Simulation and Experimental Study of Partial Discharges Considering Frequency-Dependent Properties of Cable Lines with XLPE Insulation

Analysis of partial discharges (PDs) in solid dielectric systems improves the quality of cable line insulation diagnostics by identifying installation defects, aging, and other destructive factors. However, known ABC models and single-line equivalent circuits provide an approximate description of discharges, and the distribution of amplitude-phase (APD) and amplitude-frequency (AFC) characteristics of PDs does not sufficiently agree with experiments reflecting the stochastic nature of PD-related voltage pulses recorded at the sensor output. This paper proposes a block-hierarchical approach to PD modeling in a three-phase power cable line (PCL) in LTspice. The model is based on a multilevel equivalent circuit with distributed parameters (MRLCG), supplemented by an ABC defect model. The hybrid MRLCG+ABC model takes into account the mutual influence of conductors and the sheath, the sheath grounding scheme, laying geometry, wave attenuation, distributed defects, conductor-to-conductor and conductor-to-sheath mutual inductance, the proximity effect, and the skin effect. A stochastic set of PDs is obtained using the Monte Carlo method, which reproduces the nature of the discharge process and provides stable statistical estimates. The novelty lies in the comprehensive consideration of electromagnetic and topological factors within a unified hierarchical scheme. Verification on a 10 kV cable showed convergence between numerical and physical APD/AFC according to statistical criteria: a Pearson correlation coefficient of r ≥ 0.93, an average amplitude discrepancy of ≤ 15% (over 1000 cycles), and a peak-to-peak amplitude discrepancy of single pulses of ΔU = 2–3%. This improves diagnostic quality and provides a basis for future studies on predicting the residual life of insulation.