Power-to-liquid (PtL) can serve as long-term energy storage and help maintain grid stability. In addition, the local generation of power promotes compact, small-scale, and container-based unit operations adjacent to the renewable energy source. So, it facilitates the proximity of energy sources to end users and improves energy supply security. This paper deals with the intensification of mini-scale Fischer-Tropsch synthesis and downstream upgrading units in a single divided-wall column for a PtL process. Due to the complexity of the simulated model, a surrogate model was used for optimization. Three methods of response surface methodology (RSM), least squares support vector machine, and multi-layer perceptron neural networks were used to find the best surrogate model. The results showed that RSM was the best surrogate model for the studied plant. The results of the analysis of variance sensitivity analysis showed that the most important design variable was the liquid split fraction over the wall. The proposed design was optimized for maximum variable income with a genetic algorithm. The optimal conversions of hydrogen and carbon monoxide were 82.09 % and 77.36 %, respectively. The proposed optimal plant required 3896.74 kW of energy, while the plant itself produced 4584.79 kW of energy. The optimization results showed that the optimal plant, which was fed with 85,120.08 standard cum/d of syngas, produced 569.08 kg/h of different hydrocarbon cuts, which was equivalent to 18.24 standard cum/d. This product, with its high volumetric energy density, has considerable potential for energy storage.

Keywords