An integrated CFD and machine learning analysis on pilots in-flight thermal comfort and productivity

Abstract

In the cockpit, minimising the pilots' thermal discomfort due to solar exposure while ensuring clear visibility is essential. This study numerically examined the solar radiation effect on thermal fields of an Airbus A320 cockpit and assessed global and local thermal comfort of three aircrews using the Fanger model and the Equivalent Homogeneous Temperature (EHT) model. The productivity of aircrews based on thermal comfort was further estimated. A total of 27 cases, encompassing a combination of environmental and HVAC parameters were performed. This dataset was used for data analysis, employing an explainable machine, XGBoost, to highlight the significance of each variable on the pilot's thermal satisfaction. Outcomes proved that solar radiation penetrating the cockpit domain facilitated a non-homogeneous temperature distribution and thermal stratification. With strong solar exposure, it was recommended to set cockpit's maximum supply temperature smaller than 19[Formula: see text]C, maintaining RH between 10–20% for appropriate global thermal environment. Aircrew productivity loss was estimated to decline by 2.7% to 5.6% due to solar intensity variation, with pilot being the most affected. Without solar radiation, the local comfort of lower body segments was compromised. Rising solar radiation resulted in significant EHT gradient changes in the upper torso compared to the lower body, highlighting the predominance of radiative heat transfer at the upper level. Analysis via XGBoost revealed that mean radiant temperature had the most substantial impact on thermal satisfaction in the cockpit, underscoring the efficiency of optimising cockpit materials for better absorption and reflection of radiative heat for enhanced thermal environment control.

Keywords

• Commercial airliner cockpit
• thermal comfort
• solar radiation
• machine learning
• pilots' productivity
• XGBoost