below you can see that this initial run selects over 100 impacted customers.


Identifying the location of pinch points

The prevalence of pinchable polyethylene plastic pipe enables the additional capability to reduce the number of impacted customers, by applying a gas clamp to pinch the pipe and stop the flow of gas to the location of the pipe system failure.  To represent this field capability in the GIS system, place a barrier at the location being considered for the pipe clamp.


With the proposed location(s) of the pipe clamp(s) now identified, the isolation trace is run a second time.  This time the resultant list of impacted customers has been reduced to less than 20.

The person running the analysis for both traces has so far only invested a few minutes of their time.  In that short time an accurate list of impacted customers has been created.


Defining the extent of the gas outage event

In today’s always connected, smartphone world, gas executives and managers expect to be able to access critical information that is easy to understand.  They generally do not need to see the list of individual customers impacted, often all they want to know is “where is the outage”, and “how many customers are impacted.”

By identifying the list of impacted customers with the ArcGIS tools, it is very easy to run an additional step to generate a polygon to define the boundary of the event.  In the GIS, a tool such as the Minimum Boundary Geometry geoprocessing tool will perform this task.

The creation of an event area feature provides a clear visual understanding of where this outage is occurring.  Having this singular feature representation also provides an intuitive means for managing event summary information, such as duration, and count of impacted customers.  The Esri-provided gas isolation tools automatically generate this polygon as part of the operation of the isolation trace.  In addition to the automatic generation of the polygon, a of every meter is generated and assigned an event ID to automatically relate the impacted customers to this specific event.

With the list of impacted customers defined and created, as well as the event bounding polygon, this information is ready to be electronically shared to gas operations field staff.


In the next blog, the 2nd blog of this blog series, the issue of delivering this list of impacted customers will be addressed.



ArcGIS today is deployed worldwide at many gas organizations, providing the ability to replace and improve upon non-spatial legacy processes.  Identifying impacted customers, whether they are connected by steel pipe or pinchable plastic pipe, can be accomplished in just a few minutes.  Using the ArcGIS tools can provide a more accurate list of impacted customers than is available via legacy methods.  This list not only identifies who has been impacted, it also clearly and accurately identifies where those impacted customers are located.   


PLEASE NOTE: The postings on this site are my own

and don’t necessarily represent Esri’s position, strategies, or opinions.