Insulation resistance calculation method based on neural network and electronic device

Abstract

The invention discloses an insulation resistance value calculation method based on a neural network, and the method comprises the following steps: (1) building a neural network model corresponding toa current insulation detection circuit, obtaining a database corresponding to the current insulation detection circuit, and enabling the database to have a plurality of groups of insulation detectiondata and corresponding insulation resistance values; (2) using a BP neural network error back propagation algorithm to import the data into a database, using the insulation detection data as an inputelement, using the insulation resistance value as an output value, and training a neural network model according to the data in the database; (3) acquiring insulation detection data of the current insulation detection circuit; and (4) taking the insulation detection data as an input element, and outputting an insulation resistance value by using the trained neural network model. The correspondingneural network model is trained for the specific insulation detection circuit to calculate the insulation resistance value, the insulation resistance value is high in anti-interference performance, dependence on signal acquisition precision is low, the detection result precision is high, and the calculation speed is high.

Description

technical field [0001] The invention relates to the field of power electronics, in particular to the insulation detection of power supply systems in application places such as electric vehicles and energy storage. Background technique [0002] In the field of power electronics, including application scenarios such as electric vehicles and energy storage, there are various power systems, especially for high-voltage systems, in order to ensure insulation safety, it is necessary to monitor the insulation leakage of the system in real time, so insulation detection is an extremely important task. important technology. [0003] Regarding the estimation of insulation resistance, the mainstream technical solutions currently used in the industry are: (1) Balanced bridge method: the balanced bridge is to compare the resistance to be measured with the standard resistance, and measure the resistance to be measured by adjusting the balance of the bridge Value, such as single-arm DC brid...

AI Technical Summary

Problems solved by technology

[0009] Compared with the prior art, the present invention can train the corresponding neural network model for the specific insulation detection circuit, and calculate the insulation resistance value through the neural network, which has strong fuzzy recognition characteristics and will not be affected by the parameters in the specific insulation detection circuit. Calculation results are affected by subtle influences, low dependence on signal acquisition accuracy, highly dependent on interference, and there is no fixed insulation resistance calculation formula, which also avoids applying calculation formulas due to insufficient signal acquisition accuracy or noise interference Calculation divergence problem occurs, the calculation result is highly accurate and the calculation speed is fast

Images