Multi-objective optimization of toughness and bonding strength in cement-based sealing materials using RSM coupled with NSGA-II hybrid approach: Performance evaluation and engineering application analysis

Aiming at the sealing failure caused by the insufficient toughness of traditional cement sealing materials and their poor bonding performance with coal seam interface in coal mine drilling gas extraction, this paper proposes a multi-objective collaborative optimization strategy of coupling response surface method (RSM) and non-dominated sorting genetic algorithm (NSGA-II) to achieve the synergistic improvement of the toughness and bonding performance of cement sealing materials. By constructing a performance prediction-optimization-verification system for modified sealing materials, the significance of the single performance prediction model of the toughness and bonding performance of sealing materials was analyzed, and the Pareto optimal solution set after conflict optimization of the toughness and bonding performance of cement sealing materials was obtained. And the optimal ratio of modified admixtures for sealing materials was determined: 0.52 % nano-silicon dioxide (NS), 1.92 % VAE latex powder (VAE), and 0.29 % carboxymethyl cellulose (CMC). At this time, the compression ratio of the drilling sealing material reached 0.35 and the bonding strength reached 1.39 MPa. When the axial pressure increases step by step, the cement permeability changes from 0.017 to 0.071, and the permeability change rate remains at 25 %. Finally, through engineering verification, it was found that after the application of modified sealing materials, the average borehole gas extraction concentration increased by 6.9 %, the average extraction mixed volume increased by 2.5 %, and the average gas extraction pure volume increased by 9.4 %, which has a reference significance for the research on coal mine gas extraction borehole sealing materials and the safe mining of mine resources. • The dynamic coupling of RSM and NSGA-II enables synergy between local modeling and global optimization. • A more efficient closed-loop system of "design-prediction-optimization-verification" was constructed. • The toughness and bonding performance of the sealing material are simultaneously enhanced to advance coal mine safety and green mining.