The scaling of runaway electrons (RE) generation with pre-disruption plasma current IP was tested in the JET tokamak in a set of Ohmic pulses with $I_P$ from 2.5 to 3.2 MA, where disruptions were induced by pure neon shattered pellet injection (SPI) [1]. The surprising outcome of these experiments was that RE generation did not scale with $I_P$, despite the strong increase of the RE multiplication factor with plasma current expected from the avalanche mechanism [2,3]. This suggests that other effects might have played a significant role, such as differences in primary RE generation and/or transport losses, despite the similar target plasmas. To support the interpretation of experimental findings, modelling activity is ongoing with both the DREAM and JOREK codes. In this work, we present a preliminary analysis with the 3D MHD code JOREK of the thermal quench (TQ) disruption phase in the target pulses. The aim of this study is to assess the effect of the SPI on the MHD dynamics that induce the TQ and leads to the following current quench (CQ) phase, where REs are observed. Similarly to previous mixed $Ne/D_2$ SPI validation studies on JET [4], the reduced MHD version of JOREK is employed, with a neutral gas shielding model for the ablation of SPI fragments, and a collisional-radiative model for neon radiation. Mainly due to the relatively low electron density of the target pulses (selected to promote RE generation), the assimilation of injected material is rather low, with only a few percent of the total neon atoms in the pellet being ablated in the TQ phase. While the resulting radiation cooling is still sufficient to produce the stochastization of the plasma and trigger the core temperature collapse, it is also low enough for the ongoing numerical simulations to proceed deep in the TQ phase and potentially to reach the $I_P$ spike and the following CQ phase. Reproducing the $I_P$ spike in the simulations would be important because of the related current profile relaxation, which is expected to affect RE generation [5]. Possible strategies for including RE physics in the present modelling will be also discussed. These involve using a kinetic tool for hot electrons to investigate primary RE generation at the TQ via the hot tail mechanism [6], and/or moving to the CQ phase to study secondary RE generation via the avalanche effect [7]. [1] U. Sheikh et al., in preparation. [2] M.N. Rosenbluth and S.V. Putvinski, 1997 Nucl. Fusion **37** 1355 [3] B.N. Breizman et al., 2019 Nucl. Fusion **59** 083001 [4] D. Bonfiglio et al., 2026 Nucl. Fusion **66** 026014 [5] I. Pusztai et al., 2022 J. Plasma Phys. **88** 905880409 [6] L. Puel et al., 2026 Nucl. Fusion **66** 026019 [7] C. Wang et al., 2025 Nucl. Fusion **65** 016012