Pre-sizing of cooling systems for hybridelectric aircraft. Knowledge based design tool development

Abstract

Hybrid-electric and electric aircraft offer both a solution to the decarbonization of the air-travel sector and a significant challenge for engineers to make them feasible. Component requirements for such aircraft often do not align with what is currently available on the market, necessitating rapid advances in technologies and numerous design iterations.

The cooling-system pre-sizing tool developed during this work aims to serve as the backbone of a Thermal Management System design routine that can be integrated into a larger Multidisciplinary Design Optimization (MDO) network. Besides sizing, it provides estimations of heat transfer, cooling drag, weight, and coolant parameters for a ram-air liquid-cooled system based on component models of various fidelity.

The core of the tool consists of a one-dimensional flow solver and an integral compact heat-exchanger model that analyses user-specified boundary parameters, coupled with three design-optimization algorithms: parametric design-space sweep, genetic algorithm, and SLSQP. In a typical run the routine evaluates about 1 000 designs per 25 s, while the integral heat-exchanger model exhibits a mean absolute percentage error of 1.5 – 7.4 % depending on the flow regime, empowering conceptual design.

Furthermore, CFD-based surrogate models that account for the radiator’s influence on the flow inside the preceding diffuser were developed and integrated into the design routine as an approach to increase the fidelity of pre-sizing without continuous CFD.

Finally, the tool was applied to a realistic example of the type-certified aircraft, Pipistrel Velis Electro, to compare its current design with the design from the presizing routine.