Spent water-based drilling fluids generated during the construction of technological wells impose substantial environmental, water-management, transportation, and energy burdens. Conventional practices, including storage in temporary pits, prolonged settling, and off-site disposal, do not enable process-water recovery and require repeated handling of suspensions with a high solids content. This study evaluates a pressure-driven cylindrical hydrodynamic disperser as the central component of a compact on-site treatment system. Unlike conventional mechanical mixers, the disperser contains no driven shaft within the active chamber. Particle–reagent contact is intensified through controlled jet shear, vortex-induced redistribution, and the motion of freely moving steel balls. Field-derived drilling fluids containing 30–40 wt.% solids, with densities of 1.12–1.17 g/cm3, pH values of 7.4–8.2, and median particle sizes of 15–50 μm, were treated at velocity gradients of 500–1500 s−1 for 60–180 s using Superfloc N-300 dosages of 0–100 g/t. The optimal operating conditions were G = 1300 s−1, τ = 150 s, and D = 50 g/t. Under these conditions, the separation efficiency reached 91–93%, the residual suspended-solids concentration decreased to 120–130 mg/L, process-water recovery reached 80%, sludge volume decreased by 40–60%, and specific energy consumption was approximately 0.30 kWh/m3. More intensive treatment increased the separation efficiency to 94–95% but resulted in a less favorable balance among energy consumption, reagent dosage, and resource recovery. Compared with mechanical mixing, the selected treatment system reduced flocculant consumption by 37.5%, treatment time by more than threefold, and specific energy consumption by 40%. These results support the use of modular on-site systems for process-water recirculation and reduced sludge-transport requirements at remote drilling sites.
B. Mauletbekova, B. Kaliyev, Beibit Myrzakhmetov et al.· Applied Sciences· 0 citations
This review is, to the authors' knowledge, the first organised around the current generation of tabular foundation models, and taxonomises the architectures by pretraining regime, maps the capability space across five axes, isolates the language-model-on-tabular strand for prediction, feature engineering and generation, and summarises openness and deployment.
Sergei O. Kurashkin, V. Tynchenko, Alexey S. Borodulin et al.· Machine Learning and Knowled...· 0 citations