Goyang’s Smart AI Energy Watch

Goyang Special City in South Korea developed and piloted an AI-based monitoring and conservation service focused on appliance-level energy tracking. Using Non-Intrusive Load Monitoring (NILM) technology, the system disaggregates household electricity use by device, allowing residents and facility managers to see real-time consumption, forecast bills, and adopt targeted energy-saving actions. The project reduces waste, lowers costs, and promotes voluntary conservation among citizens, while laying the groundwork for wider city-scale energy intelligence.

Goyang proceeded through a phased execution:

1. Baseline audit & instrumentation
   • Audit of existing households to select target facilities
   • Installation (or use) of energy meters, sub-meters, IoT sensors (temperature, occupancy, CO₂, lighting)
   • Integration with existing building management systems (BMS) or SCADA platforms
2. Data ingestion & platform setup
   • Central platform built to ingest streaming data from sensors, meters, weather/ambient sources
   • Data cleaning, normalization, timestamp alignment
3. AI / analytics module development
   • Algorithms for anomaly detection, pattern recognition, load forecasting
   • Rule engines and decision logic for triggering control actions
   • Feedback loops for learning (i.e. comparing predicted vs actual outcomes and adjusting)
4. Control interface & actuation
   • Implement control commands to HVAC, lighting, ventilation, etc.
   • Allow both automatic and manual override modes
   • Safety / fallback rules to prevent discomfort or system conflicts
5. Monitoring, calibration & tuning
   • Continuous monitoring of system performance
   • Calibration of models and control heuristics as more data is collected
   • Iterative refinement and responsiveness to special events
6. Operationalization & training
   • Training facility managers and operators on dashboards, alerts, and intervention
   • Defining standard operating procedures for alerts, overrides, and maintenance
7. Scaling & roll-out plan
   • Based on pilot performance, plan to scale system to additional housholds, zones, or across the city

This phased implementation ensures that the system evolves gradually, minimizes risk, and incorporates learning.

Here’s how Goyang structured its collaborations:

• Technology provider(s): AI / analytics software vendor, IoT hardware supplier
• Local utilities / energy agency: for energy data access, alignment with grid operations
• Academic / research institutions: for algorithm development, validation, performance evaluation
• Facility management companies: for integration and operational support
• Funding agencies / central government: for grants or budget co-financing

These collaborations allowed combining domain expertise, access to data, capital, and operational capacity.

Public information emphasized household benefits, lower bills and real-time appliance control, while also highlighting the city’s climate commitments. To ensure stakeholder buy-in and broader awareness, the city focuses on:

• Launch events / press briefings showcasing pilot buildings and results
• Public dashboards or portals (if feasible) showing energy savings and CO₂ reductions
• Workshops or open days at monitored buildings to engage citizens, students, staff
• Reports, infographics and media outreach to local newspapers, municipal websites
• Internal communications to municipal departments to promote adoption and feedback

The project financing model and resource mobilization was critical to its feasibility.

• Capital investment (sensors, ICT, integration, platform)
• Operational costs for software, maintenance, upgrades
• In-kind contributions (e.g. existing metering infrastructure, facility staff time)
• Subsidies / grants / central government funding or energy agency support
• Cost recovery via energy savings or avoided utility bills
• Resource allocation for training, monitoring, staffing

The pilot proved highly effective. Across households, energy consumption dropped by up to 25%, with electricity bills reduced by more than 12 % in just two months of the 2024 demonstration.

Users especially valued the ability to see real-time, appliance-specific data, helping them target which devices to switch off or use more efficiently. Satisfaction surveys confirmed strong acceptance of the technology, with more than nine in ten households reporting a positive experience.

In sum, the system proved both technically viable and strategically valuable in advancing Goyang’s energy use and costs savings.

The project faced different obstacles and lessons learned during implementation:

• Data quality and integration issues: legacy systems, inconsistent formats, missing data
• Resistance from facility managers or staff unused to automated control
• Initial model inaccuracies requiring time and calibration
• Operational constraints / occupant comfort tradeoffs when automation conflicts with human preferences
• Governance / organizational silos hindering cross-department collaboration
• Maintenance and reliability of sensors and communication networks (downtime, calibration drift)
• Scalability pressures: models trained on pilot may not generalize to all building types

These challenges underline the need for robust change management, calibration, and stakeholder alignment.

Goyang’s AI energy monitoring and control framework offers a robust path for other cities to follow: a modular, data-driven architecture that can adapt to different sensor ecosystems and building types, plus a phased pilot-to-scale governance model aligned with municipal workflows.

The key to replication is ensuring baseline metering/control infrastructure, strong city leadership support, partnership with capable tech providers, and change management for operations teams.

While adaptation will be required in each context, the Goyang model demonstrates that even in an existing urban environment, AI can deliver meaningful energy savings, operational insight, and support for climate commitments.

The experience underscores that sustainability, digital transformation, and fiscal prudence can converge, if designed thoughtfully, and that other cities can replicate this journey to accelerate energy efficiency in public assets.

• Goyang City invests 402 billion won in smart cities.
• https://urbantransitionsmission.org/advancing-urban-transitions-in-east-asia-highlights-from-the-regional-conference-in-goyang-south-korea/

DongSoo Ahn
Director of SmartCity Division, Goyang Special City