To address the challenges of low signal-to-noise ratio （SNR） and the lack of clean ground truth labels for distributed acoustic sensing （DAS） and distributed temperature sensing （DTS） data during hydraulic fracturing，a physics-informed self-supervised blind denoising method is proposed. Instead of relying on clean reference data，this method constructs self-supervised training samples using a checkerboard alternating sampling strategy on time-depth sub-blocks and enables the denoising network to learn noise distributions adaptively. Furthermore，physical constraints such as thermal conduction smoothness，edge preservation，and wavefield coherence are introduced for DTS and DAS characteristics，respectively. A spatially weighted loss function based on perforation cluster positions is also established to prioritize the preservation of key response features near clusters. It is demonstrated that the proposed method outperforms traditional filtering methods in terms of SNR and structural similarity （SSIM）. The denoised data significantly enhances the clarity of cooling fronts and acoustic energy bands at fluid entry points，which effectively improves the accuracy of fracturing event recognition and cluster efficiency evaluation.

To address the limitation of conventional smooth parallel-plate models，namely neglecting the influence of roughness on proppant transport，this study establishes a fracture geometry model based on fractal theory to embody actual reservoir roughness characteristics，where the Weierstrass-Mandelbrot function is modified using measured rock-wall profile data. The coupled computational fluid dynamics and discrete element method （CFD-DEM） is employed to simulate proppant migration and post-fracturing flowback，with a focus on the effects of the injection port number，proppant concentration，and particle size combination. The results indicate that in the case of single-port injection，fracturing fluid energy is concentrated to deliver a maximum equilibrium proppant bank height of 3.25 cm and the lowest flowback rate. When the proppant concentration increases from 8% to 11%，the proppant bank height rises from 2.60 cm to 3.28 cm，accompanied by a corresponding decrease in the flowback rate of proppants. For a particle size combination of 70/140 mesh，40/70 mesh，and 20/40 mesh at a mass ratio of 1：6：3，the equilibrium proppant bank height reaches 3.28 cm，and the proppant retention rate reaches 97.64%. It is revealed that rough-wall undulations induce localized vortex zones，increase the resistance to particle settling，and promote the formation of inter-particle force chains，thereby significantly enhancing proppant bank stability and near-wellbore placement efficiency. The proposed fractal-modified fracture model and the CFD-DEM coupling approach effectively capture the nonlinear effects of wall undulations on particle transport，providing theoretical support for optimizing proppant placement strategies in shale oil hydraulic fracturing.