Weeks might pass before a transformer can be disassembled safely to determine the reason why it exploded and incinerated. I am told this situation is not uncommon when extra-high-voltage (EHV) transformers fail. Nobody wants to be near one of these (or a breaker) when it fails, and replacing one is costly and time consuming.
To protect our employees, contractors, the public and assets, American Electric Power (AEP) needed a tool to monitor the health of vital equipment, analyze data and signal alarms whenever abnormal conditions were observed. AEP needed a better way to predict equipment failures so we could intervene beforehand.
AEP operates a 40,000-mile (64,374-km) transmission system, the largest in the U.S. Much of the equipment was installed more than 50 years ago and some equipment even 80 to 100 years ago. Some of this equipment needed to be upgraded. A tool for monitoring equipment health enabled us to prioritize asset replacement based on condition and performance, not simply age. The solution is Asset Health Center (AHC), which provides actionable information based on real-time monitoring and trending.
AEP had instrumented T&D station equipment before, but those attempts were unsuccessful. Many earlier monitoring devices required frequent and costly maintenance. AEP also saw high failure rates because the devices could not stand up to the harsh conditions at many stations across the utility’s sprawling service territory.
AEP required robust, low-maintenance monitoring devices that were simple to operate as well as easy to set up and interrogate. Monitors had to integrate with other devices and our data collection systems. They could not be, as in the past, stand-alone systems or proprietary equipment that did not work well with other systems and equipment. In many previous cases, data was held locally and had to be gathered manually. One of my colleagues referred to the unwieldy data files as the “data dumpster.”
AEP sought a better way to analyze equipment performance and help engineers identify areas requiring attention. We needed a system that enabled us to organize, query and analyze the data collected. We started our AHC in 2012 with goals to prevent equipment failures, optimize maintenance effectiveness and support asset renewal decisions. AEP’s approach focused on sending critical actionable alarms from the AHC to the transmission operations center as well as targeted notifications to engineers alerting them of anomalies.
During development, a big question we had to answer was, what exactly should we monitor? We concluded that we should monitor the things that mattered most, such as the parameters correlating to the root causes of trouble and the parameters that can be corrected through maintenance. Leveraging existing data collected through our supervisory control and data acquisition (SCADA) network — for example, transformer temperatures and loads — would be essential.
Vendors developed new on-line sensing equipment to meet our requirements. Engineers tailored monitoring schemes according to equipment type and voltage class. For transformers, AEP chose to monitor dissolved gasses (in near real time), bushing health, moisture levels, oil and core temperatures, pump and fan operation, and load tap-changer positions, among other things. Breaker monitors record the timing of the mechanism, pump starts, pressures and operating coil current draw signatures.
Simultaneously, AEP developed a governance policy for data, which included identification of the master data source, data aggregation and communications protocols, data storage and organization, plus quality requirements, checks and standards. At an ideal time in our journey, a new and more sophisticated generation of monitoring technology had evolved that was more robust and interoperable than its antecedents. Another contributor to our success was our fiber-optic network, which was expanding and made it easier to deliver large amounts of data to a central repository.
Next, we had to turn data into actionable information for system operators, field employees and asset managers. This led AEP to a partnership with ABB. In 2013, AEP began developing an analytical engine, initially focused on transformer performance, that later became our AHC software.
The algorithms in the engine are based on engineering principles, industry experience and statistical models. The tool uses real-time information from the SCADA network, on-line monitoring data, information from equipment manuals, inspection reports and prior maintenance activities to predict and prioritize equipment-specific risk of failure. The algorithms also detect changes in operating characteristics by equipment type and sometimes by manufacturer. For example, a normal operating signature for one transformer could be a precursor for catastrophe in another manufacturer’s transformer.
AHC in Action
Enterprisewide deployment of the technology started in 2016, with monitors initially fitted on EHV transformers (new and retrofitted). By April 2017, AEP had installed monitors on more than 225 EHV transformers, and today our AHC monitors perform around the clock. Benefits were immediate as our AHC detected trouble, and we prevented the failure of three EHV transformers, for an estimated savings of US$15 million to $20 million. Two were repaired in short order and returned to service. Had the AHC not been monitoring them, they would have failed, likely irreparably.
AEP’s success with EHV transformers has resulted in the expansion of AEP’s program to 138-kV transformers and led to pilot programs for circuit breakers, batteries and underground cables. We are following the same process for the expanded categories of equipment: collecting the right data, developing algorithms to predict and prioritize risk, and providing actionable information to system operators and the field organization. These pilot projects look promising, and AEP is already diagnosing problems and calling for maintenance actions before equipment failure or inoperability.
AEP’s long-range plan is to use the real-time system to help with testing obligations for North American Electric Reliability Corporation compliance, starting with station batteries. Transmission planning and asset management engineers also have started using AHC information in the planning process. As designed, AHC information informs our asset replacement prioritization process.
As we use the AHC, we continue to refine our predictive capabilities based on data and experience. Because dielectric strength is critical in EHV transformers, for example, AEP decided to monitor partial discharge. Our engineers saw a partial-discharge signature in a review of the AHC monitoring data. This discovery led to a new ultra-high-frequency partial-discharge alarming scheme.
The AHC is designed to improve safety. For example, before approaching an EHV transformer for maintenance, field personnel can contact an AHC team member for real-time, transformer-specific data. If conditions at the transformer change and become unsafe, field employees are alerted. Many transformers have sirens and strobe lights that activate when there is a partial-discharge alarm. Employees evacuate the area when an alarm sounds.
Engineers and field personnel use the AHC web-based dashboards to pinpoint trouble. The AHC automatically issues alarms to control room operators when high-risk conditions needing immediate action are detected. It also e-mails engineers about areas of concern that do not require immediate action.
The AEP transmission AHC has been a tremendous success. It is preventing equipment failures, optimizing maintenance and supporting asset renewal decisions. We have realized savings that offset the costs of the project. More importantly, we have provided a more reliable energy grid for customers and a safer work environment for employees. ♦