loading page

Maximum Likelihood Estimation of Line Parameters for a Nonlinear Model of Distribution Grids
  • +1
  • Shubhankar Kapoor,
  • Adrian G Wills,
  • Johannes Hendriks,
  • Lachlan Blackhall
Shubhankar Kapoor

Corresponding Author:[email protected]

Author Profile
Adrian G Wills
Johannes Hendriks
Lachlan Blackhall


This paper presents a method for obtaining distribution grid models based on available measurements from the grid. In particular, we propose a nonlinear approximation of the DistFlow model [1], which includes line losses between nodes and is parameterized by the unknown line parameters, namely the impedance between nodes. Based on measured voltage magnitude at each node, and net injected real and reactive powers at each node, we estimate the unknown line parameter values. To achieve this, we formulate a maximum-likelihood problem that we solve using an expectation-maximisation (EM) approach. In particular, we tailor the EM approach for the proposed nonlinear grid model, and we provide a numerically robust implementation of the resulting algorithm. The proposed method is demonstrated on the IEEE 37-node test feeder network, where we compare to state-of-the-art methods and show that the proposed approach achieves a 70% reduction in voltage error and more than 10, 000 times lower error for state variables. A final comparison uses data from a real network, and we show that the proposed approach achieves parameter estimates that are up to 100 times better than competing approaches.
19 Mar 2024Submitted to TechRxiv
29 Mar 2024Published in TechRxiv