Sequential finite element modelling of lightning arc plasma and composite specimen thermal-electric damage

Abstract

Highly complex phenomena such as lightning strikes require simulation methods capable of capturing many different physics. However, completing this in one simulation is not always desired or possible. In such instances there can be a need for a methodology to transfer loading boundary conditions from one simulation to the next while accounting for the characteristic form of the loading and the dissimilar domain and mesh geometries. Herein, the objective is to combine two models to enable the automatic sequential simulation of a lightning arc and a composite test specimen. The approach is developed using Finite Element models, with a Magnetohydrodyanmics model representing the lightning plasma and a thermal-electric model representing the specimen. The specimen mesh and loading boundary conditions are automatically generated based on the predicted output of the preceding plasma model. The precision, run-time and flexibility of the proposed approach is demonstrated, with thermal damage predictions generated in approximately 33 hours. Resulting from the integrated modelling capability is the first time prediction of damage representing the test electric boundary conditions rather than assumed specimen boundary conditions (herein using test 'Waveform B').

Coupled Thermal-Mechanical Progressive Damage Model with Strain and Heating Rate Effects for Lightning Strike Damage Assessment

“…The same python script methodology, proposed by Foster et al [15], was used to transfer the temperature profile from the thermal-electric to the dynamic, temperature-displacement, explicit simulation. A Waveform B simulation was undertaken by loading the mesh with the plasma simulation outputs from previous publications [21,49]. In this case the temperature-displacement simulation was run for a total of 0.03 ms because at this point the thermal damage had stopped increasing and pressure loading had normalised to atmospheric pressure.…”

“…The peak pressure was assumed to be 50 MPa as this was quoted by Foster et al as the peak corresponding with the experimental works of Hirano, Munoz and Chemartin et al and which was considered herein for simulation prediction verification [1,11,13,51]. Although Foster's loading data set for scaled Waveform A was based on a series of argued assumptions [13][14][15] and the loading data for Waveform B was based on simulation output [19,49] each was considered as the best available to-date for this study. The same temperature dependent mechanical properties, strain rate dependent fracture and strength properties, interlaminar properties and specimen boundary conditions were therefore used in the Waveform A analysis (matching the Waveform B analysis), Tables 3, 4 and 6. These simulations were chosen as the Waveform A variant allowed for the prediction of damage due to short time period, high peak current loads and comparison with the published, single physics simulations from Foster et al [13,15], as well as the experimental results of Hirano et al [1].…”

Understanding the Impact of Standardized SAE Waveform Parameter Variation on Artificial Lightning Plasma, Specimen Loading, and Composite Material Damage

“…However, in this work the waveform was held constant as a standard Waveform B profile. Millen et al and Chen et al are the only authors to couple their plasma model with a specimen damage model, however, both authors modelled a single, fixed waveform [24], [35].…”

“…Lightning strike thermal damage simulations have been developed with varying meshes and loading methods used within these simulations [17], [20], [22], [24], [25], [35], [41]. Ogasawara et al [20] were the first authors to present a clear thermal-electric simulation for Waveform A, replicating the experimental work of Hirano et al [1].…”

“…Authors have also used FE modelling to predict the likely damaged composite specimen due to incident current, temperature or pressure loading [17], [18], [20], [22], [23], [25], [35]- [37]. However, recently measures have been taken to combine plasma modelling and specimen material damage predictions [21], [24], [35], due to the complexity and computational burden of the required models. Thus limited understanding exists on the influence of plasma modelling on the prediction of specimen damage, and vice versa.…”