A competitive national electricity market increases the incentive to reduce maintenance costs and outages on the transmission network. There is a particular need to reduce the frequency and duration of forced outages caused by equipment defects. Effective condition monitoring provides a powerful tool to achieve these objectives.
While the practice of condition monitoring has been around for a long time, online condition monitoring is still a comparatively new phenomenon in the power industry. The emergence of new technologies, coupled with a paradigm shift in asset management from time-based to condition-based maintenance, has led to condition-monitoring devices becoming commonplace in many modern substations.
As power utilities head down this path, high-voltage (HV) equipment manufacturers are jumping on the condition-monitoring bandwagon by offering monitoring devices fitted to their equipment. In fact, a plethora of condition-monitoring devices are available that measure a range of different quantities on several types of equipment. These devices can contribute to condition-assessment and asset-management strategies, but utilities often end up installing multiple monitoring devices in several substations without a system to derive the expected benefits. This is because utilities give insufficient attention to the effective management of the information provided by these devices.
TransGrid is the major transmission utility in the Australian National Electricity Market with a dispersed network of substations well suited to condition monitoring. The company has developed an integrated real-time substation monitoring system, which manages information from a range of condition-monitoring devices across multiple substations. The system collects and stores data from sensors and brings all the data into one corporate information system.
TransGrid set out to develop an extensive condition-monitoring system in response to the asset-management drivers of reliability, availability and cost. The utility established an initial trial at the new Dumaresq 330-kV Switching Station — part of the Queensland New South Wales Interconnection project. TransGrid subsequently developed a broader philosophy for real-time condition monitoring and implemented a real-time monitoring system that incorporates several substations across the HV network.
In developing the monitoring system, TransGrid considered the following principles:
Integration — combining information from multiple devices across a number of substations into a common frontend
Accessibility — available to a range of users in a variety of locations, without the need for special software
Flexibility — in terms of queries that can be performed and the format and detail of information returned
Security — not allowing changes to data or control of substation functionality
Reliability — delivering expected features when required
Good performance — even with minimal infrastructure and regardless of the size of the system
Expandability — to potentially encompass an entire HV network.
TransGrid also considered current trends including the extension of high-bandwidth data networks to encompass substations and the use of the World Wide Web and Internet technologies for access to and sharing of information. The aim was to develop a condition-monitoring system that is in line with these trends, as well as emerging standards and best practice.
As with any information system, the design of a condition-monitoring system is critical to its success. TransGrid designed a system that could encompass an entire HV network but with initial implementation at a small number of sites, rather than designing a system for only one site and expanding later.
The utility decided not to combine condition monitoring into the SCADA system, recognizing a distinction between operational data (transmitted via the SCADA system) and maintenance data. However, alarms from various monitoring devices transmit to the operators via SCADA.
TransGrid decided to use off-the-shelf database software for the databases in each substation. It chose software that could interface with each site's control system and could be queried using standard methods. However, most off-the-shelf software was found to have the following limitations:
Much software of this nature currently available has been designed with process control and automation in mind, which are generally centralized rather than distributed systems.
Most off-the-shelf software relies on proprietary client software with, at best, a limited link to standard applications such as a Web-based system.
Most software does not cater for complex data structures such as fault waveform traces or circuit breaker travel curves, and is not sufficiently flexible in the way data is displayed.
For these reasons, TransGrid decided to develop the main frontend in-house using standard database and Web server software.
The System Architecture
The figure above shows the basic architecture of an online condition-monitoring system. Devices on HV equipment provide event-driven inputs to the data acquisition system (which may be the substation's control system). The acquisition system writes data into the data store at regular intervals; the data can be accessed at a later date as history or trend data.
The user-interface server responds to user requests for information and queries the control system (for real-time data) or data store (for history data) to return to the user. This may be a Web server on the company's intranet responding to a request from a user through its Web browser.
The TransGrid System
Substations on the TransGrid network are well dispersed geographically and vary substantially in size and design. Some sites have high-bandwidth fiber optics or microwave communications to the wider corporate data network, while others have basic communications due to the remoteness of their location and lack of infrastructure in their geographical area.
The condition-monitoring system is currently monitoring seven substations, each with their own database, using a common Web server. All substations but one have high-bandwidth communications; Balranald Substation connects via secure dial-up. All sites have a main database that stores condition-monitoring data, but these databases differ depending on the control system at each site. At some sites there are condition-monitoring devices that have their own databases. These databases are contacted by the centralized system independently of the main databases at these sites.
TransGrid selected this architecture because it best suits its needs. The company chose the Web-based interface as a standard interface that does not require proprietary software and can be accessed from many locations including offices, substations and even home via dial-up for on-call staff. The architecture was designed to minimize traffic across data networks, thereby enabling the connection to sites with low external bandwidth while retaining good performance.
