| GISdevelopment.net ---> GITA 1998 ---> Data Evolution |
Validating your GIS with Network Analysis
Patrick H. Dolan
Stoner Associates
P.O. Box 86
Carlisle, PA. 17013-0086
Patrick H. Dolan
Stoner Associates
P.O. Box 86
Carlisle, PA. 17013-0086
Introduction
Oftentimes the GIS facilities layer is called upon to support numerous applications. The challenge for GIS is to provide the correct data scheme and information needed to support the numerous applications. Questions that often arise concerning these additional applications include “How robust must the GIS facilities layer be? ’’and “Can it handle all the chores demanded of it?” If part of your GIS initiative is to perform complex analysis such as hydraulic network modeling, your facilities layer will be challenged in many ways. If the facility layer is not designed for the unique challenges of network analysis from the beginning, it is unlikely that it will be successful for modeling at the end. Therefore, it is critical to assess the data quality prior to network analysis to ensure accurate modeling results. To achieve this goal requires assessing the components necessary for network analysis. The components for hydraulic analysis are connectivity, facility information and demand information. The easiest and most complete way to test the quality of these three components is to use the network analysis tool during the data capture phase. Using network analysis as a part of your quality assurance and quality control (QAIQC) procedures fhrther ensures that your GIS facility information is complete and accurate. Incorporating network analysis as a part of your QA/QC also provides additional benefits to other related GIS applications.
QA/QC Work Flow
Incorporating network analysis as a part of the QA/QC work flow procedures is shown in the diagram below.
Figww 1: QA/QCwith Network Analysis
The diagram begins with accessing the facility information from GIS. Network analysis is then performed on the data set to identify potential problems. The results are then analyzed, and the problems are identified. The identified problems are grouped into two categories, data and design errors. Data errors are problems in the source information that are either inaccurate field information or missing information required for hydraulic analysis. Design errors are problems identified in the field where the GIS facility information is correct, but where network analysis has pinpointed potential weaknesses or breakdowns in the network’s design. Sorting the problems into these two categories helps identifi the action(s) needed to resolve each issue. If the problem is identified as a data error, the GIS source information is corrected and is once again run through network analysis until the remaining errors are within acceptable tolerances. If the problem is identified as a design error, then the modeler can further evaluate the issue and determine if corrective actions are needed. To further understand how using network analysis tools within QA/QC procedures can validate and correct GIS data, it is important to understand the three components that make up hydraulic models. Each are discussed in more detail below.
Connectivity
The first of the three major components in evaluating the ability of GIS data to support hydraulic modeling is connectivity. Connectivity plays a critical role in network analysis by enabling the model to identifi which pipes tie into one another to determine the flow direction and pressure fluctuation. Many tools have been developed within GIS applications to ensure that connectivity is being maintained during data capture and maintenance updates. These tools are generally based on tracing routines which start at a specific point along a pipe or a node and trace outward until it reaches a point were the pipe dead ends. These tracing routines are very helpfid in identifying orphan pipes segments. Orphan pipes are individual pipe segments that are isolated from the hydraulic network. However, in a network model, identifying individual isolated pipe segments is just the first level of connectivity needed for accurate analysis of a hydraulic model. Additional connectivity requirements are needed. For example:
Incorporating network analysis as a part of the QA/QC procedures will enable these additional connectivity issues to be identified and corrected. For example, if a bottleneck is identified in the GIS data set, the user must determine if the bottleneck is the result of inaccurate source information or if it is a legitimate design problem. If the problem is derived from incorrect source information, then the GIS facility layer can be updated with the correct pipe diameters. If the problem is a design problem, then the modeler can address it and determine what type of hydraulic impact this may have on the network. Identi$ing and correcting the connectivity issues ensure that the GIS data quality can support network analysis. In addition, other GIS applications will benefit. A few GIS applications that will benefit (they use the same connectivity rules used in hydraulic analysis) are: emergency valve shut down, cathodic protection and work-order design applications.
