How KEPCO is enhancing grid security with AI
When KEPCO, South Korea’s largest utility, integrated Alchera’s anomaly detection software with its existing CCTV network, the company raised the bar for facility protection and fire detection.
How does a world-class power utility use technology to enhance grid security in an era of regulatory and technical change? That’s the challenge KEPCO (Korea Electric Power Corporation) faced as it rolled out its Smart Grid program.
In the late 2000s, KEPCO launched Smart Grid, an ambitious, comprehensive, multi-year program designed to transform the country’s power system. Starting with a demonstration project on Jeju Island, the utility has field-tested numerous community or building-based initiatives. These include electric vehicle charging stations, smart meters, and zero-energy buildings.
System-wide, Smart Grid aims to optimize supply and demand management and improve grid stability by avoiding outages caused by natural disasters with IT solutions. In addition to supply-demand priorities, national guidelines require the utility to add renewables to the energy mix and achieve a 37% reduction target (compared to the business as usual case) in greenhouse gas emissions (GHG) by 2030. In effect, KEPCO is transforming the way it generates, distributes, and markets power.
Grid security was one priority on a long shopping list of things to do. Given the ambitious scope of the Smart Grid program and the projected cost ($7 billion US by 2030), finding the right partner was vital.
For KEPCO, that partner was Alchera.
Alchera enhanced facility protection and fire detection with state of the art visual anomaly detection technology (VADT). Seamlessly integrating Alchera’s VADT algorithms into the existing CCTV network, KEPCO was able to deploy a cost-effective, unmanned monitoring and detection system delivering real-time data and analysis.
Inside substations, for instance, the software monitors digital and analog gauges. In the field, edge-based detection capabilities were programmed to identify a suite of transmission tower and power line events, like smoke, fire, intrusions (human or mechanical), and galloping lines.
Alchera’s VADT system, in basic terms, in volves a three-step process. First, real-time video images are converted into a stream of digital data. Second, the data stream is continuously assessed against a baseline data set, which defines normal parameters. Third, alert protocols are automatically issued when the software detects and confirms a real-time event deviating from normal parameters.
Detection is important, but accuracy gives a fuller measure of anomaly detection software efficacy. Unusual behaviour falling outside normal parameters is not always a threat. Events that are rare and non-threatening are possible and likely. Treating every anomalous event as a threat can generate false positives. The key to efficacy is an algorithm that sees, classifies, and learns.
Powered by Alchera, KEPCO rolled out a VADT pilot project in 2018. In Phase 1, Alchera software was connected to 240 cameras. Based on strong Phase 1 results, KEPCO launched Phase 2 in 2020, which is expected to increase the total number of cameras to 700.
Though specific details have yet to be released, the benefits are human and financial. VADT lowers the risk of injuries because staff avoid or reduce time spent on dangerous tasks, such as line inspections. Worker safety is important for KEPCO because recruiting next-generation line workers is challenging. The software also generates positive financial impacts by lowering ongoing maintenance, inspection, and insurance costs.
The Smart Grid revolution might be the biggest technological change since homes were wired for power. Pushing intelligent solutions just beyond power generation, distribution, and markets, KEPCO is modernizing the way utility operators think about grid security.
“You can’t protect what you can’t see.”
– Young Hwang, Alchera Founder & CTO
You can’t protect what you can’t see.
Get peace of mind with AIIR visual anomaly detection today.