The explicit FE solver LS-DYNA is nowadays widely used in numerous engineering applications for simulating complex physical phenomena in a virtual environment, particularly those characterized by high strain-rates and large deformations, such as high-velocity impacts. These simulations require accurate and reliable material characterization to correctly reproduce material behavior. This is a fundamental step for the correct numerical reproduction of material response to severe loading conditions. In the present work, six high-strength metallic alloys were calibrated using the Johnson-Cook constitutive model and the Cockcroft-Latham and Johnson-Cook damage models. An extensive experimental campaign was conducted at the JRC in Ispra (VA), Italy, during which the alloys were subjected to static and dynamic tensile tests, respectively performed using a standard tensile testing machine and Split Hopkinson Pressure Bar (SHPB) apparatuses. From the tensile tests, high-speed images of tensile tests and the force v. displacement curves were obtained, however the latter were not used for the calibration of the materials, due to possible unwanted non-ideal effects that influenced the measurements. Instead, DIC was employed to reconstruct the alloys' stress-strain curves from the high-resolution images, and, using an iterative error-minimization fitting algorithm, the strain- and strain-rate-hardening constants were determined. In addition to the Cockcroft-Latham damage criteria, the constitutive and damage models were successfully validated by reproducing the tensile tests in LS-DYNA and compare the numerical results with experimental data. The material response and failure were correctly reproduced by the numerical model, with maximum relative errors lower than 10%. The thermal softening parameter was obtained through inverse calibration, while the Johnson-Cook damage model was calibrated by fitting its formulation — considering only the strain-hardening influence on the failure strain — to a re-formulation of the Cockcroft-Latham criteria. This approach produced acceptable parameter values and correctly reproduced material failure. The calibrated alloys, using the Johnson-Cook constitutive and damage models, were then employed in an LS-DYNA ballistic impact simulation to replicate depth-of-penetration (DOP) tests previously conducted. The finite element model was successfully validated: it correctly predicted the depth of penetration observed in the ballistic tests, with an error of 4%.
Il solutore esplicito agli elementi finiti LS-DYNA è oggi comunemente impiegato in numerose applicazioni ingegneristiche per la simulazione di fenomeni fisici complessi in ambiente virtuale, in particolare fenomeni caratterizzati da elevati tassi di deformazione e grandi deformazioni, quali gli impatti ad alta velocità. Tali simulazioni richiedono una caratterizzazione del materiale accurata e affidabile, al fine di riprodurre correttamente i complessi comportamenti dei materiali. Questo rappresenta un passaggio fondamentale per la corretta riproduzione numerica della risposta del materiale a condizioni di carico severe. Nel presente lavoro sei leghe metalliche ad alta resistenza sono state calibrate secondo il modello costitutivo di Johnson–Cook e i criteri di danneggiamento di Cockcroft–Latham e Johnson–Cook. Un’estesa campagna sperimentale è stata condotta presso il JRC di Ispra (VA), Italia, dove le leghe sono state sottoposte a prove di trazione statiche e dinamiche eseguite rispettivamente mediante una macchina universale per prove di trazione e un apparato Split Hopkinson Pressure Bar (SHPB). Dalle prove di trazione sono state ottenute immagini ad alta velocità e curve forza–spostamento; tuttavia, queste ultime non sono state utilizzate per la calibrazione dei materiali a causa di possibili effetti non ideali indesiderati che hanno influenzato le misurazioni. In alternativa, la tecnica di Digital Image Correlation (DIC) è stata impiegata per ricostruire le curve sforzo–deformazione delle leghe e, mediante un algoritmo iterativo di minimizzazione dell’errore, sono state determinate le costanti di incrudimento e di sensibilità alla velocità di deformazione. Il modelli costitutivo e di daneggiamento sono stati validati confrontando i risultati numerici ottenuti tramite simulazioni in LS-DYNA con i dati sperimentali. La risposta del materiale e la modalità di rottura sono state riprodotte correttamente dal modello numerico, con errori relativi inferiori al 10%. Il parametro dell'addolcimento termico è stato determinato mediante calibrazione inversa, mentre il modello di danneggiamento di Johnson–Cook è stato calibrato adattando la sua formulazione — considerando esclusivamente l’influenza dell'incrudimento sulla deformazione a rottura — a una riformulazione del criterio di Cockcroft–Latham. Tale approccio ha fornito parametri accettabili e ha consentito di riprodurre correttamente la rottura del materiale. Le leghe calibrate secondo i modelli costitutivo e di danneggiamento di Johnson–Cook sono state quindi impiegate in una simulazione balistica di impatto in LS-DYNA, finalizzata alla riproduzione di prove di profondità di penetrazione precedentemente condotte presso un poligono militare. Il modello agli elementi finiti è stato validato con successo, risultando in grado di prevedere correttamente la profondità di penetrazione misurata nelle prove balistiche, con un errore del 4%.