Using the System
To use the system, a user first selects the data he or she would like to view using a grid map and operating diagrams. Alternatively, a user can select the equipment from lists, which offer more powerful and flexible options, including the ability to select data from multiple substations and specify multiple date ranges for a parameter. An example would be the comparison of one week's data with a previous week or data from another substation.
When viewing the data, a user can view or download the raw data and import it into third-party applications, or send a graph or raw data via e-mail.
Even in its early stages, the system has provided a range of information on a variety of equipment, serving several purposes. To date, TransGrid has used online monitoring:
To detect and prevent failure of a current transformer
To identify defective wiring on a circuit breaker control system
To detect faulty transformer monitoring devices (for temperature and oil level)
To monitor tapchanger behavior
To monitor occurrence and nature of AC auxiliary supply disturbances.
These early results are encouraging and the range of uses found for the data has exceeded initial expectations. Furthermore, TransGird has gained sufficient confidence in the system to continue its implementation, with further learning to continue as additional devices are brought onto the market and additional incidents occur.
Changes to Maintenance Practices
An effective condition-monitoring system can reduce or eliminate time-based maintenance and the need for routine inspection. Genuine benefits are only achievable, however, if the substation has a comprehensive monitoring scheme covering all devices that require inspection.
Condition monitoring provides an opportunity for early detection of the defect. It also can improve the response to the defect. Usually, a substation alarm requires a highly experienced person to determine if an immediate response is necessary and ensure the appropriate personnel are sent to the site to rectify the defect. Often these decisions are based on little information until there is a someone at the site.
A more difficult or obscure problem could be e-mailed to an expert for diagnosis. For example, an unusual travel curve on a circuit breaker could be e-mailed to the manufacturer for diagnosis. It is possible that the manufacturer on the other side of the world could analyze the data before any person has arrived at the site.
The benefits of reduced maintenance and inspections, and improved response to defects are enhanced when the substation is in a remote location.
Maximizing the Benefits
Online monitoring provides the information necessary to change routine maintenance and inspection practices, reducing outage requirements resulting in direct savings in maintenance costs and improved availability. Continuous monitoring can alert on unusual equipment condition, degradation or impending failure in its early stages, minimizing major defects and forced outages. Online monitoring also provides real-time data for decision-making on occasions such as short-time transformer overload conditions.
Cost advantages will only be realized over a period of time because the cost of online monitoring is generally greater than short-term reductions in maintenance costs. However improved reliability, improved availability and overload monitoring provide substantial cost benefits. These benefits are of particular advantage at remote sites where travelling times and costs become significant in routine maintenance or defect response.
The value of condition monitoring for maintenance practices and diagnostics are most fully realized when such monitoring is comprehensive and considered strategically from a risk-management perspective.
TransGrid has developed a strategy for incorporating online condition monitoring into new installations and retrofitting monitoring systems at older sites. The extent of monitoring installed at a particular site depends on the following factors:
Remoteness of the site
Criticality of the equipment at site for network stability and reliability
Age, known condition and existing replacement strategy of equipment
Life-cycle likelihood of equipment failure, taking into account past experience such as failures and defects on equipment of similar types
Cost of condition-monitoring devices compared to the cost of the asset they are monitoring
Reliability of the condition-monitoring devices themselves.
As a minimum requirement, complete systems are installed at all new remote substations and important plant items are targeted at other sites. TransGrid's initial strategy for retrofitting is to target equipment with high consequential system or financial impacts of failure.
Establishing a condition-monitoring system that is integrated into corporate systems and asset-management systems will enable condition monitoring to contribute to the reliability of electricity networks and business performance. TransGrid is now considering integrating this system with other maintenance systems and simulation tools. Automatic remote data analysis and comparison between sites will be possible using artificial intelligence techniques.
The presence of an online monitoring system does not replace the careful thought given to which parameters to monitor, nor does it replace proper analysis and interpretation of the data. However, it makes the acquisition and manipulation of data substantially easier.
Steven Jones is manager design at TransGrid. He has the master's degrees in engineering and business administration. He manages the group responsible for the design of transmission lines, cable systems, telecommunications and substations including protection and control. Jones has more than 20 years experience in the power industries in Australia, Switzerland and Vietnam. He has authored many papers on substation design, insulation testing and asset management, and he is a member of the CIGRE international Study Committee on Switching Devices.
Andrew Kingsmill has worked at TransGrid for five years in both engineering design and asset-management sections. In this time, he completed a thesis on online condition monitoring and has had significant involvement in the implementation of online monitoring devices, systems and strategies within TransGrid. He has authored several technical papers and presentations on this topic, and maintains continuing involvement in deployment of new technologies within TransGrid.