Microfabricated fuel cell systems have the potential to outperform batteries for man-portable power generation. Because many electronic devices operate at various loads, with frequent start-ups and shut-downs, transient aspects are highly important and must be considered thoroughly. In this paper, the focus is on the optimal start-up of microfabricated fuel cell systems using numerical open-loop optimal control. For start-up purposes, a small rechargeable battery is used to provide the energy needed to heat up the fuel cell stack and meet the power demand when the fuel cell is unavailable or can only satisfy part of the demand. The objective of the start-up problem is to bring the system to a desired operating point with a minimal total mass of the system (battery and fuels), while meeting the nominal power demand at any time and satisfying the operational restrictions. The model for the fuel cell stack consists of partial differential-algebraic equations with multiple time scales and numerical techniques that exploit a separation of these time scales are used for efficient and reliable integration of the state and sensitivity equations. A case study of a microfabricated power generation system employing a high-temperature solid-oxide fuel cell and using ammonia and butane as fuels is presented.These micro processes have the potential to yield much higher energy densities than state-of-the-art batteries, because the abovementioned fuels have very high energy contents, and fuel cells can in principle achieve very high efficiencies.In previous work, a methodology for the design of microfabricated fuel cell systems for portable power generation has been developed [3]. This methodology is based on a decomposition into three levels of modeling: at the system level, lumped algebraic models are used to address the process synthesis, heat integration and layout considerations simultaneously [4,5]; at the intermediate-fidelity level, spatially distributed models are formulated to optimize unit sizes and operation for a given process structure, without the need to specify a detailed geometry [6, 7]; and at the detailed level [8], CFD models are used to facilitate geometrical improvements as well as to derive and validate modeling assumptions that are used in the system-level and intermediate-fidelity models. So far, mainly steady-state aspects have been studied through this methodology.As most power-consuming devices are operated periodically and have rapidly changing power demands, the dynamics and automated operation of portable power production are very important and must be considered thoroughly. In this paper, the focus is on optimization of the start-up of micro power generation processes. For start-up operation, it is most likely that the devices will be coupled with a relatively small, rechargeable battery, whose role will be to ensure that the power demand is met when the fuel cell is unavailable or can only satisfy part of the demand, as well as to provide the energy needed to heat the fuel cell stack up to a tem...