Lasers used in industrial environment do not only have to be powerful, reliable and economic, but are also expected to work properly and within the specified range over hundreds of thousands of hours. Regarding a short time-to-market, the time-consuming lifetime testing of early prototypes and components is a major drawback in the development of new products. In this article we demonstrate a way to reduce the testing effort dramatically by use of simulations.Over the past decade, fiber lasers have become increasingly important in a number of industrial applications like sheet metal cutting and welding. The reason for this is a maturation of the fiber laser technology that -besides other benefits -also sets record after record in terms of power.A simplified scheme of the fiber laser principle is depicted in Fig. 1. As in every laser, a gain medium and two mirrors are the basic ingredients, which are all realized in a monolithic all-in-fiber design in this case. The active medium is an ytterbium-doped fiber pumped by high-power laser diodes, while fiber Bragg gratings (FBGs) directly written into the fiber serve as cavity mirrors. The fiber is coiled in order to suppress higher order fiber modes that could deteriorate the beam quality.The remarkable power scaling of high power fiber lasers over the past years has become more and more challenging, because the effects of transverse mode instability (TMI) and photodarkening (PD) of the fiber can limit the output power of industrial fiber lasers. The origin of the TMI is a power transfer from the fundamental mode of the fiber to higher transverse modes via self-induced long-period gratings in the fiber due to the thermo-optical effect [1]. The excitation of higher modes can lead to temporal instability of the laser power, and to an increased power transfer from the core to the cladding via bend losses, thus reducing the effective laser power. Over the lifetime of a fiber laser the TMI threshold is decreasing due to photodarkening of the active fiber core [2]. This can lead to a degradation of the laser performance over several thousand hours of operation.Because it can take very long to observe the unwanted effects in the laboratory, the goal was to develop a reliable way to predict the performance of a certain laser setup including fiber design, pump and resonator configuration, and fiber routing, prior to the actual prototype build. Testing of the final design is not replaced by the simulation, but it enables to compare variants in order to avoid unnecessary design iterations.
Simulation approachThe main challenge of the modeling work was the huge range of time and length scales involved. Spatial dimensions range from the fiber core diameter in the order of microns to fiber lengths of some 10 meters, and the temporal scales range from the sub-microsecond optical round trip time to the aging over thousands of hours.A full 3D (or moreover time-resolved) model capturing the longitudinal variations along the fiber as well as the transverse effects with a sufficient reso...