Optimization of a vacuum cleaner fan suction and shaft power using Kriging surrogate model and MIGA

Abstract

Optimization of vacuum cleaner fan components is a low-cost and time-saving solution to satisfy the increasing requirement for compact energy-efficient cleaners. In this study, surrogate-based optimization technique is used and for the first time it is focused on maximization of Airwatt parameter, which describes the fan suction power, as an objective function (Case II). Besides, the shaft power is minimized (Case I) as another optimization target in order to reduce the power consumption of the vacuum cleaner. 11 geometrical variables of 3 fan components including impeller, diffuser and return channel are selected as the optimization design variables. 80 training points are distributed in the sample space using Advanced Latin Hypercube Sampling (ALHS) technique and the outputs of sample points are calculated by means of CFD simulations. Kriging and RSA surrogate models have been fitted to the outputs of the sample space. Through coupling of constructed Kriging models and Multi-Island Genetic Algorithm (MIGA), the optimal design for each of the optimization cases is presented and evaluated using numerical simulations. A 20.22% reduction in shaft power in Case I and an improvement of 27.73% in Airwatt in Case II have been achieved as the overall results of this study. Despite achieving goals in both optimization cases, a slight decrease in Airwatt in Case I (−6.20%) and a slight increase in shaft power in Case II (+4.82%) are observed relative to primary fan. Furthermore, the Analysis of Variance (ANOVA) determines the importance level of design variables and their 2-way interactions on the objective functions. It was concluded that geometrical parameters related to all of the fan components must be considered simultaneously to conduct a comprehensive optimization. The reasons of enhancement in optimal cases compared with the reference design have been further investigated by analysis of the fan internal flow field. Post-processing of the CFD results demonstrates that the applied geometrical modifications cause a more uniform flow through the flow passages of the optimal fan components.

Aerodynamics Optimization of Multi-Blade Centrifugal Fan Based on Extreme Learning Machine Surrogate Model and Particle Swarm Optimization Algorithm

“…For example, Zhao et al [13] adapted the reverse design method to optimize the geometric configuration of the centrifugal fan hood using the back-propagation neural networks surrogate model. Soheil et al [14] used the Kriging surrogate model and MIGA algorithm to optimize the suction and shaft power of a vacuum cleaner centrifugal fan. Zhang et al [15] used the DOE study, the RSM surrogate model, and the multi-objective genetic algorithm to optimize the fan's 13 design parameters.…”

Centrifugal Pump Design: An Optimization

“…Figure 2a depicts an image that suggests the mesh around the blade, and Figure 2b depicts a close-up view of the blade's leading edge. The algorithm (Abdolahnejad et al, 2022;Almasi et al, 2022;Bamberger et al, 2020;Derakhshan et al, 2009)begins with a modest rotational speed and progressively increases it until it achieves its nominal speed of 3000 rpm. This is done for stability reasons.…”