This paper presents two algorithms for the estimation of the nodal admittance matrix (Y ) on dc microgrids. The first methodology is a gradient-based identification algorithm whereas the second is a projection-based algorithm which includes additional constraints related to the expected structure of the matrix. These constraints are defined in a matrix space R and include symmetry and sparsity. The objective function is a matrix-valued convex function, thus the convergence of the algorithm is guaranteed via convex optimization theory. Simulations results demonstrate the feasibility of the proposed methodology for real-time practical applications considering high variability of the primary resources. BUS N×N