State-Space Driven Digital Twin for Condition Monitoring and Predictive Health Assessment in Grid-Integrated Power Converter System

Abstract

This work presents a novel Digital Twin (DT) framework, integrating state-space analysis with advanced optimization for predictive health monitoring in a two-stage, single-phase grid-connected power electronic converter system. The DT is constructed as a high-fidelity mathematical model, continuously synchronized with its Physical System (PS) through real-time sensor-driven data acquisition, enabling seamless condition monitoring and anomaly detection. A well-defined objective function ensures precise system representation by integrating PS sampled data with the DT. To enhance predictive maintenance, an advanced E2FD-HO (Electromagnetic Field and Electrostatic Discharge Hybrid Optimization) algorithm is introduced, offering superior convergence speed and parameter estimation accuracy. Additionally, real-time fault diagnostics are conducted under varying operational conditions, with the inverter’s switching control governed by αβCDSC (alpha-beta Cascaded Delayed Signal Cancellation)-UVT (Unit Vector Template)-based IMPC (Integral Model Predictive Control) strategy, improving transient stability and system adaptability. The FPGA-based OPAL-RT real-time DT validation demonstrates a Percentage Similarity (PST) exceeding 98.55%, underscoring the DT’s robustness in predictive maintenance for system. These findings contribute to pre-fault detection, condition-based monitoring, and secure DT implementations, offering scalable solutions for Industrial Cyber-Physical Systems (ICPS).