Design-Oriented Modeling of Active Fibers for High-Power Laser Applications

Fiber lasers have become the leading optical source in high-power industrial and medical applications, thanks to their unique combination of efficiency, robustness, and high beam quality. These systems rely on active optical fibers doped with rare-earth ions and pumped by semiconductor laser diodes. For their optimal design and operation, accurate numerical models are essential. However, the predictive capability of such models critically depends on the knowledge of fiber parameters, such as emission and absorption cross sections or rare-earth dopant concentration, which vary significantly with fabrication processes and structural design and are rarely disclosed by manufacturers. This lack of reliable parameter data makes calibration procedures indispensable to ensure meaningful simulations and guide optimization efforts. This thesis addresses the development of a systematic procedure for recovering the key parameters of active fibers to improve the design of high-power fiber lasers. The proposed methodology is based on fitting experimental results to the predictions of an accurate physical model, with the parameter retrieval carried out using advanced optimization algorithms, in particular particle swarm optimization. To this end, a general simulation framework was implemented in MATLAB, designed to be applicable to virtually any rare-earth dopant. Beyond parameter fitting, the model was employed to investigate the sensitivity of laser performance to variations in simulation parameters, highlighting the most critical quantities on which optimization efforts should concentrate. The approach was validated through an experimental study on a high-power ytterbium-doped double-cladding fiber. Simulations showed good agreement with experimental measurements, confirming the accuracy, robustness, and practical relevance of the proposed calibration procedure. Overall, this work demonstrates that, starting from a limited set of experimental data, it is possible to extract reliable parameters for use in simulation and optimization of fiber lasers. The developed methodology thus represents an effective tool for the design and analysis of rare-earth-doped fiber lasers, with particular importance for applications requiring efficient, stable, and scalable high-power light sources.