Environmental conditions driven method for automobile cabin pre-conditioning with multi-satisfaction objectives.

Abstract

The optimal initial pre-conditioning parameter is essential to properly adjust the temperature within the cabin in an effective and accurate way, especially while passengers' thermal comfort and energy-saving properties are both considered. Under the various environmental thermal loads, the pre-conditioning solutions resulting from those pre-fixed cooling parameters are unfeasible for achieving accurately passengers' comfort temperature. In addition, it is also difficult in such a narrow car space to identify a lot of local attributes due to the different material properties and sizes of a variety of structural parts that have various thermal responses to environmental conditions. This paper presents a data-driven decision model to numerically identify the degrees of the cabin thermal characteristic to determine satisfactory pre-conditioning parameter schemes. Initially, based on the thermal data within a vehicle recorded through the whole year at a selected hot climate region of the Middle East, the study levels multiple climate scenes corresponding to change in the cabin air temperature. Then three classification algorithms (Support Vector Machines, Decision Tree, and K-nearest neighbor model) are used to comparatively identify climate levels according to the input conditions. Based on the identified climate level, an appropriate parameters scheme for this level is applied. A comprehensive evaluation index (CEI) is proposed to characterize the passengers' satisfaction in numerical computation, on considering multi-satisfaction objectives including Predicted Mean Vote (PMV), local temperature, air quality, and energy efficiency; and it formulates the pre-conditioning parameter scheme for each climate scene with CEI. Several scene cases are carried out to verify the effectiveness of the proposed models. The result shows that the pre-conditioning schemes of the model can effectively satisfy passengers in multi-satisfaction objectives.