Turbomachinery variable working condition performance prediction method based on flow field reconstruction

Abstract

The invention discloses a turbomachinery variable working condition performance prediction method based on flow field reconstruction. The turbomachinery variable working condition performance prediction method comprises: determining turbomachinery geometrical parameters, working fluid and variable working condition performance prediction input parameters; performing CFD calculation on the sample points to obtain initial flow field data; preprocessing the CFD result to obtain real flow field data and performance data; constructing a flow field reconstruction network, and training a deconvolution neural network by using a flow field reconstruction loss function; constructing a performance prediction network, and training a convolutional neural network by using a performance prediction loss function; and predicting the variable working condition performance of the turbomachinery based on the trained flow field reconstruction network and the performance prediction network. According to themethod, the deep convolutional neural network is utilized to realize direct conversion between variable working condition parameters of the turbomachinery and flow field information and performance,and the method has the advantages of accuracy, quickness, universality, flexibility, easiness in implementation, no need of human intervention and the like, and has a good application prospect in theaspects of real-time control and design optimization of the turbomachinery.

Description

technical field [0001] The invention belongs to the field of energy and power, and in particular relates to a method for predicting the performance of turbomachinery under variable working conditions based on flow field reconstruction. Background technique [0002] Turbomachinery is a mechanical device that converts the energy in the fluid working medium into mechanical energy. It is the core component of the power cycle. Because of its complex structure and harsh working environment, its development and production level is an important indicator of a country's scientific and technological strength. Its performance will directly affect the power and efficiency of the circulation system, so it is an important research direction in the industrial field to carry out design optimization research on turbomachinery. [0003] In actual operation, the operating state of turbomachinery is affected by many uncertain parameters such as motor speed, incoming flow parameters, power, etc....

AI Technical Summary

Problems solved by technology

The traditional CFD (Computational Fluid Dynamics) solution based on the physical model is often used to predict the performance of turbines under variable operating conditions. Although CFD calculations can predict turbine performance more accurately at present, it requires a large number of calculation operating points, which greatly increases the Calculation cost and time consumption

On the one hand, this will lead to a significant increase in the turbine design cycle; on the other hand, in the actual operation and control of the system, it is difficult to grasp the off-design performance of the turbine unit in real time, and the system cannot be adjusted in time

[0004] In summary, using the traditional CFD analysis method to predict the performance of turbomachinery under variable operating conditions requires a large workload and a long design cycle

Images