To address customer satisfaction and supply reliability, utilities now value the importance of collecting and correctly classifying the causes of every system fault outage. Reliability worth and continuity index assessment are important factors in distribution system planning and operation.

Assessing system reliability is a two-step process that involves measuring past performance and predicting future performance. This is an exercise that requires complex data collection from the event site followed by detailed analysis. The structure of this data-collection system is entirely related to the technique used to collect and validate the forced-outage events, because the data is used to determine the utility's reliability index, and to establish standards for operational and maintenance planning.

To enhance this process, we chose to interact with the utility's maintenance and field staff to develop a reliable computational system. If staff input is used to develop ways to improve the data-collection system and information analysis, they will be more likely to support the final system implementation.

In the course of its research and development program, distribution utility Rio Grande Energia S.A. (RGE; Porto Alegre, Rio Grande do Sul, Brazil) enlisted the help and support of the Electrical Energy System Group (GSEE) of Pontifical Catholic University of Rio Grande do Sul (PUCRS) to establish a fault-outage-management program that uses mobile computer technology. RGE is now implementing this system.

COLLECTION SYSTEM

The main objective for establishing a fault-outage-management system is to provide the distribution utility with forced-outage-causes inspection information on the performance of the network components to ensure that the financial resources available to improve system reliability are invested for maximum benefit. This could require increased resources on data collection that, given an automated detection process, should be equipped to provide accurate information on the performance of system components. The data collected forms the comprehensive database to determine:

• Network sections with the worst performance
• Equipment quality used on the distribution system
• Outages attributable to switching errors
• Equipment performance on a regional basis.

An overview of the computational system used to collect data is a three-stage process. First, the field electricians are trained to use handheld software to collect the system-forced-outage information at the event site. This application is a formal questionnaire implemented as personal digital assistant (PDA) software. Second, the desktop PC software uses the collected information to identify forced-outage causes. The final stage uses OpenGL technology to process the historical database of event information to identify the most critical areas in the overhead network, providing reports and statistical information to enhance resource allocation and maintenance policies.

DATA COLLECTION

At the site of the fault, the field electrician answers the questionnaire while working on the fault repair. This provides information on network conditions and surrounding variables. This information is sent to a historical database. The desktop PC analysis application uses this information — along with information from other databases — to compose an event diagnostic, providing a possible cause, details of all previous maintenance work in the locality and the consequences of the forced-outage event (for example, the number of consumers without supply).

Although the data collection at the fault location is considered important because of its content and accuracy, the line worker's main function is to repair the fault damage and restore supply. For these reasons, the application for collecting information must be quick and easy to use.

The questionnaire software interface used by RGE contains several features that simplify the process such as tree views, combo boxes and check boxes. RGE utility engineers developed the questionnaire and all of the questions were created based on their experiences. The questions cover almost every possible equipment condition and surrounding environment variables necessary to identify the cause of an outage.

The data-collection software has been developed in several steps, two of which have special features. The Java implementation, one of the first versions, has the advantage of portability because it runs on structures called virtual machines. Java virtual machines are commonly implemented on most of the recently released hardware, from PCs to PDAs to mobile phones. After that, for benchmarking reasons, another implementation was introduced that uses complicated code in C++ language. This version is platform dependant (in RGE's case, dependant on Palm OS PDAs), but has the benefit of running nearly 30 times faster than Sun Java's mobile information device profile (MIDP).

Following the collection of information on the environment such as weather conditions, presence of humans, vegetation conditions and so forth, the field electrician receives a list of components that contains the following elements:

• Load-break switch
• Public lighting
• Open fuse cutout
• Electric meter
• Load interrupter switch
• Pole cable termination
• Cable
• Lightning arrester
• Connection
• Pole
• Fuse link
• Voltage regulator
• Connection
• Recloser
• Guy
• Automatic line sectionalizer
• Metal structure
• Distribution transformer.

The field electrician must indicate the defective components and their conditions. For each component listed, there is a sublist of the most common component conditions for which the field electrician must choose the most appropriate condition.

ANALYSIS

The analysis application consists of a graphical system for distribution network visualization, constructed under the methodology of object oriented programming (OOP) using the C++ as programming language, Microsoft Windows as the platform and the Borland C++ Builder as the integrated development environment (IDE). The application uses the OpenGL API technology to draw the distribution network graphical representation on the screen. OpenGL API comprises a library that provides an interface for developers to ease the implementation of graphical software.

The graphical information and reports are necessary tools for operation and maintenance planning. The developed application provides an easy way to obtain such information. The user can select a feeder area on the screen or select a feeder from a tree view, and the application can retrieve the forced-outage historical information. The application uses information from the master database, the maintenance historical database and the failure historical database to compose the reports and the graphical form.

BENEFITS

The first key benefit of this system to emerge was the geographical presentation of fault outages attributable to vegetation factors. This information has been used effectively to significantly improve the reliability of RGE's overhead distribution network. This computer-aided process replaces a paper-based system that provides more reliable information on fault outages. To date, RGE has not reduced the number of field staff.

Mobile computing technology has improved the efficiency and accuracy of many processes in electric utilities. The mobile technology offers a wide range of new possibilities to enhance the utility's internal processes, making the information flow both easy and reliable. In this application, the technology has been used as the foundation for building a fault-outage-management system to improve RGE's system reliability. To date, there are 20 PDAs collecting data from fault-outage events in the RGE area. Results indicate that the utility's field electricians have completely adapted to the new technology, and that the quality of information has improved. For this reason, RGE is keen to extend the project to cover all areas of the utility.

Introduced following a research and development project conducted by the GSEE of PUCRS, this improved form of data acquisition has enabled the utility to identify which sections of its network have problems to achieve system reliability performance. The fault-outage-management system provides RGE's operational managerial staff with the information required to make informed decisions on the allocation of resources and capital investments.

ACKNOWLEDGMENT

This system was developed by the joint efforts of the Electrical Energy System Group of Pontifical Catholic University of Rio Grande do Sul and Rio Grande Energia S.A. (RGE) under the ANEEL research and development project. The authors wish to thank all the project members in RGE for their work and best efforts to complete the pilot project, especially engineer Rodrigo Bertani for his support and helpful discussions.

---

Carlos Oliva Pretto earned a degree in control and automation engineering in 2002 and a master's degree in 2005 from Pontifical Catholic University (Porto Alegre, Brazil). In 2001, he worked as a research assistant at the Fraunhofer Institute for Manufacturing Engineering and Automation at IPA in Stuttgart, Germany. His main areas of expertise include OOP, Java, mobile computing and robotics. cpretto@ee.pucrs.br

Flávio Antonio Becon Lemos earned his BSEE degree from the Universidade Federal de Santa Maria in Brazil in 1988, and his master's and Ph.D. degrees in 1994 and 2000, respectively, from the Universidade Federal de Santa Catarina. He worked for the state electricity utility in Porto Alegre, Brazil, from 1988 to 1992. In 1996, he was a research fellow at Brunel University (U.K.). Lemos is head of the Electrical Energy System Group and his research interests include power-system operation and stability, voltage control and distribution systems. flavio.lemos@pucrs.br

Mauro Augusto da Rosa earned his BSEE and MSEE degrees from Pontifical Catholic University in 1998 and 2003, respectively. He is currently a Ph.D. student at INESC (Porto at FEUP) Universidade do Porto in Porto, Portugal. His areas of expertise include planning and operation of distribution systems. marosa@ee.pucrs.br