Facility Information
The second component in assessing data quality for modeling is the validity of the GIS facility information. Facility information is attribute data associated with each piece of equipment stored within the GIS. The validity of the facility information is a measure of how complete and accurate the information is. The validity of the data can be compromised during data conversion, the process of capturing source information from legacy data (paper maps, service cards). Data conversion focuses on ensuring that all critical information from the source documents are populated into the GIS database. The key is to have a one-to-one relationship between legacy data and the GIS database for all objects identified for conversion. This serves well in reproducing the current workflow processes that the legacy data once supported. However, since the legacy data was never designed to support network analysis, it is critical to identifi what additional information will be required to support these future GIS applications. A few examples of missing hydraulic information required for network analysis are:
Therefore, performing network analysis as a part of the QA/QC process can assist early on in resolving these missing or incorrect pieces of information. For example, for regulator stations, it is critical to determine whether or not the information about stations is complete for hydraulic modeling. After network analysis, the errors are identified as either a data or design error. If the error is identified as a data problem (missing regulator size) the additional information is acquired from other source documents and populated back into the GIS database. If the error is identified as a design error (too small of a regulator) the modeler can assess the situation and determine whether corrective action is necessary. In addition to hydraulic analysis, other applications and reporting tools can benefit from network analysis. For example, regulator and valve maintenance/inspection and inventory reporting are just a few applications that benefit from having complete and correct hydraulic information.
Demand Information
Demand information is the third component in evaluating the ability of GIS data to support hydraulic modeling. Demand information is the consumption information for each individual customer served by the network. The demand information determines what flows and pressures are required to sustain each customer along the network. When determining the magnitude of each customer’s load, it is critical that the load is properly assigned to the appropriate location in the network. The more accurately these assignments are made, the better the model will reflect real-world flow conditions, and the better the assessment will be of the impact of localized load changes. The process that is generally used within a GIS to assign customers to facilities is done through geocoding. During geocoding, you assign demand information defined by street address, or other address information, to facilities. This process is extremely useful in quickly assigning and identifying demands along the hydraulic model. However, this process has its own inherited accuracy issues that must be understood to ensure demand information is assigned appropriately for a hydraulic model. A few of those issues concerning demand assignment are:
Performing network analysis as a part of the QA/QC process can assist in identifying the accuracy of the demand assignment within the GIS facility layer. For example, once demand information is geocoded to the GIS facility layer, network analysis will be performed to derive the pressure and flow calculations. The modeler can then assess the results and determine if the pressure and flow calculations fall within the anticipated measured ranges. If the results do not fall within the ranges due to incorrect demand assignment from missing address information (data error), the demands can be pinpointed in GIS and reassigned appropriately. Also, if the modeler sees results that are out of range as a result of oversized demands at a particular point along the hydraulic model (design error), the individual can assess the situation and determine what corrective action(s) needs to take place. The benefits derived from using network analysis as a part of the QA/QC process ensure that an accurate model of the individual gas use characteristics exists and an accurate assignment of each customer to the hydraulic model exists. In addition to the benefits for hydraulic modeling, other GIS applications benefit as well. A few examples are, outage analysis, one call and marketing applications.
Conclusion
Network analysis has always demanded the highest level of data quality for hydraulic modeling. Harnessing the rigorous and thoroughness required for hydraulic analysis and incorporating it into the QA/QC process ensures that complete, accurate and efficient models are built from the GIS. In addition, other applications can benefit from implementing network analysis as a part of the QA/QC process. These applications benefit by using the same connectivity, facility and demand information required for hydraulic modeling. Yet, the greatest benefit derived from validating your GIS with network analysis is ensuring the GIS database quality needed to perform today’s complex analysis routines, and ensuring the ability to support tomorrow’s next generation of applications.
Oftentimes the GIS facilities layer is called upon to support numerous applications. The challenge for GIS is to provide the correct data scheme and information needed to support the numerous applications. Questions that often arise concerning these additional applications include “How robust must the GIS facilities layer be? ’’and “Can it handle all the chores demanded of it?” If part of your GIS initiative is to perform complex analysis such as hydraulic network modeling, your facilities layer will be challenged in many ways. If the facility layer is not designed for the unique challenges of network analysis from the beginning, it is unlikely that it will be successful for modeling at the end. Therefore, it is critical to assess the data quality prior to network analysis to ensure accurate modeling results. To achieve this goal requires assessing the components necessary for network analysis. The components for hydraulic analysis are connectivity, facility information and demand information. The easiest and most complete way to test the quality of these three components is to use the network analysis tool during the data capture phase. Using network analysis as a part of your quality assurance and quality control (QAIQC) procedures fhrther ensures that your GIS facility information is complete and accurate. Incorporating network analysis as a part of your QA/QC also provides additional benefits to other related GIS applications.
