Determination of vibration-related spinal loads by numerical simulation

Objective. Dynamic spinal loads due to human whole body vibrations are extremely difficult to determine experimentally. However, they can be predicted by numerical simulation. This paper presents an approach for the prediction of dynamic spinal loads caused by whole body vibrations, as well as some basic considerations concerning the process of numerical simulation. Background. Long-term whole body vibrations have been found to cause health risks for the lumbar spine. As an increasing percentage of the population is exposed to whole body vibrations at work, more and more people have to face the risk of whole body vibrations-related injury. Knowledge about the actual loads in the lumbar spine is essential when spinal loads are to be compared with spinal strength in order to assess the possible health risks caused by whole body vibrations. Methods. Since an extrapolation of results to unknown data such as spinal loads can only be done using anatomical models of the human body, a simplified finite-element model is presented which is adaptable to body height, body mass, and posture of any specific subject under investigation. The model has been built by reducing a very detailed, nonlinear finite-element model of seated man in its complexity (number of degrees of freedom). Furthermore, the simplified model has been linearised to avoid nonlinear solution procedures. Results. The model has been Verified for vertical and horizontal excitation at the seat. Model results have been compared to measurements on subjects. Individual exposure-effect relationships may be predicted by this model, due to the adaptability to a specific subject. Additionally, a new phenomenological method of eliminating the influence of local skin-accelerometer vibrations on vibration measurements on the skin surface is discussed. This method may provide data about bone acceleration that can be used in the process of model verification. Conclusions. Integral loading measures, such as spinal loads, may be predicted with simplified finite-element models. Quantitative judgements of these loads may be performed for individual conditions. Linearised models may be used for limited ranges of excitation intensities. Energy dissipation should be modeled by discrete dashpot elements instead of proportional damping.