Hybrid Structural Health Monitoring for Impact Damage in PLA Plates Using SLDV and the Electromechanical Impedance Method ^†

Abstract

With the growing use of 3D-printed polymers in structural applications, understanding their damage response under impact is critical for reliability and safety. This study investigates the impact response and damage progression in Fused Deposition Modelling (FDM)-printed Polylactic Acid (PLA) plates with varying infill densities (40%, 60%, and 100%) using a combination of scanning laser Doppler vibrometry (SLDV) and electromechanical impedance (EMI) techniques. Progressive impacts were applied in four stages, and damage was evaluated through wave attenuation, impedance deviation, and phase distortion metrics. Results show that lower infill densities exhibit more severe degradation, with increased damping and poor wave transmission, while 100% infill demonstrates higher damage resistance and better detectability. The findings underscore the importance of infill design in optimizing mechanical performance and structural health monitoring in additively manufactured components.

1. Introduction

Applications in polymer and composite additive manufacturing (AM) range from the creation of prototypes for art, architecture, medicine, and aerospace to the fabrication of large structures, molds, and dies. In the medical industry, AM is used for custom-made fixtures, surgical tools, and implants [1]. The increasing adoption of AM in structural applications necessitates advanced structural health monitoring (SHM) techniques to ensure reliability and detect damage. The fused deposition modeling (FDM) process is the most widely used AM process for polymers. A variety of thermoplastics ranging from polylactic acids (PLA) to engineered high-temperature polymers like nylon6 and polycarbonate (PC) are used in the FDM process [2]. Although, FDM is relatively fast and cost efficient compared to SLA and SLS, higher porosity levels are observed in the final printed parts. Jin et al. have distinguished the infill densities in 3D-printed acrylonitrile butadiene styrene (ABS) samples using effective density images by ultrasonic imaging [3]. Scheyer et al. monitored the progressive effect of simulated damage and drilled holes on the overall response of the printed PLA plate with embedded sensors using EMI method [4]. Recently, Fakih et al. established the relationship between damage sensitivity and variation in infill density in PLA plates based on guided wave measurements using a scanning laser Doppler vibrometer [5].

However, a significant challenge in AM structures, particularly polylactic acid (PLA) plates, is their susceptibility to impact-induced damage, which can degrade mechanical performance. Traditional damage detection methods often fail to capture subtle internal defects and variations in structural integrity caused by impact events. To address this, the present study employs a hybrid SHM approach that integrates wave-based sensing via scanning laser Doppler vibrometer (SLDV) and electromechanical impedance (EMI) measurements to assess impact damage in PLA plates with varying infill densities (40%, 60%, and 100%). The successive impacts with gradually increasing impact energy are performed on each sample to understand the effect of infill density on the structural response and sensitivity of damage identification.

2. Experimental Setup

PLA plates with a size of 325 mm × 235 mm × 4 mm were printed using FDM with a 40%, 60%, and 100% infill density. The plates were printed by an S5 printer (Ultimaker, Zaltbommel, The Netherlands) using 0.4 mm infill line width and by varying the infill line distances to 2 mm and 1.33 mm for the 40% and 60% infill density samples, respectively. A circular 10 mm diameter Lead Zirconate Titanate (PZT) transducer (CeramTec, Plochingen, Germany) was placed at the middle side edge, as shown in Figure 1, which was used for the excitation of a five-cycle Hann-windowed sinusoidal signal of 200 kHz central frequency. An area scan of 101 mm × 101 mm around the center of the sample was performed using an SLDV (Polytec, Waldbronn, Germany) to receive the propagating waves. The complete experiment set is shown in Figure 2. Also, the impedance signatures were measured after each loading (impact).

The specimen was impacted gradually at the center. A 20 mm diameter steel ball of 55 gms mass was used as the impactor. Impacts were applied in four distinct stages, with increasing energy input at each stage to simulate progressive damage at the center of the plate within the scanned area, as shown in Figure 1. Stage 1 involved a single low-energy impact of 0.54 J, intended to initiate minor surface damage. Stage 2 increased the impact energy to 1.03 J, while Stage 3 involved a further increase to 1.37 J; both were applied as single impacts to the same location to promote crack propagation and deeper structural damage. Stage 4 consisted of a severe loading scenario, where each impact simulated the effect of 30 consecutive hits at 1.37 J, totaling 41.1 J of energy per impact site. This high-energy stage was intended to replicate extensive mechanical fatigue and material failure.

Due to these impacts, progressive damage was observed at the impact location, as shown in Figure 3. The intensity of the damage differed for the samples with different infill densities. In the 40% low infill density sample, which was the weakest of the samples, an indentation was seen at Stage 2. The damage increased with successive impacts, and after Stage 4 a 2.7 mm indentation could be seen, as shown in Figure 3. However, the 60% infill specimen showed only moderate damage. After Stage 4, an indentation of 1.5 mm occurred. For the 100% infill specimen, which was the strongest, even after Stage 4, only a small indentation of 0.4 mm was observed, as shown in Figure 3.

