Related Experiment Video
Updated: Aug 15, 2026
13:19
Deep Neural Networks for Image-Based Dietary Assessment
Published on: March 13, 2021
Performance Optimization Design of Stirred Tanks Based on Deep Neural Networks
Laifa Lu1,2, Jiali Wang1,2, Shuiqing Zhou1,2
1College of Mechanical Engineering, Zhejiang University of Technology, Hangzhou 310023, China.
ACS Omega
|August 14, 2026
Summary
This study introduces an intelligent design framework for efficient stirring impellers. The optimized blades enhance particle suspension by 16.05% while reducing energy consumption by 10.34%.
Area of Science:
- Chemical Engineering
- Fluid Dynamics
- Computational Science
Background:
- Stirring equipment is crucial in energy and chemical industries.
- Efficient mixing and energy-saving designs are key research focuses for industrial stirring operations.
Purpose of the Study:
- To develop an integrated optimization framework for designing energy-efficient stirring impellers.
- To enhance particle dispersion and reduce energy consumption in industrial mixing processes.
Main Methods:
- Utilized Class/Shape Function Transformation (CST) for low-dimensional blade parametrization.
- Employed Latin hypercube sampling to generate blade design space.
- Developed a deep neural network (multilayer perceptron) surrogate model for rapid performance prediction.
- Applied a multiobjective genetic algorithm for design optimization.
Main Results:
- Optimized blades reduce wake region energy dissipation and strengthen tip turbulence.
- Enhanced off-bottom particle starting and axial transport capabilities.
- Achieved a 16.05% increase in solid suspension height, meeting complete suspension criteria.
- Reduced impeller power consumption by 10.34% compared to the original blade.
Conclusions:
- The integrated CST and deep learning framework is feasible for energy-efficient stirring impeller design.
- This methodology offers intelligent optimization and energy-saving solutions for complex multiphase flow equipment.
