| GISdevelopment.net ---> GITA 2001 ---> A Tangled Web of Pure Opportunity |
Trouble call intigration without an outage
Larry R. Schantell
Senior Application Specialist
UtiliCorp United
Kansas City, Missouri
Robert W. Rozen
Senior Technical Manager
GE Smallworld Spatial Services
Westminster, CO
Larry R. Schantell
Senior Application Specialist
UtiliCorp United
Kansas City, Missouri
Robert W. Rozen
Senior Technical Manager
GE Smallworld Spatial Services
Westminster, CO
Introduction
UtiliCorp United is an international energy company with its headquarters in Kansas City, Mo. It operates in the United States, Canada, United Kingdom, Spain, Germany, Norway, New Zealand and Australia. UtiliCorp has over 4 million customers in its domestic and international operations. For 1999 it was ranked the second-largest wholesale marketer of electricity in the U.S. and the third-largest marketer of natural gas. On the Fortune 500 list, UtiliCorp ranked 90th based on 1999 sales. The company is also included in Fortune's list of America's Most Admired Companies. UtiliCorp’s United States operations are concentrated in the Mid-West, largely in Colorado, Kansas, Nebraska, Missouri, Minnesota and Michigan.
Like many long established utilities, UtiliCorp’s computer systems had evolved in an unplanned manner, resulting in a variety of systems on different computers, ranging from PCs to mainframes. Moving and maintaining information between all these disparate systems was a chore. UtiliCorp wanted to have a common solution for its call centers, field resource centers (FRC), and other facilities. They also wanted to improve customer service by providing outage status, estimated restoration times, and identification of problems areas. The reduction of maintenance expenses was another important consideration. Whatever the solution, it was very important that current operations not be affected.
Existing processes and components
A major component in UtiliCorp’s existing architecture is the Facilities Management & Mapping Enabler (FAME). FAME is a GE Smallworld-based AM/FM software system that is used to inventory, model and help manage gas and electric networks. The FAME system accesses data within the Smallworld GIS that maps and inventories UtiliCorp’s gas and electrical facilities throughout its Mid-Western territory. Some of the projects that were being implemented within the FAME framework included; outage incidence recording, distribution interruption reporting, customer information system (CIS) interface, and SCADA data availability.
UtiliCorp operates two service centers staffed by UtiliCorp personnel that handle the normal volume of service/trouble calls. Trouble call tickets were entered into a Microsoft Outlook e-mail form and then sent, via a series of Microsoft Exchange servers, to all the appropriate UtiliCorp sites and personnel. UtiliCorp also contracts with 21 st Century Communications to provide High Volume Call Answering, (HVCA), services. 21 st Century Communications can process 20,000 calls per hour. The HVCA process is used by UtiliCorp to help in answering problem calls when the demand becomes greater than what UtiliCorp’s trouble call centers can handle. Typically this process is utilized whenever serious, widespread outages occur. In UtiliCorp’s Mid-Western territory the widespread outages are most often weather related and are caused by the serious thunderstorms, tornadoes and ice storms common to that region. The initial implementation of the HVCA system generated an ASCII file which was passed every five minutes from the vendor to UtiliCorp, where the ticket information was parsed, entered in the Exchange form and sent using Smallworld OLE processing.
Part of the existing trouble call reporting system was a PC based database system that was written using FoxPro. This was a single user system that required someone to manually enter outage data after an incident had occurred. This was a very slow, time-consuming effort and one that allowed the introduction of errors during the data entry process. Compounding this data entry problem, the FoxPro system did not have any data validation routines and there was no link between the FoxPro system and the facility model. Thus, there were problems with the timeliness of the data, the accuracy of the data, and the data existed in its own space without any links to other systems
UtiliCorp was also facilitating a link between the mainframe based Customer Information System, (CIS), and FAME. This link would allow access to customer information and usage via the graphic facility model.
Problems with Existing Processes and Components
There were many problems with the existing trouble call process. One problem was the slowness of the Exchange server. This was mainly the result of having to replicate large volumes of data across the network to all the Exchange servers. Another problem was the reliability of the Exchange network.
The FoxPro outage incidence recording and outage incidence reporting routines were being re-engineered as separate and independent projects, utilizing an Oracle database. This project was being implemented with no tie to the trouble call process. The same pieces of information were being captured and stored with different names and different data types throughout the unrelated systems. Using FAME to facilitate an outage management system required existing and valid linkages between the graphic data and the CIS data. Only 10% of the customer data were linked with graphic data, making it very difficult to associate trouble calls to electric devices or predict and close an outage. It was estimated that it would take three years to validate graphic data, verify electric network connectivity, and create valid linkages between FAME and CIS.
