Key takeaways
A structured network integration testing framework for SSTs was developed
Supports scalable SST architectures (250 kW → 1 MW) through structured integration testing
Identifies critical risks in grid interaction and DC system behaviour
Aligns with international standards and best practices
Enables future deployment readiness through robust validation
The Challenge
Artificial Intelligence (AI) data centres are placing increasing demands on electrical infrastructure due to rapidly rising power density and reliability requirements. These facilities require stable, efficient power delivery capable of supporting high-performance computing workloads while maintaining resilience under dynamic load conditions.
One emerging solution is the Solid-State Transformer (SST), which enables direct conversion from medium-voltage AC to a regulated DC supply. Compared to conventional transformer systems, SSTs offer enhanced controllability, reduced footprint, and improved power quality.
However, the lack of SST-specific standards presents a major barrier to adoption—particularly when it comes to safe and reliable integration into grid-connected environments.
Project Overview
This project focused on developing a structured framework for network integration testing of an SST supplying power to an AI data centre environment.
The system under study was designed to:
- Convert medium-voltage three-phase AC directly into 800 V DC
- Deliver approximately 250 kW of power
- Scale to ~1 MW via parallel operation
- Operate within a grid-connected environment supporting complex loads
The work aimed to:
- Review applicable standards and technical literature
- Define a structured integration test philosophy
- Develop a draft test schedule
- Identify key technical risks and constraints
“Due to the lack of established SST-specific standards, organisations require clear guidance on testing and validation requirements.”
Andreas Avras MEng MIET, R&D Engineer, PNDC
PNDC’s Approach
PNDC provided independent technical expertise to develop a comprehensive network integration testing framework for SST-based systems.
- Key activities included:
- Reviewing international standards for power electronics and grid integration
- Defining system-level test boundaries and objectives
- Developing a test philosophy focused on externally observable behaviour
- Creating a draft schedule covering steady-state, dynamic, and fault scenarios
- Identifying risks such as grid stability and DC bus behaviour
- Upstream grid conditions
- Downstream DC loads
- System-level power network modelling
- Hardware-in-the-loop (HIL) test methodology development
- Grid disturbance simulation
- DC load emulation
- Power quality and protection coordination
The testing philosophy ensures the SST is evaluated based on how it interacts with both:
PNDC capabilities applied:
Outcomes
The project delivered a robust and scalable framework for SST integration testing in data centre environments.
A structured review of international standards
A defined network integration test philosophy
A detailed draft test schedule
Identification of key risks and operational constraints
Recommendations for future testing and deployment
This framework enables SST systems to be tested and validated under representative operating conditions, improving confidence in performance and reliability while mitigating integration risks and supporting deployment readiness.
Through this project, delivered in collaboration with the Semiconductor Catapult, PNDC enabled structured validation of emerging power technologies, delivering measurable value.
Semiconductor Catapult is proud to have supported this work with PNDC. This project demonstrated the value of collaboration in advancing solid-state transformer technology for emerging applications such as AI data centres. Work of this kind is strategically important in positioning the UK as an early developer and adopter of these technologies.
Connect with PNDC’s project lead, Andreas Avras, on LinkedIn.