PoC for AI-Powered Drone Solution for Smarter Energy Grid Inspection | Insights – Reshape Digital 

At the Cleantech Hackathon during Energy Expo Bucharest, organized by Techcelerator.co, Reshape Digital presented a forward-thinking drone-based inspection platform for utility infrastructure, designed to increase safety, efficiency, and sustainability across power grid maintenance. The hackathon challenged teams to solve real-world problems raised by major energy companies, including PPC Romania, Delgaz Grid, Romgaz, and Nuclearelectrica. Our autonomous drone solution was recognized with a Special Mention for Innovation and High Potential in Energy and Sustainability, reinforcing our mission to apply AI and automation to high-impact industries.

The Challenge: Grid Maintenance That’s High-Stakes, High-Cost

Power grid infrastructure is vast, aging, and essential. Regular inspections are costly, risky, and often delayed by manual effort and limited access. Traditional maintenance methods can't keep up with the demand for faster, more accurate diagnostics in the face of growing climate and energy pressure.

Our Proposed Solution

An Autonomous Drone Platform for Intelligent, Scalable Grid Monitoring

We designed a system that reimagines aerial inspections using AI, edge automation, and innovative energy management. Here’s what it looks like in action:

Autonomous Flight Path Optimization

Using Traveling Salesman Problem solvers, the drone calculates optimal routes within flight-time constraints (e.g., ~40 minutes for DJI Matrice 30T), ensuring maximum coverage with minimal energy usage.

In-Field Charging Innovation

Inspired by recent academic work, we proposed electromagnetic perch-and-charge capabilities, allowing drones to recharge directly from power lines, extending mission duration without landing.

High-Fidelity Data Capture

Equipped with color, thermal, and LiDAR sensors, the drone captures rich media for infrastructure diagnostics. Data is compressed using H265/RAW/TIFF algorithms to reduce storage needs without compromising analysis quality.

Onboard AI Flight Control

An AI agent, trained through Reinforcement Learning or Model Predictive Control, dynamically adjusts flight, avoids obstacles, and triggers inspection actions at optimal points, maximizing data quality and autonomy.

Image Analysis and Fault Detection

Using supervised learning, imagery is processed to:

• Detect network elements
• Classify faults by severity and type
• Automatically generate annotated maintenance reports

From Data to Insight

The system doesn’t just collect data, it creates actionable insight:

• Fault frequency by location or component
• Maintenance cost projections
• Historical reliability metrics (e.g., SAIFI-like indices)

Why It Matters

By combining autonomy, AI, and innovative hardware integration, this solution brings major benefits to energy companies:

• Fewer truck rolls and human risk
• Faster fault response and maintenance
• Better decision-making through historical fault data
• Lower carbon footprint via reduced manual interventions

Recognition for Innovation

Our solution was recognized at the Energy Expo Cleantech Hackathon in Bucharest, where we received a Special Mention for Innovation and High Potential in Energy and Sustainability. This validation reinforces our belief in using technology to tackle systemic infrastructure challenges with real-world impact.

We extend our deep thanks to Techcelerator.co for organizing the hackathon, and to the entire team of mentors, jury members, sponsors, and energy partners who made the event so valuable. Collaborating in an environment of shared purpose and technical challenge was an incredible experience, and we’re honored to have contributed ideas that push the boundaries of sustainability, innovation, and digital transformation in the energy sector.