Part-load performance prediction of a novel diluted ammonia-fueled solid oxide fuel cell and engine combined system with hydrogen regeneration via data-driven model

This paper proposes a novel diluted ammonia-fueled solid oxide fuel cell and X-type rotary engine (SOFC-XRE) combined system with hydrogen regeneration sub-system to realize high energy conversion efficiency. The combined system key design parameters and geometric parameters are optimized. The influences of SOFC inlet temperature, stack fuel utilization ratio, stack excess air ratio and pressure ratio across membrane for permeating hydrogen on system off-design performance are investigated. Meanwhile, the improved off-design performance of combined system is revealed by comparing with standalone SOFC. The data-driven neural network model (NNM) is established for predicting system part-load performance based on the off-design performance datasets. The optimal system operation parameter combinations and performances in a wide power load range are ob-tained by combining data-driven NNM with particle swarm optimization (PSO) algorithm. The results show that the cell stack temperature gradient decreases with decreasing fuel flow rate coefficient. The stack fuel utilization ratio has a larger effect on combined system energy efficiency than the fuel flow rate coefficient. The optimal membrane pressure ratio and fuel flow rate coefficient decrease with decreasing power load. The combined system energy efficiency is increased by 5.94% as the power load is decreased by 75%, while it is decreased by 2.1% as the power load is increased by 20%. The combined system energy efficiency is 10.9-12.2% higher than standalone SOFC in a wide combined system power load range of 120-25% without carbon emission.