Proposed Solution
After examining the problems, UtiliCorp decided it needed an integrated solution that could be accessed from the call centers, field resource centers, and the main UtiliCorp offices. An initial investigation of the GE Smallworld PowerOn product generated several roadblocks. The PowerOn application was perceived to be very disruptive to existing processes, provided more functionality than currently needed, required more training and more time to implement, and the cost was not acceptable to the user community.
It was clear that a more compatible and cost efficient implementation was needed. To achieve this goal it was decided to use the power of the Internet to link these sites to a common database. A thin-client/thick server setup could provide a common solution to all users and help in providing better customer service, identification of problem areas, minimization of maintenance expenses, and making outage statuses and restoration times available. After a thorough investigation, it was determined that an Oracle database would be the central component in the new architecture. This web-based application was later named ServiceOn.
Proposed Processes & Components
Another component of the proposed solution was to link the trouble call database to the FAME GIS system. Using the outage information gathered by the services, the data could then be used to create real-time displays of outage s areas, help predict problem devices, help in dispatching field crews, and help relay service information back to the customer.
ServiceOn
ServiceOn is a web-based trouble call system that was developed by UtiliCorp and GE Smallworld personnel to help solve some of the problems with the existing trouble call system. Some advantages of UtiliCorp’s ServiceOn implementation are that the input screens have the same look and feel as the old trouble call data entry screens, and the process flow is the same. These features, combined with the browser interface, present the service centers users with a system that requires a minimal amount of training and documentation. Because of the thin-client architecture, ServiceOn is faster and more reliable than the old system.
One of the key features of ServiceOn is that it uses an Oracle database as its data storage server. Having the data in a server-based Oracle database means that the data is more accessible to a wide variety of UtiliCorp users. Also, the users now have a faster response time because of the capabilities of the Unix server that hosts the Oracle database.
On the ServiceOn data query panel, the user has the ability to select pre-defined queries of the database using one or more of the following items: city name, state abbreviation, zip code, address, phone number, ticket label, status, and the definition of the begin and end date/time. From this query panel the user can select fields with which to sort the report. The user is also given a number of report display types to choose.
The results of the ServiceOn query are displayed on a series of reports. The Query Results Panels provide users the ability to review the original trouble ticket and to update that information. There is a monitoring option on the Query Results Panel where the data can be refreshed every 30 seconds. This option allows the users to watch for new trouble tickets without any input. Another Query Results Panel option is to query and display the trouble tickets on their status of “First Call”, “Predicted”, “Confirmed”, “Dispatched”, “Resolved”, or “Closed”. Combined with the date/time option, this gives UtiliCorp the ability to only look at tickets that are currently been worked on or that have been recently completed.
One of the ServiceOn reports is the Field Resource Center (FRC) display. This report uses an existing FRC format, minimizing user training. The report allows the users to access individual trouble tickets and to change the status on them. Another report type is the Review by Feeder report, which displays information by substation and feeder data. On the Review by Ticket Label Report, the users can group individual trouble tickets by ticket label. On this report, the user can summarize trouble tickets with same label display count. They have the ability to state transition tickets by ticket label and can reassign tickets to new or existing ticket labels.
ServiceOn provides the user the ability to capture the outage incidence information via the Distribution Interruption Report form. On this form the user can “close” the outage form and enter in data pertaining to the customers affected, such as weather conditions or components that failed, and select the cause of the outage from a set list. Data entered on this form is used to create outage incidence reports that show when and where an outage occurred, what was the failing component, what customers were affected, the feeder involved, and the duration of the outage including the first call, dispatched, and completed times. These reports can be sent to regulatory authorities and used for other performance-based asset management reports.
ServiceOn procedures are not dependent on FAME for graphic validation, but enhanced by the graphic display. If a valid link to a graphic device is not found, FAME will create a default label based on work district and zip code and/or the user can assign a ticket label to associate related trouble tickets. If valid linkage is found, FAME will associate related data, assign an intelligent ticket label, and create graphic components to display the outage.
