A Comprehensive Optimization Method for Centrifugal Impeller Based on Digital Twin and Reinforcement Learning

A technology of reinforcement learning and optimization methods, applied in machine learning, multi-objective optimization, design optimization/simulation, etc., can solve problems such as poor aerodynamic performance of blades, cumbersome design, simulation, manufacturing process, and inability to be considered in a unified manner, to ensure The effect of consistency

Active Publication Date: 2021-09-24
BEIHANG UNIV
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] At present, in the CI design process of centrifugal impellers, both aerodynamic performance and processability are taken into account. Most of the existing research results only partially reveal the influence of developable ruled graining on aerodynamic performance, and do not start from a comprehensive optimization strategy to find out the relationship between the two. The best balance point; secondly, in terms of improving the processing performance of the centrifugal impeller CI approximate straight graining, the existing research cannot accurately express the surface error of the centrifugal impeller before and after CI straightening after the straight graining of the impeller profile; now Some aerodynamic analysis and manufacturing performance evaluation are poorly coupled, separate analysis cannot be considered uniformly, and information data cannot be shared in time
At present, CI blade shape optimization of centrifugal impellers is mainly based on aerodynamic analysis, followed by processing performance analysis, resulting in poor aerodynamic performance of the actually processed blades, cumbersome design, simulation, and manufacturing processes, and poor economy

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A Comprehensive Optimization Method for Centrifugal Impeller Based on Digital Twin and Reinforcement Learning
  • A Comprehensive Optimization Method for Centrifugal Impeller Based on Digital Twin and Reinforcement Learning
  • A Comprehensive Optimization Method for Centrifugal Impeller Based on Digital Twin and Reinforcement Learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0073] A comprehensive optimization method for centrifugal impellers based on digital twins and reinforcement learning proposed by the present invention will be further described in detail below in conjunction with the accompanying drawings.

[0074]For a schematic diagram of applying the above method to an existing digital twin model, see the attached figure 1 , using the CI digital geometric model of the centrifugal impeller as the link to connect the modules of the digital twin, connect it with the processing performance optimization module and the aerodynamic performance evaluation module through data, and input the processing performance and aerodynamic performance indicators into the reinforcement learning. Learn to formulate optimization criteria to realize iterative optimization of centrifugal impeller CI blade shape. Based on the digital twin model, iteratively optimize the processing performance and aerodynamic performance of the centrifugal impeller CI with the help...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention relates to a comprehensive optimization method for centrifugal impellers based on digital twins and reinforcement learning, including step 1: establishing a digital geometric model of centrifugal impeller CI, and the ruled surface of the initial blade surface of centrifugal impeller CI; step 2: Machining performance evaluation, generating the tool position surface of the ruled surface in side milling and simulating the side milling process trajectory, analyzing the error of the centrifugal impeller CI before and after straight graining; Step 3: Aerodynamic performance evaluation , comparative analysis of the difference in the aerodynamic performance of the centrifugal impeller CI before and after straightening; Step 4: using the processing performance and aerodynamic performance as evaluation indicators, combined with reinforcement learning to develop an optimization method, the centrifugal impeller CI The ruled surface of the digital geometric model is optimized. Based on digital twin and reinforcement learning, the integrated design and optimization of centrifugal impeller CI digital geometric model, processing performance evaluation and aerodynamic performance evaluation are realized.

Description

technical field [0001] The invention relates to the field of optimization of centrifugal impellers for small gas turbine engines or superchargers, in particular to a comprehensive optimization method for centrifugal impellers based on digital twins and reinforcement learning. Background technique [0002] The centrifugal impeller (Centrifugal Impeller, referred to as CI) has become the core rotating part of a power system such as a small gas turbine engine or a supercharger due to its simple structure and high efficiency at a small flow rate. For centrifugal impeller CI, although performance parameters such as efficiency and pressure ratio are particularly important, from the perspective of application objects, manufacturing cost is more important for small power devices. The airfoil surface of the centrifugal impeller CI is mostly a free-form surface. Although the design freedom is large, it can only be processed by the relatively low-efficiency end milling method (the bott...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Patents(China)
IPC IPC(8): G06F30/27G06F30/17G06N20/00G06F111/06
CPCG06N20/00
Inventor 周煜丁水汀邢通刘晓静宋越
Owner BEIHANG UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products