A Loop-Space Embedded Optimization Framework for Energy-Efficient Underground Mine Ventilation Networks

Underground mine ventilation is a safety-critical engineering system and one of the largest electricity consumers in mining, typically accounting for 40–60% of a mine’s electrical demand. Allocating airflow economically without violating safety limits is therefore central to energy-efficient ventilation regulation. This study proposes an optimization framework that couples loop-space embedding with a sine–cosine and Cauchy mutation-enhanced sparrow search algorithm (SCSSA). Loop-space embedding reparameterizes the branch airflow vector on the null space of the network incidence matrix, so that nodal mass conservation holds identically rather than approximately, and the decision dimension falls from the number of branches m to the number of independent loops L = m − n + 1. The remaining airflow, fan-operation and safety constraints are treated by a hierarchical exterior penalty function. Within SCSSA, refraction-based opposition learning improves the initial population, a sine–cosine strategy with adaptive inertia replaces the rigid producer step, and Cauchy mutation supplies the heavy-tailed perturbations needed to leave local optima. On ten classical benchmark functions, together with ablation and parameter-sensitivity studies, SCSSA attains better convergence accuracy and lower run-to-run variability than SSA, GWO and PSO. For a synthetic but engineering-scale 15-node, 21-branch ventilation network, the framework lowers fan shaft power from 32.604 kW to 14.784 kW relative to a 1.3-times over-ventilation design baseline, a scenario-based saving of 54.65% that at 8000 operating hours corresponds to about 1.43 × 105 kWh of electricity, about 8.55 × 104 CNY and roughly 87 t of CO2 per year, with both working-face airflow requirements satisfied. The framework thus offers a practical route to energy-efficient mine ventilation regulation in which mass conservation is structurally exact.