FPGA-Based Digital Twin Implementation for Power Converter System Monitoring

Implementing closed-loop control requires ensuring a robust system response against undesired disturbances and random fault events. To overcome these challenges, the controllers must be able to detect and adapt the process behavior against undesired events by adjusting its parameters accordingly. Developers have utilized the concept of a digital twin—a real-time representation of the physical process—to design and update the physical controller effectively. However, traditional digital twin implementations often involve significant data exchange between the digital system and the real asset through cloud platforms, leading to data latency and privacy issues. To mitigate these concerns, we propose an FPGA-based digital twin implementation where the information is directly sourced into the Digital Twin from the physical asset, which runs in parallel with the real system. This setup eliminates the need for big data transfers and cloud uploads, ensuring enhanced data privacy and facilitating a faster and more efficient digital twin implementation and update process. To demonstrate the capabilities of embedded Digital Twin, we present a case study involving monitoring a power converter system during a sensor fault scenario.

I. Introduction

The concept of real-time controller improvement has gained significant importance in achieving effective process in the loop control despite unpredictable disturbances. Until recently, classical control methods relied on approaches like model predictive control (MPC) [1], adaptive control methods [2], and offline optimization methods [3]. However, these methods were often either too mathematically complex or impractical for implementation on microcontrollers.