The victim of the aforementioned workplace accident sustained a skull fracture and various injuries to the brain. These injury types are associated with different injury mechanisms and have therefore been studied separately in this study – see a detailed motivation behind this methodological choice in Supplementary Appendix D.

A schematic overview of the study approach is provided in Fig. 3. In the following sections, the reconstruction procedure will be described in detail starting with a description of the current accident case and any made inferences, followed by a presentation of the skull fracture and brain injury prediction approach. All referred-to reconstructions were preprocessed in LS-PrePost v4.8 and simulated in LS-DYNA v13 using multiple CPUs. Postprocessing was done in MATLAB v2021b. 3D Slicer v4.11 (open-source software available at www.slicer.org) was used for segmentation of medical images.

Study overview. Firstly, three plausible impact scenarios (Scenario A, B and C) were outlined for further investigation. Secondly, a parametric study was performed using the subject-specific head model (ADAPT) until the fracture was predicted in agreement with the victim’s medical images. Third, the head kinematics and brain deformations generated by the impact was studied with whole-body models (THUMS) positioned with two different postures. The final impact scenario was reconstructed both with and without a worn helmet for a comparative analysis

Linear skull fractures, i.e. skull fractures that follow a line and are not splintered or depressed, have for long been hypothesized to emerge due to tensile stresses inherited from the outwards bending of the cranium at a distance away from the impact region [19]. Because of this outward bending, a linear fracture usually extends towards the impact point, as well as away from the impact point, towards areas that are structurally weaker, such as the skull base [20]. In experimental studies, researchers have also observed how microfractures in the skull bone’s inner cortical layer occur right below the impact point, before fracture is initiated away from the impact point and propagated towards it[17, 21]. With this fundamental knowledge about skull fracture mechanisms, an estimation was made regarding the impact location of the falling jack post. The impact location, highlighted in Fig. 2, was assumed to be in the region of the “interruption” of the fracture on the crown of the victim’s head.

Before impacting the victim’s head, the jack post was assumed to have fallen freely from upright position. The jack post’s free falling velocity was calculated using FE analysis. A meshed jack post, 5 m long and 34.8 kg heavy, was positioned upright with an initial incline of \(\:_\) = 3°. The jack post was allowed to fall freely with only gravity as load. Translational constraints were applied at the bottom node to create an axis of rotation. The resulting angular velocity of the jack post is included in Fig. 1. The angular velocity can also be derived analytically by applying conservation of energy. By expressing the potential energy at an arbitrary angle ? and relating it to the rotational kinetic energy, the following function can be derived:

$$\:\omega\:\left(\alpha\:\right)=\sqrt\left(\text_-\text\alpha\:\right)}$$

(1)

with L denoting the length of the falling beam. The analytical function is included in Fig. 1 and aligns well with the simulation results.

It is not known how far away the victim was standing from the falling jack post, as no measurements were taking from the scene. The translational impact velocity would consequently differ depending on the victim’s position relative to the jack post’s point of rotation. Three hypothetical impact scenarios, presented in Fig. 3, were evaluated in this study. Scenario A would translate as the most severe case, where the victim would be standing at a distance from the jack post so that the top of the post would hit the head with a resultant velocity of about 9.4 m/s. In Scenario B and C, the victim would be placed so that the resultant impact velocity would be approximately 7.1 and 4.6 m/s respectively. All three scenarios are associated with different impact rotational velocities ? and with different inclines ? of the jack post. Note that due to the design of the jack post (Fig. 1), the cross-section geometry of the impacting part of the jack post differ in Scenario A compared to Scenario B and C.

It is not known how the victim fell after impact and if any secondary head impacts occurred. All head injuries were assumed to be associated with the initial jack post impact, and potential ground impacts were not included in the analysis.

To reconstruct and investigate the victim’s skull fracture in-depth, the ADAPT model [22, 23] was used. The ADAPT model, visualized in Fig. 4, is an anatomically detailed FE head model that includes the brain and several of its constituents, as well as an enclosing cranium. The elements of the ADAPT model are between 0.2 and 0.5 mm in size, a resolution high enough to consider fracture propagation.

The ADAPT head model and its constituents

The human skull can be regarded as a three-layered structure, with two layers of compact cortical bone covering the outer and inner layer of the skull, and a more porous bone layer in-between referred to as trabecular bone or diploë. This bone structure, which is very significant for the mechanical behavior of the cranium, is included in the ADAPT model. The cranial model, defined with a strain-rate dependent material, has previously been shown to be successful in predicting skull fractures in fall accidents [17].

