Our Insights

Utility Network vs. Geometric Network (First in a series)

September 19, 2019

Utility Network – Not as Scary as it Seems

By Brittany Dizdar, Senior Consultant and Glenn Farrow, Senior Consultant, POWER Engineers

First in a series of articles with insights into the Utility Network.

Now that the Utility Network has been released and is being used by utilities in production, let’s gain some insight by discussing some of the similarities and differences in the geodatabase between the geometric network and the Utility Network.

A Utility Network contains one or more domain networks (electric, gas, water, etc..) and exactly one structure network. We will elaborate on what domain and structure networks are in a later article, but for now let’s focus on the feature classes that make up the Utility Network in the geodatabase. To aid in presenting the material in this first article we will focus on the electric model.

The electric Utility Network consists of standardized feature classes that are configured to model a real-world electric utility — as the facilities actually exist in the field. The ability to more accurately depict the real-world network is one of the major selling points of the Utility Network.

In examining the electric Utility Network in the geodatabase, one of the first things we notice is that the number of feature classes has been significantly reduced from what we would see in a geometric network model. It is important to realize that while there are fewer feature classes than before, they are the same point, line, and polygon feature class types with which we are all familiar. The Utility Network is built on the same Esri geodatabase technology that we use today with our geometric networks.

Figure 1: Utility Network feature classes for the electric domain at left compared with typical feature classes in an electric geometric network at right

If you are currently using one of the standard electric models for the Geometric Network, there may be 18 or more feature classes utilized in the model. Contrast that with the Utility Network which utilizes five electric feature classes, plus supporting classes for service areas and structures.

One principal benefit of having fewer feature classes is that it requires fewer round trips to the geodatabase when retrieving or updating data. This is especially important for maintaining performance given the services-based architecture of the Utility Network. The remainder of this article will focus on how all of the electric assets and facilities are modeled in the constrained set of Utility Network feature classes.

Data Modeling in the Utility Network

What is the purpose of the core electric feature classes in the Utility Network and how are they utilized to model the electric network?  Where are the transformers, switches, conductors and other facilities that we are accustomed to dealing with? In the geometric network it was easy to identify the different assets because it was reflected in the name of the feature class, but in the Utility Network this is not the case. Let’s explore the usage of these Utility Network feature classes further.

The ElectricAssembly feature class is used to represent containment facilities, or features that contain other Utility Network features. For example, consider a transformer bank that contains individual transformer units. Or think of an electric substation that contains transformers, switches, busbars and other assets. The contained facilities are first class features in their own right, they are just understood to be contained within a container or assembly. We will discuss containment and containment associations in more detail in a later article.

The ElectricDevice feature class is used to model point features that can influence the electrical circuit flow such as transformer units, individual switches/fuses, voltage regulators, or service points. In the geometric network some of these devices were represented in non-graphical tables or object classes, but now in the Utility Network they are all maintained in the device feature class and are part of the physical model as first-class spatial features This helps us to be able to accurately model the real-world network connections within the GIS.

The ElectricJunction feature class can be thought of as modeling non-operational connections or device entry points such as grounds, line ends, terminals and risers. Junctions are used to connect lines where there is no intervening device, or to model line ends. For example, a riser might be placed at a location where there is a transition from overhead lines to underground lines.

The ElectricLine feature class is used to model all of the linear electric network features, such as primary conductors, secondary conductors and busbar. Finally the ElectricSubnetLine feature class, which is system maintained and read-only, is used strictly for network visualization purposes.

Asset Groups and Asset Types in the Utility Network

So how is it possible to model a complex electrical network that previously required 18 or more feature classes, in just five Utility Network feature classes? This is where the Utility Network concepts of Asset Groups and Asset Types apply.

Think of Asset Groups as analogous to the numerous feature classes in the old geometric network. There is an Asset Group for each type of device that previously had its own feature class in the geometric network, such as transformers or switches. Asset Group is the major classification applied to the feature classes that make up the Utility Network.

Figure 2: Assembly Asset Groups in the electric domain model

While the Utility Network Asset Groups categorize each electric component within the five feature classes at a high level, this is still not specific enough. In the geometric network we were able to further refine component categorization by specifying a subtype. It is not sufficient to recognize a device as a transformer, we need to distinguish between service transformers, step transformers, grounding transformers, etc. In the Utility Network this level of categorization is accomplished using Asset Types, which are analogous to the subtypes used in the geometric network. Asset Types in the Utility Network allow us to more finely classify our electric assets as shown below for transformers.

Figure 3: Transformer Asset Types

Note that since all devices reside in the same feature class, differentiated by Asset Group, all devices share a common set of attributes. The same applies to all linear features maintained in the single line feature class and all junctions maintained in the single junction feature class. We will discuss the ramifications of this data modeling choice further in a future article.

Hopefully this article has helped you to gain a better understanding of the Utility Network feature classes for the electric domain by relating to concepts we are all familiar with. While the Utility Network does introduce new concepts, it is still based on much of the same Esri technology we have been using for decades. Note that the gas and water domains have a similar set of five tables as described here for electric and may be present within the same Utility Network as the electric domain.

The standardization of the electric Utility Network as modeled in the feature classes described above has great potential benefit for all electric utilities. In our experience, each electric utility currently has their own unique data model and it is often difficult to collaborate because of these differences. While the Utility Network models are designed to be extensible, the hope is that the data models will remain more standardized across utilities, eliminating some of that difficulty and promoting more collaboration between utilities and vendors.

Next in this series:  We will explore additional Utility Network concepts such as domain networks and tiers. 

About the authors:

Brittany Dizdar is a senior consultant with the Geospatial and Asset Management Division at POWER Engineers, Inc. She is currently responsible for the electric network and equipment data from many repositories that feed the advanced distribution management systems at several utilities. Specializing in geographic information, her work spans across both Information Technology (IT) and Operational Technology (OT). Prior to her seven years at POWER, she had long stints at SCANA and Arizona Public Service. Brittany has a B.S. in Computer Science from University of South Carolina (which she refers to as “the real USC”).

Glenn Farrow has been a Senior Consultant with the Geospatial and Asset Management Division at POWER Engineers since 2009. He has been developing solutions using Esri software in the environmental, municipal, and utility sectors since 1984, with a focus on utilities and telco since 2000. Glenn has a BSc. in Forestry from the University of Alberta.