An efficient design method of indoor ventilation parameters for high-speed trains using improved proper orthogonal decomposition reconstruction. (15th July 2023)

• Record Type: Journal Article
• Authors: Lu, Yibin; Wang, Tiantian; Zhao, Changlong; Zhu, Yan; Jia, Xiaoping; Zhang, Lei; Shi, Fangcheng; Jiang, Chen
• Is Part Of: Journal of building engineering. Volume 71(2023)
• Publication Date: 2023-07-15
• Subjects: Thermal comfort -- High-speed train -- Ventilation parameter design -- POD -- Orthogonal experiment; Building -- Periodicals; 690.05
• DOI: 10.1016/j.jobe.2023.106600
• Languages: English
• ISSNs: 2352-7102

Abstract

Abstract: The ventilation parameters inside a compartment of a high-speed train (HST) are of great significance to the thermal comfort and air quality around passengers. Additionally, parameter design relying on traditional numerical simulations takes considerable time and resources. This study proposed an improved inverse design approach based on proper orthogonal decomposition (POD) to the design of supply air parameters of a compartment for an HST; this method was combined with the radial basis function (RBF) model and the orthogonal experimental design (OED) to facilitate the process. The feasibility and accuracy of this improved method applied to ventilation design inside the HST were validated by experimentally verified numerical simulations. Three design parameters (i.e., the temperature of the inlets, the ratio of the upper air supply, and the angle of the upper inlets) and evaluation indexes (i.e., predicted mean vote, draught rate and air age) were taken into consideration, and the sensitivity of the design parameters to the evaluation indexes was analysed by orthogonal cases using range analysis. The proposed POD reconstruction method has a remarkable advantage over traditional numerical simulation methods with respect to the time and efficiency of design and can take approximately 40% less time than the current POD method to obtain numerical simulation samples. Moreover, the reconstructed results by the RBF model were more consistent with the actual simulation data, and the average determination coefficients ( R 2 ) of all test cases for all evaluation indexes are more than 0.97. Highlights: The improved POD method that combines the radial basis function model and orthogonal experiment is proposed. The feasibility of the proposed method in ventilation design inside the train is well validated. The sensitivity of design parameters on the evaluation indexes is analysed. The time for obtaining CFD data drops by approximately 40% compared to the existing POD method.