By using morphing techniques [13, 22], the ADAPT geometry can be individualized to match a subject of interest. The CT images of the victim’s head were used to reshape (morph) the ADAPT model to have the same morphology as the victim. Using this type of subject-specific models is highly important for fracture prediction, as the skull morphology and thickness has shown to be very influential for fracture propagation [17, 24].

The scalp thickness was measured to be 6 mm at the derived region of impact, observed in the CT scans. Thus, a 6 mm thick scalp (modeled with six layers of solid elements) was applied to the ADAPT model. The scalp was modeled with an exterior layer of dermis and an inner layer representing the two adipose layers of the scalp and the intermediated galea aponeurotica [17]. Since the ADAPT model does not include any facial structures, an extra weight of 0.6 kg as added to the skull base, leading to a total head mass of 4.0 kg, which is in better agreement with earlier reports on human head mass [25].

The jack post was meshed and then placed above the subject-specific ADAPT model with different impact angles ?, in all instances targeting the initial contact at the derived impact region. The impact angle, denoted ?, corresponds to different orientations of the victim relative to the jack post, see Fig. 3.

To study the head kinematics during and after impact, an HBM was used to reconstruct the event. The HBM used in this study was the Total HUman Model for Safety (THUMS) [26], an HBM distributed by Toyota. THUMS is aimed to represent an average sized male (178.6 cm, 77. kg) and is widely used within the field of impact biomechanics as well as the vehicle industry. The HBM is anatomically detailed with skeleton, organs, muscles and brain, however with a significantly larger element size (element size between 3 and 5 mm) in comparison to the ADAPT model.

Two postures that were brought up by witnesses were investigated. The two postures represent scenarios where the victim was standing upright and where the victim was positioned standing leaning forward, in a slightly hunched posture, see Fig. 3.

After the accident had been reconstructed using the HBM, the kinematics pulses (the angular velocity and acceleration of the HBM’s head center of gravity) were applied to the ADAPT model and the strains of the brain tissue were studied.

Once a plausible accident scenario was found, the same impact configuration was also simulated with a worn helmet. This way, a comparison of the injury outcome with and without wearing a helmet could be made. The helmet represents a construction helmet currently available in the Swedish market, equipped with an outer shell of ABS and an inner lining of EPS foam. It includes an adjustable strap for the head and chin, as well as a Mips low friction layer (LFL). The helmet model is validated against standard impact tests EN397 and EN12492s, involving drop impacts at different heights against the crown on the head, see Supplementary Appendix A for further details. The impacts were reconstructed with the worn helmet both with and without the LFL.

Based on the victim’s medical reports, the CT and MRI revealed the following types of head- and brain injuries: skull fracture, subdural hemorrhages (SDH), subarachnoid hemorrhages (SAH), contusions and edema. Several experimental studies using post mortem human subjects (PMHS) have been published over the decades, investigating appropriate metrics and thresholds of some of these injuries. However, human tolerance thresholds, meaning, the biomechanical limits of the human body beyond which injury is likely to occur, is a subject under debate. A selection of relevant experimental findings regarding human injury thresholds are presented in Table 1. Note that the listed thresholds are not definite and consensus on human tolerance levels has not been reached.

Skull fractures are categorized as a contact induced injury and they occur when the applied contact load reaches a magnitude that cause the skull to break. Thus, as seen in Table 1, skull fractures are associated with metrics in terms of forces and linear accelerations.

Brain hemorrhages are accumulated blood within the brain as a result of tearing of veins or arteries. In SDHs, the bleeding is found between the dura and arachnoid meninges. In SAHs, the bleeding is accumulated in the subarachnoid space. SDHs, which are one of the most lethal of head injuries, are believed to occur due to rapid and high rotational accelerations. The most common mechanism of SDH is ruptured veins that bridge the brain surface and intracranial subdural space [44, 45]. It has been thus been suggested that SDH is produced by short duration and high amplitudes of angular acceleration, see Table 1. To the authors knowledge, no thresholds specific for SAH have been reported.

Diffuse Axonal Injury (DAI) are characterized by wide-spread damage of the white matter of the cerebral hemispheres. DAI occurs due to rotational and/or rapid accelerations of the head causing damage to many of the brain’s axons. DAI is one of the most severe brain injuries, where a majority of patients with severe DAI never regain consciousness [46, 47]. There were no obvious signs of DAI according to the medical reports. Experimental findings relating to the human tolerance levels for DAI are presented in Table 1.

Brain contusions are generally found at the site of impact or the opposite side, so-called coup contrecoup injury. In general, contusions are believed to originate from straining of brain tissue that occur due to rapid head rotations [48, 49]. Brain contusions occur have been attributed to bleeding from the continuous flow of injured microvessels that ruptured during impact [50]. If a skull fracture is present, however, contusions are believed to occur due to the skull pressing against the underlying brain tissue [45, 48].