A cable defect data enhancement method and system based on physical consistency constraints

By constructing a PCDH-VAE generative model and combining a dynamic potential well mechanism with adaptive Hamiltonian Monte Carlo sampling, enhanced cable defect samples that satisfy multi-physics constraints are generated. This solves the problems of unreliable data and insufficient coupling under the condition of scarce samples, and improves the stability and accuracy of the defect diagnosis model.

CN121997053BActive Publication Date: 2026-06-09SHANDONG UNIV
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
SHANDONG UNIV
Filing Date
2026-04-10
Publication Date
2026-06-09

AI Technical Summary

Technical Problem

Existing technologies for cable defect identification suffer from problems such as unreliable physical data generated under conditions of scarce samples, insufficient maintenance of multi-physics coupling relationships, poor training stability, and disconnection from the actual physical state of the cable.

Method used

By constructing a PCDH-VAE generative model, introducing a conditional variational autoencoder network, embedding a dual-field potential function, constructing a Hamiltonian energy function, and employing an adaptive kinetic energy term, combined with a dynamic potential well mechanism and adaptive mass Hamiltonian Monte Carlo sampling, defect enhancement samples that satisfy multiphysics constraints are generated.

Benefits of technology

It achieves the generation of physically consistent, well-coupled multi-physics, stable and high-fidelity cable defect enhancement samples under the condition of scarce samples, thereby improving the performance of downstream defect diagnosis models.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121997053B_ABST
    Figure CN121997053B_ABST
Patent Text Reader

Abstract

The application belongs to the field of cable defect data enhancement, and relates to a cable defect data enhancement method and system based on physical consistency constraints. The system outputs balanced training sets and test sets through a data preprocessing module; a generative model construction module embeds double potential functions in a latent space based on a conditional variational auto-encoding network to encode observable vectors and conditional labels, constructs a Hamilton energy function that fuses double potential function constraints, and introduces an adaptive kinetic term to unify the PCDH-VAE generative model training target into a Hamilton energy minimization problem; a sample generation module introduces a dynamic potential well mechanism and uses adaptive mass Hamilton Monte Carlo sampling to generate defect enhancement samples that meet multi-physical field constraints through Hamilton dynamics evolution and Metropolis correction; and a sample evaluation module evaluates the generated defect enhancement sample set. The application can alleviate the problem of a lack of cable defect data samples.
Need to check novelty before this filing date? Find Prior Art