The impacts of evaluation indices and normalization methods on E-TOPSIS optimization of return vent height for an impinging jet ventilation system

Stratified air distribution (STRAD) systems have been intensively investigated in recent decades for their energy-saving potential and good indoor air quality performance. However, the evaluation indices used to optimize STRAD systems and the normalization methods for weight calculation vary from one research to another. This study aims to investigate the impacts of evaluation indices on the optimal return vent height of a room cooled by an impinging jet ventilation system (one type of STRAD system). The effects of several widely used normalization methods (i.e., vector normalization, sum normalization, min-max normalization, and no normalization) on indices weights are investigated. The evaluation indices are cooling coil load ( Q coil ), energy-saving potential (Δ Q coil ), mean age of air (MAA), CO 2 mass fraction, temperature difference between the head and ankles (Δ T 0.1–1.1 ), predicted mean vote (PMV), predicted percentage of dissatisfied (PPD), and draft rate (DR). The multi-criteria optimization method is the entropy-based technique for order preference by similarity to ideal solution (E-TOPSIS). As a result, the min-max normalization method evens the weight of each index and results in unreasonable relative weights. Consequently, the raw matrix (i.e., the normalization is omitted) is suggested for weight calculation. Among these indices, Δ T 0.1–1.1 and PPD play critical roles. Without Δ T 0.1–1.1 , the optimal return vent height changes from mid-level to near-floor, while without PPD, it changes to near-ceiling. Another important result is that the Q coil plays the most trivial role, followed by MAA and DR. Therefore, the optimal return vent height is not determined by energy-saving performance but by performances of thermal comfort.