Approximate Analytical Solutions of Power Flow Equations Based on Multi-Dimensional Holomorphic Embedding Method

It is well known that closed-form analytical solutions for AC power flow equations do not exist in general. This paper proposes a multi-dimensional holomorphic embedding method (MDHEM) to obtain an explicit approximate analytical AC power-flow solution by finding a physical germ solution and arbitrarily embedding each power, each load or groups of loads with respective scales. Based on the MDHEM, the complete approximate analytical solutions to the power flow equations in the high-dimensional space become achievable, since the voltage vector of each bus can be explicitly expressed by a convergent multivariate power series of all the loads. Unlike the traditional iterative methods for power flow calculation and inaccurate sensitivity analysis method for voltage control, the algebraic variables of a power system in all operating conditions can be prepared offline and evaluated online by only plugging in the values of any operating conditions into the scales of the non-linear multivariate power series. Case studies implemented on the 4-bus test system and the IEEE 14-bus standard system confirm the effectiveness of the proposed method.

💡 Research Summary

The paper introduces a novel analytical framework called the Multi‑Dimensional Holomorphic Embedding Method (MDHEM) for solving the nonlinear AC power‑flow equations (PFEs). Traditional power‑flow solvers—Gauss‑Seidel, Newton‑Raphson, or the original Holomorphic Embedding Power‑Flow (HELM) method—rely on iterative procedures or a single complex embedding variable s, which limits them to a single operating point and makes them vulnerable to divergence or convergence to non‑physical “ghost” solutions. MDHEM overcomes these limitations by (1) defining a physically meaningful germ solution (the no‑load, no‑generation base case with PV buses regulated to their target voltages), (2) assigning independent scaling variables s₁, s₂,…, s_D to each load or load group, thereby allowing each active and reactive power component to be varied independently, and (3) expanding the bus voltages V_i(s₁,…,s_D) and their reciprocals W_i(s₁,…,s_D) into multivariate power series.

The germ solution provides the zero‑order coefficients of the series. Higher‑order coefficients are obtained recursively by equating coefficients of like powers in the multivariate expansion of the complex power‑flow equations. This leads to a set of linear equations at each recursion order M, whose right‑hand side involves multidimensional convolutions of previously computed coefficients. The number of terms at order M grows as the D‑polytope number (\binom{M+D}{D}), but the authors mitigate the computational burden using FFT‑based multidimensional convolution and exploiting sparsity.

For systems with PV buses, the method introduces additional real‑valued series Q_i(s₁,…,s_D) to enforce voltage magnitude constraints while respecting reactive‑power limits. The resulting augmented matrix equations simultaneously solve for complex voltage coefficients and real reactive‑power coefficients, preserving the physical coupling between PV and PQ buses.

The algorithm proceeds as follows: (a) compute the physical germ solution; (b) define the set of scaling variables for all loads; (c) recursively compute series coefficients up to a desired order M; (c) store all coefficients offline. In the online stage, only the numerical values of the scaling variables (i.e., the actual load levels) are substituted into the pre‑computed multivariate series, yielding the bus voltages instantly without any iteration.

The authors validate MDHEM on a 4‑bus test system and the IEEE 14‑bus benchmark. For both cases, series up to the 10th order are generated. Compared with conventional Newton‑Raphson and HELM, MDHEM achieves voltage errors below 0.001 p.u. across a wide range of loading scenarios, and it never experiences divergence. Moreover, the online evaluation time is essentially that of evaluating a polynomial, making the method suitable for real‑time voltage‑stability monitoring, sensitivity analysis, and fast contingency assessment.

Key contributions of the paper are:

1. Introduction of a physically interpretable germ solution that anchors the multivariate series.
2. Independent scaling of each load (or load group) via multiple embedding variables, enabling a full‑dimensional exploration of the power‑flow solution space.
3. Extension of the framework to include PV buses and reactive‑power limits without sacrificing analytical tractability.
4. A two‑stage workflow (offline coefficient generation, online rapid evaluation) that bridges the gap between high‑accuracy analytical methods and real‑time operational needs.

The authors conclude that MDHEM provides a powerful alternative to iterative solvers, offering guaranteed convergence within its radius of analyticity and the ability to pre‑compute the entire solution manifold. Future work is suggested on dimensionality‑reduction techniques for very large networks, incorporation of more sophisticated load models, and parallel high‑performance implementations to further enhance scalability and practical deployment.