Comparison between linear frequency-modulated ultrasonic excitation signals and pure tones for concrete applications

Abstract

Non-destructive testing has supported prognosis and health management in the construction industry for decades. Among them, ultrasonic tests (UT) are increasingly used to assess the structural integrity of the concrete. Typically, the UT uses pure tones as the excitation signal. However, an alternative is to use a linear frequency-modulated pulse (LFMP) followed by compression. This process improves the signal-to-noise ratio, provides broadband frequency responses, and is less susceptible to propagation dispersion. This work compared the ultrasonic transit times (UTT) of LFMP and pure tones. Three nondispersive models for concrete materials were numerically simulated using the k-Wave MATLAB toolbox. Additionally, a sample of reactive powder concrete (RPC) was experimentally tested. The frequencies ranged from 50 kHz to 750 kHz for both simulations and experiments. The simulation results presented LFMP relative errors in UTTs ranging from 0.0 % to 2.7 % compared to the theoretical values, whereas for pure tones, the RE was 2 % to 63 %. Experimentally, as no theoretical value can be considered, the quantitative comparison was the measurement dispersion under repeatability conditions. LFMPs disclosed a standard deviation in transit time of 35 ns, whereas for pure tones, it ranged from 43 to 51 ns. This research reinforces the potential of LFMP excitation compared to traditional pure tones, providing more accurate and precise ultrasonic velocity measurements in heterogeneous environments.