High-precision motor multi-physics field coupling simulation calculation method

Abstract

The invention discloses a high-precision motor multi-physics field coupling simulation calculation method. The method comprises the steps of 1, establishing a deep learning model based on a Keras framework, and training existing ferromagnetic loss data; evaluating the model by adopting an out-of-order repeated K-fold verification method; expanding the training sample by establishing two deep learning models, and taking the expanded sample as the training sample of the other deep learning model; 2, taking electromagnetic loss as an input source for temperature field calculation; and 3, calculating and simulating a thermal stress borne by the motor due to temperature rise through multi-physics joint simulation of an electromagnetic-temperature-stress field. According to the multi-physics field coupling simulation calculation method, the electromagnetic loss can be accurately calculated, an accurate input source is provided for subsequent multi-physics field joint simulation, and the state of the motor in actual operation is accurately simulated.

Description

technical field [0001] The invention relates to the technical field of coupling field simulation analysis, in particular to a high-precision multi-physics field coupling simulation calculation method of a motor. Background technique [0002] The motor is usually composed of stator, rotor, air gap and other parts. The three-phase symmetrical rotating magnetic potential generated by the stator coil cuts the rotor coil to realize the conversion of electrical energy and mechanical energy. In the process of energy conversion, due to the non-ideality of the stator and rotor iron core and coil materials, iron loss and copper loss will be generated during the operation of the motor, and the electromagnetic loss will cause the temperature rise of the motor, and the temperature rise will affect the silicon steel sheet. Changes in magnetic permeability and coil resistivity; further lead to changes in motor loss; at the same time, changes in temperature will change the thermal expansion...

AI Technical Summary

Problems solved by technology

This method does not effectively simulate the actual operation of the motor: it does not consider the relationship between the magnetic permeability of the silicon steel sheet and the conductivity of the copper wire in the actual situation, and during the actual operation of the motor, the iron loss and copper loss of the motor will be The temperature of the motor will increase, and the increase of the motor temperature will lead to the change of the material performance, that is, the change of the motor loss. On the other hand, in the traditional electromagnetic calculation, the Bertotti iron loss discrete calculation model is used for the motor loss, in which the iron loss calculation Coefficients are often estimated empirically, which also leads to inaccurate calculations of electromagnetic losses

[0005] Publication No. CN103400010A proposes a method for calculating the temperature rise of permanent magnet synchronous motors based on multi-physics field coupling. This patent conducts finite element electromagnetic modeling of the motor and imports the electromagnetic analysis results into the temperature field for fluid-solid coupling analysis. , this method considers the actual operation of the motor in terms of heat dissipation in more detail, and simulates the heat dissipation of the motor through fluid-solid coupling analysis, but this method does not consider the influence of temperature changes on different materials of the motor during real-time operation

Using this method to analyze the heat dissipation of the motor has the following problems: (1) In the electromagnetic analysis, the modeling of the iron loss model is inaccurate, and the calculation error of the iron loss coefficient is relatively large; (2) the silicon steel sheet, copper The magnetic permeability and electrical conductivity of conductor materials change with temperature, which further leads to inaccurate calculation of motor loss and large calculation errors of motor temperature rise distribution

Images