K-water is top level total water service company.
HOMETechnologyAI Water Treatment Plant
트위터 페이스북

AI Water Treatment Plant


K-water has developed an AI Water Purification Plant that operates autonomously using big data-based AI technologies, which is a significant change from traditional water treatment plants that were operated and managed by human analysis and judgment. The AI water purification plant is operated based on four core technologies: ‘autonomous operation, energy management, predictive facility maintenance, and intelligent video safety technologies.’ K-water completed a successful pilot project at the Hwaseong Water Purification Plant in April 2022, whose excellence was recognized and selected as a World Economic Forum Global Lighthouse as the first public organization in the water industry. We plan to expand the AI water purification system in 43 metropolitan water purification plants by the end of 2024 and to develop ISO international standards by leading the global water industry through continuous technological advancement.

Project Progresses

  • Complying with the management policy of “freezing tap water rates” by reducing costs, such as electricity, chemical, and maintenance costs, even in difficult situations such as rising electricity bills and chemical costs
  • Selected as a “Global Lighthouse*” by the World Economic Forum for the first time in the water industry and as a public organization for its global excellence in AI water purification and leadership in water technologies.
    * Pioneer enterprises introducing 4th Industrial Revolution technologies, such as AI and Big Data (153 globally, 5 domestically)

AI Autonomous Operation


An operating system that allows real-time decision-making for all water purification processes (8 units) by utilizing big data-based AI technology for water purification plant operation factors, such as water quantity and quality, etc.


  • ① On-site sensing

    • measure the quantity and quality of tap water with sensors installed in the facilities such as pumps and pipes from the water source to the consumers
  • ② Data collecting

    • transmit and collect data collected from the sensors to the SCADA system in the water purification plant in real-time
  • ③ Big data analysis

    • AI, which has learned using a massive amount of big data over the past three years, analyzes data in real-time, and creates optimal control values
  • ④ AI operation

    • AI operates and monitors equipment, such as the turning on or off of pumps


  • Production costs have been reduced through process optimization (more than 4% reduction in chemical costs, etc.)

Energy Management System (EMS)


A system that suggests optimal operation plans through real-time electricity power monitoring, data analysis, and facility management for high-energy consumption facilities


Experience-centered operation > minimum required pipe pressure + optimal pump combination > optimal energy consumption

  • (Pump control) optimal control of water pumps using minimum required pipe pressure
  • (Power peak) an alarm is generated when the target peak value exceeds the power demand prediction


Electricity costs have been reduced by more than 5% by optimum pump operation

Predictive Maintenance System (PMS)


The predictive maintenance system detects failures and suggests causes through real-time operation data analysis for major facilities


Sensor construction, Real-time monitoring, Detection of abnormal symptoms, Self-diagnosis, Alarm and cause analysis

  • (Self-diagnosis) detect failures and suggest potential causes by analyzing the vibration data of pump motors in real-time
    * Malfunction items: shaft misalignment, mass imbalance, impeller/motor/bearing defect, cavitation
  • (Detection of abnormal symptoms) early detection and alarm generation of abnormal symptoms through the analysis of facility operation data
    * Applied facilities: 3-phase motor (stirrer, etc.), distribution board, ozone generator, salt chlorine generator


Improvement of facility safety by detecting and diagnosing abnormal symptoms in real-time

AI CCTV for Safety Management


A safety management system that applies real-time AI video analysis and auto-rotation technology for 8 items in three major fields to the existing CCTV and detector sensors


  • Fire

  • Trespassing

  • wandering

  • Collapse

  • Alarm alert for 8 items in the three major fields, including facilities, security, and industrial safety
    * (Facility) fire, pipe leak (Security) trespassing, wandering (Safety) collapse, rescue signal, electric shock, and hard hat, etc.


Accident reduction and prompt action through accident recognition without time and space constraints.
* AI model accuracy of 96%, KTL performance test, and KISA performance certificate acquired