ServiceOn Data Loader
Part of the proposed solution was to devise a method by which data from the High Volume Call Answering (HVCA) was input into the ServiceOn database. This method was a custom built program called the ServiceOn Data Loader. The HVCA process collects data in a text file. This file is sent to a server on UtiliCorp’s network via an FTP process. As soon as one of these files is written to the server, it is read-in by the ServiceOn Data Loader. This program reads in the text file, translates the data from the HVCA format to the ServiceOn format, inserts the data into the ServiceOn Oracle database, and archives the original file. This whole process takes only seconds for a file that may contain a thousand records.
Fame Enhancements
The original intent to use the FAME Smallworld GIS system for facilitating outage management was re-engineered so that FAME provides a real-time view of active outages. This real-time background processing of each trouble ticket record identifies the graphic location of an outage. FAME outage prediction logic associates outage locations to the electric facility model and graphically displays the network affected by the outage. This FAME prediction logic also groups the individual trouble tickets with a common ticket label. FAME outage analysis assists in the management of outages by changing the graphic display relative to the outage life cycle status: First Call, Predicted, Confirmed, Dispatched, Resolved, or Closed.
Another advantage of easily locating network outages is improved customer service. Now customer can be told of the extent of the outage, the probable cause, and the estimated recovery time. The maps can also be used to identify problem feeders and/or components. This ability helps minimize maintenance expenses by initiating needed maintenance projects before an outage occurs.
Reports Enhancements
BusinessObjects is used to generate reports and statistics on outage incidences from the Oracle database. These reports show when and where the outage occurred, the failing components, the customers affected, and the feeders involved. The reports also contain duration information such as the times of the first call to a service center, when the crews were dispatched, and how long it took to complete the repairs. This information can be given to regulatory authorities to show response time and be used in performance-based asset management reports. Outage Distribution Interruption Reports are archived in the Oracle Data Warehouse.
Data Structure Enhancements
FAME was modified to include Smallworld objects for premise, service points and load groups, creating links to the Customer Information System (CIS). A Smallworld Connectivity object was created to capture and maintain the device network association and the total number of customers on each device within the network. Load Device and Control Device objects were created to facilitate the graphic display of the electric model affected by the outage incidence.
The proposed solution also required the implementation of an Oracle database with common data structures between the various components. This common data structure facilitates the sharing/maintenance of data between the ServiceOn web based application, FAME / Smallworld application, and BussinessObjects reporting tools.
Summary
UtiliCorp has created a comprehensive and cost efficient system by implementing basic building blocks, and has laid the foundation for a more robust outage management system. This work was managed without affecting the service centers and was done in such a way to minimize training the service center personnel. UtiliCorp is confident in the capabilities and flexibility of their system, and believes the simple components it contains could potentially be used at other utilities.
UtiliCorp United is an international energy company with its headquarters in Kansas City, Mo. It operates in the United States, Canada, United Kingdom, Spain, Germany, Norway, New Zealand and Australia. UtiliCorp has over 4 million customers in its domestic and international operations. For 1999 it was ranked the second-largest wholesale marketer of electricity in the U.S. and the third-largest marketer of natural gas. On the Fortune 500 list, UtiliCorp ranked 90th based on 1999 sales. The company is also included in Fortune's list of America's Most Admired Companies. UtiliCorp’s United States operations are concentrated in the Mid-West, largely in Colorado, Kansas, Nebraska, Missouri, Minnesota and Michigan.
Like many long established utilities, UtiliCorp’s computer systems had evolved in an unplanned manner, resulting in a variety of systems on different computers, ranging from PCs to mainframes. Moving and maintaining information between all these disparate systems was a chore. UtiliCorp wanted to have a common solution for its call centers, field resource centers (FRC), and other facilities. They also wanted to improve customer service by providing outage status, estimated restoration times, and identification of problems areas. The reduction of maintenance expenses was another important consideration. Whatever the solution, it was very important that current operations not be affected.
Existing processes and components
A major component in UtiliCorp’s existing architecture is the Facilities Management & Mapping Enabler (FAME). FAME is a GE Smallworld-based AM/FM software system that is used to inventory, model and help manage gas and electric networks. The FAME system accesses data within the Smallworld GIS that maps and inventories UtiliCorp’s gas and electrical facilities throughout its Mid-Western territory. Some of the projects that were being implemented within the FAME framework included; outage incidence recording, distribution interruption reporting, customer information system (CIS) interface, and SCADA data availability.
