Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

High-dimensional optimization and selection method of impeller for multi-layer perceptual structure

A multi-layer perception and impeller technology, applied in the field of big data learning model, can solve the problem that the impeller of rotating machinery cannot obtain the optimal solution accurately and efficiently, so as to shorten the design cycle and ensure the design accuracy

Active Publication Date: 2021-11-05
ZHEJIANG UNIV
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to solve the problem that the optimal solution cannot be obtained accurately and efficiently during the optimization of the rotating machinery impeller in the prior art, and to provide a high-dimensional optimization and type selection method for the impeller facing the multi-layer perception structure

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
  • High-dimensional optimization and selection method of impeller for multi-layer perceptual structure
  • High-dimensional optimization and selection method of impeller for multi-layer perceptual structure
  • High-dimensional optimization and selection method of impeller for multi-layer perceptual structure

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0089] In this example, the optimally designed original rotating machine is a low specific speed centrifugal pump with a single-stage cantilever structure. The impeller of the pump has 5 blades, and the design speed is n =1450rpm. Key performance parameters include design flow Q d =100m3 / h, lift H d =50m, specific speed N S =46.9. NPSH ( NPSH) The rated value is recorded as the rated NPSH ( NPSH r), in this example NPSH r is 3m. figure 2 and Table 1 give the geometry of the LSSCP and the main profile parameters of the impeller, respectively. The calculation domain adopts a non-simplified model considering the flow field between the volute and the impeller gap to improve the calculation accuracy.

[0090] Table 1 Main profile parameters of the impeller

[0091] Model Line Parameters value impeller inlet diameter D 1 (mm)

[0092] The range of the main profile parameters is determined according to the fluid characteristics, design specific...

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 discloses a high-dimensional optimization and type selection method of an impeller facing a multi-layer perceptual structure, which belongs to the field of big data learning models. The invention firstly reduces the dimensionality of design variables through sensitivity analysis, and then uses the distributed computing capability of the cloud platform to obtain a large number of simulation samples based on coupled computational fluid dynamics (CFD) to train the multi-layer perceptron (MLP) and obtain big data learning model. Then, based on the big data learning model, continue to use the non-dominated sorting genetic algorithm (NSGA-Ⅲ) on the cloud platform to perform cloud computing and high-dimensional multi-objective global optimization of the design variables, and then use the adjoint method to continue to optimize the overall impeller type for the global optimization results. The line is improved to realize the high-dimensional multi-objective optimization of the performance of rotating machinery. The invention organically combines cloud computing and big data learning models, can quickly and accurately converge to the global optimal solution of the high-dimensional optimization target, and at the same time ensures the design accuracy and greatly shortens the design cycle of the rotating machinery.

Description

technical field [0001] The invention belongs to the field of big data learning models, and in particular relates to a high-dimensional optimization and type selection method of an impeller facing a multi-layer perceptual structure. Background technique [0002] Rotating machinery converts rotational kinetic energy into fluid power to transport fluid. Optimizing the efficiency, output, and cavitation resistance of rotating machinery is of great significance to energy saving. Rotating machinery generates energy through the rotation of the impeller, so the impeller has a great influence on the performance of the rotating machinery. The key to optimizing the performance of the rotating machinery is to improve the profile of the impeller. [0003] Traditional design optimization methods mostly use the combination of experience and simulation. The shape of rotating machinery is complex and there are many design variables. Therefore, the calculation period is often long and the amo...

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): G06F30/17G06F30/28G06F30/23G06N3/04G06F119/14
CPCG06F30/17G06F30/28G06F30/23G06F2119/14G06N3/045
Inventor 童哲铭辛佳格童水光
Owner ZHEJIANG UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products