3. Result and Discussion

3.1. SLDV Wavefield Visualizations

A one-dimensional scanning laser Doppler vibrometer (1D SLDV) was employed to measure the out-of-plane component of wave velocity (v) along the direction of the laser beam reflected from the surface of a structure. This technique enabled the visualization of ultrasonic wave propagation across the surface of the tested specimen.

The frames of the propagating wavefield were processed using wavenumber-based adaptive filtering and signal analysis, following the approach outlined in [6]. In this method, the out-of-plane velocity field $v (x,y,t)$, measured at a dense grid of points using the scanning laser Doppler vibrometer, is first transformed into the frequency–wavenumber domain using a three-dimensional Fourier transform. Wavenumber adaptive filtering is then applied to isolate specific Lamb wave modes, which improves the signal-to-noise ratio by suppressing reflections and undesired wave components. The filtered signals are subsequently transformed back into the time–space domain, where energy-based root mean square indices (ERMSF) are calculated for each spatial point. These indices quantify the local scattering of wave energy due to damage, thus serving as a damage-sensitive feature. To account for amplitude reduction caused by geometric spreading and material damping, attenuation compensation is applied in the time domain, independent of propagation direction. The resulting ERMSF values are assembled into spatial maps, where localized increases in index magnitude correspond to damage regions. For consistency, all maps are displayed with a uniform color scale (Figure 4), thereby enabling direct comparison across plates with different infill densities. The main objective of this process is to amplify damage-related features and provide clear visual contrast between intact and damaged zones.

Given that Stage 4 of impact loading resulted in the most significant damage across all PLA specimens, wavefield visualizations and corresponding waveform patterns at approximately 91 µs (the time when the guided wavefront approaches the impact site) are presented in Figure 4 for detailed comparison. As shown earlier in Figure 3, the extent of physical damage varies considerably across specimens with different infill densities, despite being subjected to the same impact energy. This variation is also reflected in the full-field SLDV data.

In the 40% infill specimen, although the indentation is visibly the most severe (2.7 mm deep), the ability to localize damage using SLDV becomes challenging. The low infill density leads to increased scattering and attenuation of wave energy, which diminishes the clarity of wavefront reflections and interferometric patterns associated with damage zones. The disrupted wave propagation reduces contrast in the amplitude-based maps, thereby masking the damage signature. Conversely, in the 60% infill sample, the moderate internal density provides a better balance between structural stiffness and wave transmission efficiency. As a result, the impact-induced discontinuities are more distinctly visible in the SLDV data, with localized amplitude drop and waveform distortion near the damaged region indicating the presence of structural degradation. In the case of the 100% infill specimen, despite only a minor indentation of 0.4 mm observed post-impact, the SLDV measurements clearly capture the subtle wave disturbances at the impact site. The superior structural integrity and consistent material distribution in this specimen allow the guided wave to propagate with minimal distortion, thereby enhancing the detectability of even small damage features. This highlights that higher infill not only improves mechanical resistance but also enables better damage localization through wave-based diagnostic methods.

3.2. Damage Metrics

The damage evolution in PLA specimens subjected to progressive impact loading was assessed using a multimodal monitoring approach combining guided wave propagation (SLDV) and electrical impedance (EMI) analyses. The results, detailed across three key damage-sensitive metrics as normalized root mean squared (RMS) amplitude, root mean square deviation (RMSD), and mean absolute phase deviation (MAPD), highlight the role of infill density in governing the mechanical resilience and damage detectability of the structures.

3.2.1. Normalized RMS Amplitude

Normalized RMS amplitude, extracted from SLDV-acquired wavefields, serves as a sensitive measure of signal integrity and energy transport. A pronounced amplitude reduction is observed for all infill densities as impact severity increases, but the rate and extent of attenuation are strongly dependent on the infill level. For a conductance signal with damage $D$ over n frequency points, normalized RMS can be calculated against healthy H (without impact) conductance signal as,

$$\mathrm{N}\mathrm{o}\mathrm{r}\mathrm{m}\mathrm{a}\mathrm{l}\mathrm{i}\mathrm{z}\mathrm{e}\mathrm{d} \mathrm{R}\mathrm{M}\mathrm{S}={\displaystyle \frac{\sqrt{\frac{1}{n}\sum _{i=1}^n{D}^2}}{\sqrt{\frac{1}{n}\sum _{i=1}^n{H}^2}}}$$

(1)

The waveforms were selected near the impact location for all specimens (shown as red dots in Figure 4). The 40% infill specimen exhibits a steep reduction, with the normalized amplitude dropping to 45% of its initial value by Impact 15, as shown in Figure 5. This substantial energy loss is attributed to the accumulation of internal damage mechanisms such as matrix cracking, void growth, and delamination, all of which act as scattering and absorbing sites for the propagating wave.