QA/QC Work Flow
Incorporating network analysis as a part of the QA/QC work flow procedures is shown in the diagram below.
Figww 1: QA/QCwith Network Analysis
The diagram begins with accessing the facility information from GIS. Network analysis is then performed on the data set to identify potential problems. The results are then analyzed, and the problems are identified. The identified problems are grouped into two categories, data and design errors. Data errors are problems in the source information that are either inaccurate field information or missing information required for hydraulic analysis. Design errors are problems identified in the field where the GIS facility information is correct, but where network analysis has pinpointed potential weaknesses or breakdowns in the network’s design. Sorting the problems into these two categories helps identifi the action(s) needed to resolve each issue. If the problem is identified as a data error, the GIS source information is corrected and is once again run through network analysis until the remaining errors are within acceptable tolerances. If the problem is identified as a design error, then the modeler can further evaluate the issue and determine if corrective actions are needed. To further understand how using network analysis tools within QA/QC procedures can validate and correct GIS data, it is important to understand the three components that make up hydraulic models. Each are discussed in more detail below.
Connectivity
The first of the three major components in evaluating the ability of GIS data to support hydraulic modeling is connectivity. Connectivity plays a critical role in network analysis by enabling the model to identifi which pipes tie into one another to determine the flow direction and pressure fluctuation. Many tools have been developed within GIS applications to ensure that connectivity is being maintained during data capture and maintenance updates. These tools are generally based on tracing routines which start at a specific point along a pipe or a node and trace outward until it reaches a point were the pipe dead ends. These tracing routines are very helpfid in identifying orphan pipes segments. Orphan pipes are individual pipe segments that are isolated from the hydraulic network. However, in a network model, identifying individual isolated pipe segments is just the first level of connectivity needed for accurate analysis of a hydraulic model. Additional connectivity requirements are needed. For example:
- Bottlenecks: Sections of pipe where reduction occurs from a large pipe diameter to a smaller pipe diameter. This results in a “bottleneck” effect where gas is so constrained at that point that pressure is unable to be maintained downstream due to the constriction.
- Subsystems: Small network of pipes disconnected from the main network. These subsystems can impact your analysis by not accurately representing all of the associated connections to the main model.
- Pressure system interconnections: This occurs when a low pressure system is tied back to the high pressure system without any device regulating or isolating the two separate pressure systems. This affects the model by not accurately representing the pressure and flow directions.
Incorporating network analysis as a part of the QA/QC procedures will enable these additional connectivity issues to be identified and corrected. For example, if a bottleneck is identified in the GIS data set, the user must determine if the bottleneck is the result of inaccurate source information or if it is a legitimate design problem. If the problem is derived from incorrect source information, then the GIS facility layer can be updated with the correct pipe diameters. If the problem is a design problem, then the modeler can address it and determine what type of hydraulic impact this may have on the network. Identi$ing and correcting the connectivity issues ensure that the GIS data quality can support network analysis. In addition, other GIS applications will benefit. A few GIS applications that will benefit (they use the same connectivity rules used in hydraulic analysis) are: emergency valve shut down, cathodic protection and work-order design applications.
Facility Information
The second component in assessing data quality for modeling is the validity of the GIS facility information. Facility information is attribute data associated with each piece of equipment stored within the GIS. The validity of the facility information is a measure of how complete and accurate the information is. The validity of the data can be compromised during data conversion, the process of capturing source information from legacy data (paper maps, service cards). Data conversion focuses on ensuring that all critical information from the source documents are populated into the GIS database. The key is to have a one-to-one relationship between legacy data and the GIS database for all objects identified for conversion. This serves well in reproducing the current workflow processes that the legacy data once supported. However, since the legacy data was never designed to support network analysis, it is critical to identifi what additional information will be required to support these future GIS applications. A few examples of missing hydraulic information required for network analysis are:
- Regulator Stations: Regulator stations are hydraulic structures which regulate pressure. The legacy data generally identifies the geographic location of these stations. However, performing network analysis requires specific information about the individual regulators within the station. Generally this information is either not stored in the GIS, or the information is limited in its content.
- Valves: Valves are devices which isolate or direct flow. These devices play a critical role in assessing the ability to control the pressure and flow of gas to customers. Generally, the GIS data will identifi the location of these devices and the associated pipe on which the valve is located. However, if valve information is missing or incorrect, the hydraulic results from network analysis will not reflect the true conditions in which the system is operating.