UtiliCorp operates two service centers staffed by UtiliCorp personnel that handle the normal volume of service/trouble calls. Trouble call tickets were entered into a Microsoft Outlook e-mail form and then sent, via a series of Microsoft Exchange servers, to all the appropriate UtiliCorp sites and personnel. UtiliCorp also contracts with 21 st Century Communications to provide High Volume Call Answering, (HVCA), services. 21 st Century Communications can process 20,000 calls per hour. The HVCA process is used by UtiliCorp to help in answering problem calls when the demand becomes greater than what UtiliCorp’s trouble call centers can handle. Typically this process is utilized whenever serious, widespread outages occur. In UtiliCorp’s Mid-Western territory the widespread outages are most often weather related and are caused by the serious thunderstorms, tornadoes and ice storms common to that region. The initial implementation of the HVCA system generated an ASCII file which was passed every five minutes from the vendor to UtiliCorp, where the ticket information was parsed, entered in the Exchange form and sent using Smallworld OLE processing.
Part of the existing trouble call reporting system was a PC based database system that was written using FoxPro. This was a single user system that required someone to manually enter outage data after an incident had occurred. This was a very slow, time-consuming effort and one that allowed the introduction of errors during the data entry process. Compounding this data entry problem, the FoxPro system did not have any data validation routines and there was no link between the FoxPro system and the facility model. Thus, there were problems with the timeliness of the data, the accuracy of the data, and the data existed in its own space without any links to other systems
UtiliCorp was also facilitating a link between the mainframe based Customer Information System, (CIS), and FAME. This link would allow access to customer information and usage via the graphic facility model.
Problems with Existing Processes and Components
There were many problems with the existing trouble call process. One problem was the slowness of the Exchange server. This was mainly the result of having to replicate large volumes of data across the network to all the Exchange servers. Another problem was the reliability of the Exchange network.
The FoxPro outage incidence recording and outage incidence reporting routines were being re-engineered as separate and independent projects, utilizing an Oracle database. This project was being implemented with no tie to the trouble call process. The same pieces of information were being captured and stored with different names and different data types throughout the unrelated systems. Using FAME to facilitate an outage management system required existing and valid linkages between the graphic data and the CIS data. Only 10% of the customer data were linked with graphic data, making it very difficult to associate trouble calls to electric devices or predict and close an outage. It was estimated that it would take three years to validate graphic data, verify electric network connectivity, and create valid linkages between FAME and CIS.
Proposed Solution
After examining the problems, UtiliCorp decided it needed an integrated solution that could be accessed from the call centers, field resource centers, and the main UtiliCorp offices. An initial investigation of the GE Smallworld PowerOn product generated several roadblocks. The PowerOn application was perceived to be very disruptive to existing processes, provided more functionality than currently needed, required more training and more time to implement, and the cost was not acceptable to the user community.
It was clear that a more compatible and cost efficient implementation was needed. To achieve this goal it was decided to use the power of the Internet to link these sites to a common database. A thin-client/thick server setup could provide a common solution to all users and help in providing better customer service, identification of problem areas, minimization of maintenance expenses, and making outage statuses and restoration times available. After a thorough investigation, it was determined that an Oracle database would be the central component in the new architecture. This web-based application was later named ServiceOn.
Proposed Processes & Components
Another component of the proposed solution was to link the trouble call database to the FAME GIS system. Using the outage information gathered by the services, the data could then be used to create real-time displays of outage s areas, help predict problem devices, help in dispatching field crews, and help relay service information back to the customer.
ServiceOn
ServiceOn is a web-based trouble call system that was developed by UtiliCorp and GE Smallworld personnel to help solve some of the problems with the existing trouble call system. Some advantages of UtiliCorp’s ServiceOn implementation are that the input screens have the same look and feel as the old trouble call data entry screens, and the process flow is the same. These features, combined with the browser interface, present the service centers users with a system that requires a minimal amount of training and documentation. Because of the thin-client architecture, ServiceOn is faster and more reliable than the old system.
One of the key features of ServiceOn is that it uses an Oracle database as its data storage server. Having the data in a server-based Oracle database means that the data is more accessible to a wide variety of UtiliCorp users. Also, the users now have a faster response time because of the capabilities of the Unix server that hosts the Oracle database.
On the ServiceOn data query panel, the user has the ability to select pre-defined queries of the database using one or more of the following items: city name, state abbreviation, zip code, address, phone number, ticket label, status, and the definition of the begin and end date/time. From this query panel the user can select fields with which to sort the report. The user is also given a number of report display types to choose.
