Comparing Loss Allocation Methods: Fairness, Accuracy & Regulatory Acceptance

🎯 Learning Objectives

✓ Calculate loss allocation shares using the Proportional Sharing Method (PSM) and Marginal Loss Coefficient (MLC) method
✓ Analyze fairness implications of different allocation methods using sensitivity to load/generation location and topology
✓ Explain regulatory acceptance criteria for loss allocation in ISO-operated markets (e.g., NYISO, PJM, ENTSO-E)
✓ Apply the Current Injection Method (CIM) to a 3-bus radial system and interpret nodal loss contributions
✓ Compare accuracy and computational complexity of five standard loss allocation methods under varying loading conditions

📖 Why This Matters

In modern deregulated power systems, who pays for line losses isn’t just an engineering question—it’s a multi-million-dollar billing, fairness, and market design issue. A poorly allocated loss can distort investment incentives (e.g., discouraging distributed generation near load centers), trigger disputes among market participants, and violate FERC Order No. 888 or EU Regulation (EU) 2019/943. Understanding *how* and *why* losses are assigned—not just *how much* is lost—is essential for engineers designing tariffs, participating in ISO markets, or advising regulators.

📘 Core Principles

Loss allocation rests on two foundational ideas: (1) Losses are nonlinear functions of power flows, making exact attribution impossible without approximation; and (2) Fairness requires satisfying key axioms—symmetry, efficiency, and null-player (no contribution → no allocation). The Proportional Sharing Method (PSM) treats losses as shared proportionally to injected/withdrawn power—a simple but topology-agnostic approach. In contrast, marginal methods (e.g., MLC) compute incremental loss change per unit injection at each node, yielding location-sensitive allocations aligned with economic dispatch principles. Advanced methods like the Current Injection Method (CIM) use circuit theory to decompose losses into contributions from each generator’s current phasor, preserving Kirchhoff’s laws and enabling physical traceability. Regulatory frameworks increasingly require transparency, reproducibility, and consistency across timeframes—making method choice a governance as much as a technical decision.

📐 Marginal Loss Coefficient (MLC) Method

The Marginal Loss Coefficient quantifies how much total system loss changes when 1 MW is added at a specific bus. It’s derived from the Jacobian of total loss with respect to real power injection and is used in locational marginal pricing (LMP) frameworks. MLC is computed via power flow sensitivity analysis and forms the basis for loss components in LMP calculations.

Marginal Loss Coefficient (MLC)

MLC_i = ∂L_{total}/∂P_i

Measures incremental system loss (MW) per unit real power injection (MW) at bus i.

Variables:

Symbol	Name	Unit	Description
MLC_i	Marginal Loss Coefficient at bus i	MW/MW	Dimensionless sensitivity factor
L_{total}	Total system real power loss	MW	Sum of I²R losses across all branches
P_i	Real power injection at bus i	MW	Net generation minus load at bus i

Typical Ranges:

Strongly meshed HV grid (e.g., German UCTE): 0.01 – 0.05

Radial distribution feeder: 0.10 – 0.35

Weakly connected offshore wind interconnection: 0.40 – 1.20

💡 Worked Example

Problem: Given a 3-bus system where total system loss L = 0.02P₁² + 0.015P₂² + 0.008P₃² + 0.006P₁P₂ (MW), with P₁ = 120 MW (Gen at Bus 1), P₂ = 80 MW (Gen at Bus 2), P₃ = −180 MW (Load at Bus 3). Calculate MLC at Bus 1.

1. Step 1: Differentiate total loss L w.r.t. P₁: ∂L/∂P₁ = 0.04P₁ + 0.006P₂

2. Step 2: Substitute P₁ = 120, P₂ = 80 → ∂L/∂P₁ = 0.04(120) + 0.006(80) = 4.8 + 0.48 = 5.28 MW/MW

3. Step 3: Interpret: Adding 1 MW at Bus 1 increases total system loss by ~5.28 MW — unusually high due to quadratic coupling; verify convergence and check for numerical instability.

Answer: The MLC at Bus 1 is 5.28 MW/MW, indicating strong sensitivity—common near weakly connected buses or low-X/R lines. Typical MLCs range from 0.01–0.15 in well-conditioned systems; values >1 suggest model or topology issues requiring review.

🏗️ Real-World Application

In PJM Interconnection’s 2023 tariff filing (FERC Docket No. ER23-2793), the MLC-based loss allocation was validated against the Z-bus method for 2,100+ buses over 8760 hours. PJM found MLC yielded <0.8% average deviation from physical loss tracing but reduced computation time by 92% vs. full CIM. Crucially, PJM mandated MLC be recalculated hourly with updated base cases—and required public disclosure of MLC matrices to satisfy FERC’s transparency rule (Order No. 2222). This case demonstrates how regulatory acceptance hinges not only on technical soundness but also on auditability, timeliness, and stakeholder accessibility.