The presented study is an extended computational investigation of a developed model describing thermally driven processes in anisotropic wood particle. The model is an in-house implementation written in Fortran, formulated as a transient two-dimensional approach that incorporates conservation equations for mass, energy, pressure, and gas velocity within the particle. The analysis possibilities is significantly enhanced through parametric simulations enabled by cloud computing infrastructure. The research uses container based computing on Amazon Web Services. The developed container allows the simulations to be run efficiently, consistently, and to switch across many computing resources. The model incorporates two kinetic parameters, including activation energy (Ea) and the pre-exponential factor (A), and anisotropic thermal conductivity (λ), which reflects the directional heat transfer properties of wood. A parametric analysis was conducted, including 9200 simulation cases generated from 20 values of A, 20 values of Ea, and 23 values of λ. The thermal conductivity λ is defined to preserve anisotropic characteristics of wood structure. The extensive computations enabled systematic exploration of parameter interactions and sensitivity. Such resolution of parametric study would not be achievable with conventional computational approaches. The results clearly indicate that heat transfer governs the pyrolysis process, with the system showing high sensitivity to variations in thermal conductivity. This demonstrates that the rate of heat penetration into the material is the key factor controlling the overall reaction dynamics. Changes in λ strongly affect temperature evolution and, consequently, the onset and progression of pyrolysis. Only after reaching a specific temperature, kinetic parameters influence the system behavior, with activation energy (Ea) determining reaction initiation thresholds and the pre-exponential factor (A) affecting reaction intensity and temporal evolution. The use of container-based cloud computing allowed efficient execution and management of the large simulation set, reducing computational time and enabling robust data generation. The findings confirm that integrating scalable cloud infrastructure with in-house models provides a powerful approach for comprehensive sensitivity analysis and improves the predictive capability of the models.