The results of the ServiceOn query are displayed on a series of reports. The Query Results Panels provide users the ability to review the original trouble ticket and to update that information. There is a monitoring option on the Query Results Panel where the data can be refreshed every 30 seconds. This option allows the users to watch for new trouble tickets without any input. Another Query Results Panel option is to query and display the trouble tickets on their status of “First Call”, “Predicted”, “Confirmed”, “Dispatched”, “Resolved”, or “Closed”. Combined with the date/time option, this gives UtiliCorp the ability to only look at tickets that are currently been worked on or that have been recently completed.
One of the ServiceOn reports is the Field Resource Center (FRC) display. This report uses an existing FRC format, minimizing user training. The report allows the users to access individual trouble tickets and to change the status on them. Another report type is the Review by Feeder report, which displays information by substation and feeder data. On the Review by Ticket Label Report, the users can group individual trouble tickets by ticket label. On this report, the user can summarize trouble tickets with same label display count. They have the ability to state transition tickets by ticket label and can reassign tickets to new or existing ticket labels.
ServiceOn provides the user the ability to capture the outage incidence information via the Distribution Interruption Report form. On this form the user can “close” the outage form and enter in data pertaining to the customers affected, such as weather conditions or components that failed, and select the cause of the outage from a set list. Data entered on this form is used to create outage incidence reports that show when and where an outage occurred, what was the failing component, what customers were affected, the feeder involved, and the duration of the outage including the first call, dispatched, and completed times. These reports can be sent to regulatory authorities and used for other performance-based asset management reports.
ServiceOn procedures are not dependent on FAME for graphic validation, but enhanced by the graphic display. If a valid link to a graphic device is not found, FAME will create a default label based on work district and zip code and/or the user can assign a ticket label to associate related trouble tickets. If valid linkage is found, FAME will associate related data, assign an intelligent ticket label, and create graphic components to display the outage.
ServiceOn Data Loader
Part of the proposed solution was to devise a method by which data from the High Volume Call Answering (HVCA) was input into the ServiceOn database. This method was a custom built program called the ServiceOn Data Loader. The HVCA process collects data in a text file. This file is sent to a server on UtiliCorp’s network via an FTP process. As soon as one of these files is written to the server, it is read-in by the ServiceOn Data Loader. This program reads in the text file, translates the data from the HVCA format to the ServiceOn format, inserts the data into the ServiceOn Oracle database, and archives the original file. This whole process takes only seconds for a file that may contain a thousand records.
Fame Enhancements
The original intent to use the FAME Smallworld GIS system for facilitating outage management was re-engineered so that FAME provides a real-time view of active outages. This real-time background processing of each trouble ticket record identifies the graphic location of an outage. FAME outage prediction logic associates outage locations to the electric facility model and graphically displays the network affected by the outage. This FAME prediction logic also groups the individual trouble tickets with a common ticket label. FAME outage analysis assists in the management of outages by changing the graphic display relative to the outage life cycle status: First Call, Predicted, Confirmed, Dispatched, Resolved, or Closed.
Another advantage of easily locating network outages is improved customer service. Now customer can be told of the extent of the outage, the probable cause, and the estimated recovery time. The maps can also be used to identify problem feeders and/or components. This ability helps minimize maintenance expenses by initiating needed maintenance projects before an outage occurs.
Reports Enhancements
BusinessObjects is used to generate reports and statistics on outage incidences from the Oracle database. These reports show when and where the outage occurred, the failing components, the customers affected, and the feeders involved. The reports also contain duration information such as the times of the first call to a service center, when the crews were dispatched, and how long it took to complete the repairs. This information can be given to regulatory authorities to show response time and be used in performance-based asset management reports. Outage Distribution Interruption Reports are archived in the Oracle Data Warehouse.
Data Structure Enhancements
FAME was modified to include Smallworld objects for premise, service points and load groups, creating links to the Customer Information System (CIS). A Smallworld Connectivity object was created to capture and maintain the device network association and the total number of customers on each device within the network. Load Device and Control Device objects were created to facilitate the graphic display of the electric model affected by the outage incidence.
The proposed solution also required the implementation of an Oracle database with common data structures between the various components. This common data structure facilitates the sharing/maintenance of data between the ServiceOn web based application, FAME / Smallworld application, and BussinessObjects reporting tools.
Summary
UtiliCorp has created a comprehensive and cost efficient system by implementing basic building blocks, and has laid the foundation for a more robust outage management system. This work was managed without affecting the service centers and was done in such a way to minimize training the service center personnel. UtiliCorp is confident in the capabilities and flexibility of their system, and believes the simple components it contains could potentially be used at other utilities.