Experimental and numerical characterization of metallic alloys for impact simulations
Barbieri, Sara
2025/2026
Abstract
The explicit FE solver LS-DYNA is nowadays widely used in numerous engineering applications for simulating complex physical phenomena in a virtual environment, particularly those characterized by high strain-rates and large deformations, such as high-velocity impacts. These simulations require accurate and reliable material characterization to correctly reproduce material behavior. This is a fundamental step for the correct numerical reproduction of material response to severe loading conditions. In the present work, six high-strength metallic alloys were calibrated using the Johnson-Cook constitutive model and the Cockcroft-Latham and Johnson-Cook damage models. An extensive experimental campaign was conducted at the JRC in Ispra (VA), Italy, during which the alloys were subjected to static and dynamic tensile tests, respectively performed using a standard tensile testing machine and Split Hopkinson Pressure Bar (SHPB) apparatuses. From the tensile tests, high-speed images of tensile tests and the force v. displacement curves were obtained, however the latter were not used for the calibration of the materials, due to possible unwanted non-ideal effects that influenced the measurements. Instead, DIC was employed to reconstruct the alloys' stress-strain curves from the high-resolution images, and, using an iterative error-minimization fitting algorithm, the strain- and strain-rate-hardening constants were determined. In addition to the Cockcroft-Latham damage criteria, the constitutive and damage models were successfully validated by reproducing the tensile tests in LS-DYNA and compare the numerical results with experimental data. The material response and failure were correctly reproduced by the numerical model, with maximum relative errors lower than 10%. The thermal softening parameter was obtained through inverse calibration, while the Johnson-Cook damage model was calibrated by fitting its formulation — considering only the strain-hardening influence on the failure strain — to a re-formulation of the Cockcroft-Latham criteria. This approach produced acceptable parameter values and correctly reproduced material failure. The calibrated alloys, using the Johnson-Cook constitutive and damage models, were then employed in an LS-DYNA ballistic impact simulation to replicate depth-of-penetration (DOP) tests previously conducted. The finite element model was successfully validated: it correctly predicted the depth of penetration observed in the ballistic tests, with an error of 4%.| File | Dimensione | Formato | |
|---|---|---|---|
|
Tesi_SaraBarbieri_10995745Final.pdf
non accessibile
Descrizione: The explicit FE solver is widely used in numerous engineering applications to simulate complex physical phenomena in virtual environments, particularly those characterized by high strain rates and large deformations, such as high-velocity impacts. These simulations require accurate and reliable material characterization to correctly reproduce material behavior. This is a fundamental step for the accurate numerical reproduction of the material response under severe loading conditions. In the present work, six high-strength metallic alloys were calibrated using the Johnson-Cook constitutive model and the Cockcroft-Latham and Johnson-Cook damage models. An extensive experimental campaign was conducted at the JRC in Ispra (VA), Italy, during which the alloys were subjected to static and dynamic tensile tests, performed using a standard tensile testing machine and a Split Hopkinson Pressure Bar (SHPB) apparatus, respectively. From the tensile tests, high-speed images and force-displacement curves were obtained; the latter were not used for material calibration due to possible non-ideal effects that could have influenced the measurements. Instead, DIC was employed to reconstruct the alloys' stress-strain curves from the high-resolution images, and, using an iterative error-minimization fitting algorithm, the strain- and strain-rate-hardening constants were determined. In addition to the Cockcroft-Latham damage criteria, the constitutive and damage models were successfully validated by reproducing the tensile tests in the numerical environment and comparing the numerical results with experimental data. The material response and failure were correctly reproduced by the numerical model, with maximum relative errors lower than 10\%. The thermal softening parameter was obtained via inverse calibration, while the Johnson-Cook damage model was calibrated by fitting its formulation—considering only the strain-hardening influence on the failure strain—to a re-formulation of the Cockcroft-Latham criteria. This approach produced acceptable parameter values and correctly reproduced material failure. The calibrated alloys, using the Johnson-Cook constitutive and damage models, were then employed in ballistic impact simulations to replicate depth-of-penetration (DOP) tests previously conducted. The finite element model was successfully validated: it correctly predicted the depth of penetration observed in the ballistic tests, with an error of 4\%.
Dimensione
14.46 MB
Formato
Adobe PDF
|
14.46 MB | Adobe PDF | Visualizza/Apri |
|
Executive_Summary_10995745Final.pdf
non accessibile
Descrizione: High-velocity impacts are complex phenomena characterized by high strain rates, large deformations, and rapid energy exchange. To model such events in a numerical environment, all materials involved must be accurately and reliably calibrated to ensure accurate prediction of material behavior and structural response. The most widely used finite element (FE) code for this type of simulation is LS-DYNA, an explicit FE solver specialized in simulating material response under severe loading conditions. In this work, LS-DYNA will be used to replicate ballistic tests of an armor-piercing projectile against a multilayer steel target. The aim of the simulations is to correctly predict the depth of penetration of the ammunition: the numerical DOP will be compared with the experimental DOP measured during the ballistic tests. However, before developing the impact simulations, the materials used must be accurately and reliably calibrated. The experimental campaign was conducted at the Joint Research Center in Ispra (VA), Italy, where six metallic alloys were subjected to static and dynamic tensile tests at four different strain rates, using a standard tensile testing machine and Split Hopkinson Pressure Bars, respectively. Based on the experimental data, the materials were calibrated using the Johnson-Cook constitutive model and the Cockcroft-Latham and Johnson-Cook damage models. Because of the sensitive nature of the materials studied in this project, the calibrated parameters will be reported in normalized form. Furthermore, stress or force values will not be reported in any figure depicting the materials' mechanical response.
Dimensione
1.74 MB
Formato
Adobe PDF
|
1.74 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/252854