Power system harmonic state estimation via sparsity maximization

This paper presents a new system-wide harmonic state estimation method with the capability to identify harmonic sources with fewer meters than state variables. Note there are only a few simultaneous harmonic sources among the suspicious buses. By extending the concept of observability, the underdetermined system can be observable when considering the sparsity of harmonic sources. We formulate harmonic state estimation as a constrained sparsity maximization problem. It is solved by linear programming using L1-norm equivalence to L0-norm. Meter placement is optimized to enhance the robustness of the solution. Our numerical experiments in IEEE 14-bus power systems show the effectiveness of the proposed method