Displacement Estimation of Reactor Containment Buildings Using Accelerometer and FMCW Millimeter-Wave Radar

Abstract

This study presents an advanced fusion-based sensing methodology designed specifically for precise displacement and acceleration monitoring of Reactor Containment Buildings (RCBs). Reliable measurement of structural displacement in RCBs is critical for ensuring structural integrity and operational safety. To address challenges associated with traditional displacement monitoring techniques, this research integrates data from a three-axis MEMS accelerometer and phase-shift measurements obtained from millimeter-wave frequency-modulated continuous-wave (FMCW) radar. By fusing these distinct sensor modalities, we significantly enhance measurement accuracy, achieving displacement estimation resolutions superior to 0.1 mm for both static and dynamic structural responses. The proposed sensor's hardware architecture emphasizes cost-efficiency, compactness, and reliability. The sensor system employs an Infineon 60 GHz radar chipset, combined with a MEMS accelerometer and a microcontroller supporting multiple wireless communication protocols (2.4/5.0 GHz Wi-Fi, Bluetooth, and Gigabit Ethernet). This robust yet economical sensor design results in a compact device (100 × 80 × 50 mm³) and is projected to cost under $1,000 per unit when mass-produced, making it highly suitable for wide-scale deployments in nuclear facilities. A key innovation of the developed fusion sensor is its automated besttarget selection capability, which utilizes initial short-duration radar and accelerometer readings to efficiently determine the most stable structural reflections and establish accurate conversion factors between line-of-sight displacement and actual structural displacement. Additionally, the accelerometer data aids in addressing the radar phasewrapping issue through an adaptive phase-unwrapping algorithm, thus significantly enhancing measurement reliability and accuracy. To further improve displacement estimation, radar-derived displacement data is combined with accelerometer measurements using a finite impulse response (FIR) filter, which mitigates noise and drift typically observed in acceleration-based displacement estimations. Through experimental validations conducted on structural models, this sensor fusion method demonstrated improved displacement measurement accuracy with errors consistently below 1 mm. Given its capacity for automatic calibration, effective target selection, and enhanced resilience against occlusions and environmental interferences, the sensor offers a reliable solution for continuous monitoring of RCBs. Ultimately, the proposed sensor fusion approach presents significant potential for adoption in nuclear power facilities, enhancing predictive maintenance, optimizing structural health management strategies, and reinforcing overall operational safety.