6-1 Abstract
In 2014, Taiwan Power Company (TPC) began using Intelligent Electronic Device (IED) operation information from Remote Terminal Units (RTU) and IEC 61850 substations as trigger signals for calculating the Sequence of Events (SOE) times for auxiliary switches from circuit breakers. An online circuit breaker inquiry system was then established, and more than 830,000 pieces of information have been accumulated so far. In 2018, AI big data analysis was introduced using equipment brand, age of use and past operation timing. This led to the successful diagnosis of 410 abnormal circuit breakers, particularly those with dull or delayed mechanisms. The prediction accuracy rate reached 30%.
In 2021, the median number of the close/open operation timing was used as the diagnostic standard and the construction of IEC61850 standard substations was promoted. This will effectively improve accuracy by 40% and result in the successful diagnosis of 70% more abnormal circuit breakers.
To allow for ongoing status monitoring, a web-based management platform and an notification app were developed (as shown in Figure 1). When a circuit breaker experiences an abnormal operation, the system will immediately notify maintenance personnel so that the issue can be resolved quickly. This greatly improves the security of the power grid.
6-2 Big Data Statistical Analysis of Online Circuit Breaker Operation Timing
The operating times for circuit breaker that was obtained through the RTU/IED triggered SOE information technology (see above) varies depending on the sampling circuit and the auxiliary switch of the operating link. According to analysis, performed using the AI big data XGBOOST model, under normal circumstances the close/open operation timing is less than 5ms regardless of whether the circuit breaker is closed or open. Consequently, the median number of the close/open operation timing was introduced in 2021 as the diagnostic standard, allowing for greatly shortened AI big data analysis times.
6-3 Construction management platform and practical application results
The large database of circuit breaker operation times was combined with data management functions to build a management platform. This platform allows for the reporting of abnormalities in real-time and issues notifications via app. This quickly alerts maintenance personnel and appraises them of the situation so that maintenance can quickly be arranged in accordance with the problematic equipment’s operational trends, brand, type and use history (see Figure 2). At the time of writing, 223 circuit breakers with institutional exception have been diagnosed and processed.
Figure 1 – The online circuit breaker operation timing management platform and action notification app
Figure 2 – Online circuit breaker operation timing exception handling results