In contrast, the 100% infill sample retains around 60% of its initial amplitude, indicating that the denser internal architecture provides a more continuous load path and impedes the development of major scattering centers. This difference is especially evident in the nonlinear progression: while the initial impacts produce minor reductions in amplitude, later stages show disproportionately larger losses, particularly in the lower infill specimen. This suggests that damage does not accumulate linearly, and once a critical threshold is exceeded, especially in less dense structures, the mechanical integrity rapidly degrades.

3.2.2. Root Mean Square Deviation

The RMSD computed from EMI conductance signatures further supports the wave-based observations. The RMSD values increase monotonically with impact severity across all infill levels, but with markedly different magnitudes. The 40% infill specimen shows an RMSD of 24% at Stage 4, indicating substantial deviation from the baseline, driven by global stiffness reduction and altered boundary conditions caused by progressive internal failure, as shown in Figure 6. The RMSD for the conductance signals of damaged and healthy specimens is calculated as,

$$\mathrm{R}\mathrm{M}\mathrm{S}\mathrm{D}=\sqrt{{\displaystyle \frac{{\sum }_{i=1}^n{(D_i-H_i)}^2}{H_i^2}}}$$

(2)

where n is the number of frequency spectrum samples, H is the healthy state conductance signal, and D is the damaged state conductance signal after impact.

In contrast, the 100% infill sample remains below 7%, even after high cumulative energy input (411 J). This limited deviation aligns with the smaller amplitude loss, reaffirming the mechanical robustness of the dense infill structure. The RMSD pattern emphasizes that electrical impedance methods are effective in capturing not only localized damage but also global system-level degradation, making them a valuable complement to wave-based methods. The sharp contrast between the RMSD trends of 40% and 100% infill specimens underscores the critical role of internal density in governing structural response under repeated impact.

3.2.3. MAPD

The MAPD metric tracks the deviation in phase angle and is particularly sensitive to changes in structural stiffness and the emergence of nonlinear dynamics. The MAPD for the conductance signals of damaged and healthy specimens is calculated as,

$$\mathrm{M}\mathrm{A}\mathrm{P}\mathrm{D}={\displaystyle \frac{100}{n}}\sum _{i=1}^n{{\displaystyle \frac{{(D}_i-H_i)}{H_i^2}}}^2$$

(3)

where n is the number of frequency spectrum samples, H is the healthy state conductance signal, and D is the damaged state conductance signal after impact.

A sharp rise in MAPD to 35% is observed in the 40% infill specimen at Stage 4, reflecting significant alteration in the dynamic stiffness and possible boundary condition changes due to local softening, as shown in Figure 7. This is consistent with the visible 2.7 mm indentation observed in the specimen and aligns with both RMSD and wave amplitude observations, together conclusive of severe structural compromise.

The 100% infill specimen shows a MAPD increase to only 10%, indicating limited phase distortion and retention of the original structural response. The relatively lower deviation further supports the idea that the dense infill matrix prevents the development of large, stiffness-altering damage modes. Importantly, the nonlinear growth of MAPD across impact stages, particularly in the 40% infill specimens, arises from the complex redistribution of internal stresses and geometry-induced anisotropy in the porous FDM lattice. As impacts progress, the thin PLA filaments within low-infill regions undergo localized buckling, filament fracture, and collapse of internal voids. These mechanisms cause irregular variations in local stiffness and wave speed, leading to abrupt phase shifts in the EMI response. This nonlinear MAPD evolution is, therefore, not just a function of surface indentation but a signature of distributed internal reconfiguration of the printed architecture under cumulative impact loading.

4. Conclusions

This work demonstrates the effectiveness of combined guided wave and electromechanical impedance techniques for assessing impact-induced damage in FDM-fabricated PLA structures with varying infill densities. The experimental results confirm that infill density significantly influences both mechanical degradation and the ability to detect damage. Low-density (40%) infill specimens suffer early and severe structural compromise, which reduces wave clarity and impairs diagnostic resolution. Medium-density (60%) structures show moderate damage progression with reasonably detectable wave interactions. In contrast, fully dense (100%) PLA plates resist damage effectively and maintain strong, interpretable diagnostic signals. The ability to detect even subtle damage in high-infill specimens highlights their suitability for structural applications where both mechanical integrity and continuous monitoring are essential. These insights offer a foundation for optimizing the design of additively manufactured components where durability and inspectability are equally critical.