Transformers play a critical role in ensuring stable and reliable power supply.

Among onepredict’s solutions, ‘guardione® substation’ is specifically designed to accurately monitor oil-immersed transformers based on vast amounts of data. For those interested in learning more about guardione® substation, we spoke with Jae-kyung Shin, PM at onepredict and an original team member who has been part of the solution’s history from the very beginning.

Q. Could you introduce yourself, your role at the company, and the path you’ve taken to get here?

Hello, since joining onepredict, I’ve been working as a Product Manager (PM), planning and developing various AI-based solutions. I have a special attachment to guardione® substation, as I’ve been involved since the initial design phase. My responsibilities include setting the product direction and developing innovative features that solve real problems for our clients.

Q. You've been involved since the initial design stage of guardione® substation, which adds credibility to this interview. For those unfamiliar with the product, could you give us a simple explanation?

guardione® substation is an AI-based O&M (Operation & Maintenance) solution that accurately diagnoses and predicts the condition of critical equipment such as transformers. It analyzes dissolved gas (DGA) data generated inside transformers to diagnose the current condition, predict potential failures in advance, and provide maintenance recommendations.

It is primarily adopted by power plants, large-scale manufacturing plants, and utility companies—businesses where stable equipment operation and management are crucial. guardione® substation is optimized to reduce downtime, lower maintenance costs, and support efficient decision-making for operators. To add a bit of pride, our existing clients have said that the solution not only solves real problems but also elevates their maintenance strategies to a new level.

Q. Managing critical equipment like transformers sounds like a solution that many companies would need. How did guardione® substation come to be, and what were the key focuses during its development?

guardione® substation was born from the technical know-how gained through a joint development project with KEPCO (Korea Electric Power Corporation).

Traditional diagnostic methods based on international standards relied on single gas values, which made it easy to miss faults. To address this, we designed an AI-based diagnostic algorithm that analyzes the correlation between gas values (composition ratios) to improve diagnostic accuracy.

We also listened to voices from the field during development. Many sites lacked a proper data management system for transformers, highlighting the need for a solution that could systematically manage dissolved gas data.

Based on these insights, guardione® substation was developed with a focus on (1) systematic data management, (2) more accurate AI-based analysis, and (3) fault prediction capabilities.

Q. It sounds like guardione® substation was born from extensive research and feedback from the field. Did the final product fully reflect those insights?

Yes, after a great deal of effort from many people, guardione® substation was designed with a strong focus on meeting the real needs of the field.

In particular, it was designed not only to diagnose transformer conditions but also to manage maintenance history in an integrated way, helping operators efficiently manage equipment and plan maintenance schedules. Based on AI algorithms verified through joint development with KEPCO, it offers more precise analysis of transformer conditions and predictive capabilities that go beyond what international standards provide—making it a solution that can prevent failures in advance.

It also incorporates features such as human error detection and correction algorithms, as well as an intuitive user experience.

Q. guardione® substation continues to evolve into an even better solution. After the upgrade to version 2.0 in 2023, what has changed?

The v2.0 upgrade focused heavily on improving the UI/UX to enhance user convenience.

We redesigned the dashboard and introduced equipment search and card view features, allowing users to intuitively check transformer conditions and maintenance history. In addition, we strengthened the maintenance recommendation and history management features so that operators can easily understand and act on recommendations based on condition changes. Notably, we added a measurement reliability evaluation function that detects and corrects errors that may occur during insulating oil measurement, increasing data reliability and maximizing work efficiency.

Q. It sounds like a very promising solution. What are the future upgrade plans for guardione® pdx?

Going forward, guardione® pdx will focus on enhancing its CMMS capabilities and offering a more customized environment tailored to client needs.

We plan to incorporate AI-powered analysis using existing PM (Preventive Maintenance) data, integrating equipment condition and historical maintenance data to automatically prioritize maintenance tasks. This will allow technicians to focus on the most critical equipment and make maintenance scheduling and resource management more efficient.

In addition, we will improve the BI-based user experience by allowing customizable dashboard layouts and highlighting equipment based on priority, depending on field requirements. This will enable us to provide optimized data for each client’s operational environment and goals, supporting more strategic asset management.

Q. Expanding the contract from 24 to 100 units clearly shows the high level of client satisfaction. That must have been very rewarding. While guardione® substation has already shown impressive results, we’re curious about what’s next. How do you plan to further develop the solution?

guardione® substation is set to evolve with a focus on (1) real-time data analysis and (2) enhanced reliability.

First, we plan to integrate online DGA to enable real-time data collection and analysis. This will allow for quicker identification of transformer conditions and early detection of potential faults, ultimately improving field responsiveness.

Second, we aim to introduce explainable AI (XAI) to make our currently black-box AI models more transparent and trustworthy for users. By clearly presenting the AI’s decision-making process and results, we hope to help users utilize the solution more effectively.

We spoke with the guardione® substation team, who are dedicated to solving the real challenges customers face in the field every day. PM Jaekyung Shin emphasized that guardione® substation is not just about technical excellence, but about deeply understanding the ‘real’ pain points on site and continuously seeking ways to resolve them. We could truly sense the team’s sincerity in striving to improve operational efficiency and safety for their clients.