The nation inaugurated the Carbon Neutrality Committee under the direct control of the president.
The government announced national determined contribution (NDC) goals of reducing greenhouse gas emissions by more than 40 percent.
In particular, the power generation sector’s NDC goal stood at 44.4 percent higher than others. Korea Electric Power Corp. (KEPCO) and other power companies have now entered an era of energy transition by announcing plans to phase out coal-fired power generation by 2050.
One of the technologies, which lead industrial trends, combining trends of the times and the 4th Industrial Revolution, is artificial intelligence (AI).
KEPCO KDN’s AI-Vision has been utilized most in the power sector.
Since 2016 when the so-called “AlphaGo Shock” realized the potential for deep learning. AI has a greater impact on all sectors of society.
Ahead of this, AlexNet, using CNN, a deep learning algorism, improved the conventional image recognition limitations.
ResNet, a deep learning algorism more power than human beings, emerged in 2015. KEPCO KDN, a representative institution of the power ICT sector, is implementing a paradigm shift with an “eco-friendly, digital-oriented, ICT platform.”
AI-enabled technologies are expected to be introduced fast. KEPCO KDN is nimbly coping with AI by inaugurating an R&D body specializing in AI. KEPCO KDN established a related R&D roadmap in 2017.
Here is a look into a field in which AI technologies are employed to further advance the development of the power industry.
Jung Joon-hong, general manager with KEPCO KDN Research Institute at KEPCO KDN headquarters in Naju, introduced KEPCO KDN’s leap as an energy ICT platform company.
“AI Vision,” the most used in the power utilization sector, counts numbers in real time as a worker moves.
An LED number counter bulletin board installed on the façade of a power unit plays a pivotal role in protecting crewmembers in the case of an emergency.
KEPCO KDN’s AI-Vision, counting the number of crewmembers remaining in a special zone, employs Auto People Counting (APC) technology.
The AI-Vision technology also ascertains the safety of crewmembers and determines whether electricity cables are broken or bended in real time.
KEPCO KDN began to develop AI technologies in 2017. The institution is now in an initial stage of building a whole power grid process.
The number of studies on AI technologies that have been conducted for the past five years sands at 26. Among them, 17 studies on AI were conducted last year, realizing the popularity of the technology. KEPCO KDN’s AI-Vision has been utilized most in the power sector.
Jung said, “The reason for KEPCO KDN’s more active research on AI is that AI stands in the text of an extension of ICT.”
Human beings can watch TV monitors and concentrate on recognizing changes for only 20 minutes, and a focus is on the AI-Vision, which automatically analyzes video information taken via cameras, CCTVs and drones using the AI technology, he said.
KEPCO KDN’s AI-Vision is used in a variety of areas, such as the diagnosing of the abnormality of optical ground wires of power transmission towers.
KEPCO KDN’s photovoltaic power generation amount prediction system is based on AI. The institution employs AI technologies in predicting photovoltaic power amounts and evaluating power generation unit prices.
Fuels for generating electricity account for 80 percent of electricity production costs, and importing bituminous coal costs a few trillion won.
The power generation unit price prediction system, using an AI technology, helps fuel procurement officials of power companies decide on optimal timing of purchasing by forecasting bituminous coal and LNG prices, Jung said. The system is in place at Korea South-East Power’s Smart Fuel Center.
AI technologies have been advanced, but there are hurdles which have to be addressed. Jung said the core AI technologies include self-supervised machine learning, edge computing and data augmentation technology.
_한전KDN 사장 김장현.png)
_한전KDN AI Vision.png)