Therefore, performing network analysis as a part of the QA/QC process can assist early on in resolving these missing or incorrect pieces of information. For example, for regulator stations, it is critical to determine whether or not the information about stations is complete for hydraulic modeling. After network analysis, the errors are identified as either a data or design error. If the error is identified as a data problem (missing regulator size) the additional information is acquired from other source documents and populated back into the GIS database. If the error is identified as a design error (too small of a regulator) the modeler can assess the situation and determine whether corrective action is necessary. In addition to hydraulic analysis, other applications and reporting tools can benefit from network analysis. For example, regulator and valve maintenance/inspection and inventory reporting are just a few applications that benefit from having complete and correct hydraulic information.
Demand Information
Demand information is the third component in evaluating the ability of GIS data to support hydraulic modeling. Demand information is the consumption information for each individual customer served by the network. The demand information determines what flows and pressures are required to sustain each customer along the network. When determining the magnitude of each customer’s load, it is critical that the load is properly assigned to the appropriate location in the network. The more accurately these assignments are made, the better the model will reflect real-world flow conditions, and the better the assessment will be of the impact of localized load changes. The process that is generally used within a GIS to assign customers to facilities is done through geocoding. During geocoding, you assign demand information defined by street address, or other address information, to facilities. This process is extremely useful in quickly assigning and identifying demands along the hydraulic model. However, this process has its own inherited accuracy issues that must be understood to ensure demand information is assigned appropriately for a hydraulic model. A few of those issues concerning demand assignment are:
- Data quality of street addresses: The quality of street address information drives how successful and accurate demands will be assigned to facilities. If the street address information is limited in its information (lack of alternate streets names) or is incomplete (has missing address ranges) the geocoding process will result in either demands being assigned incorrectly or demands going unassigned to the network model. Demand information that is incorrectly assigned or missing will erroneously alter the flows and pressures of the hydraulic model.
- Front or back lot assignment: When assigning loads to the network, the concern is always that the demand information might be assigned to the wrong pipe segment. The geocoding process pinpoints customers based on address ranges which are generally stored on a street centerline layer within the GIS. Since the address ranges are not stored directly on the pipeline, the customer assignment must go through two levels of identification. The first level assigns the customer to a street. The second is to assign the customer to the nearest pipe. This process of assigning demands to the nearest pipe along a street centerline can result in customers being assigned to a main running in front of their house when they are actually serviced by the gas line running next to or behind their property. Though this occurrence may be minimal, the impact of this incorrect assignment will alter the flow and pressure of the hydraulic model. Since large loads have the greatest impact on the hydraulics of a system, extra care must be taken to ensure that demands are properly assigned.
Performing network analysis as a part of the QA/QC process can assist in identifying the accuracy of the demand assignment within the GIS facility layer. For example, once demand information is geocoded to the GIS facility layer, network analysis will be performed to derive the pressure and flow calculations. The modeler can then assess the results and determine if the pressure and flow calculations fall within the anticipated measured ranges. If the results do not fall within the ranges due to incorrect demand assignment from missing address information (data error), the demands can be pinpointed in GIS and reassigned appropriately. Also, if the modeler sees results that are out of range as a result of oversized demands at a particular point along the hydraulic model (design error), the individual can assess the situation and determine what corrective action(s) needs to take place. The benefits derived from using network analysis as a part of the QA/QC process ensure that an accurate model of the individual gas use characteristics exists and an accurate assignment of each customer to the hydraulic model exists. In addition to the benefits for hydraulic modeling, other GIS applications benefit as well. A few examples are, outage analysis, one call and marketing applications.
Conclusion
Network analysis has always demanded the highest level of data quality for hydraulic modeling. Harnessing the rigorous and thoroughness required for hydraulic analysis and incorporating it into the QA/QC process ensures that complete, accurate and efficient models are built from the GIS. In addition, other applications can benefit from implementing network analysis as a part of the QA/QC process. These applications benefit by using the same connectivity, facility and demand information required for hydraulic modeling. Yet, the greatest benefit derived from validating your GIS with network analysis is ensuring the GIS database quality needed to perform today’s complex analysis routines, and ensuring the ability to support tomorrow’s next generation of applications.