Smooth Dual-Mode Service Fee Mechanism With Negative-Impact Voltage Assessment for Grid-Friendly P2P Energy Trading

Abstract

Peer-to-peer (P2P) energy trading can intensify voltage deviations in distribution networks, whereas conventional fixed service fees do not distinguish market participants according to their voltage impacts. This paper proposes a smooth dual-mode voltage-impact service-fee mechanism for bilateral-auction-based P2P energy trading. LinDistFlow-based Voltage Sensitivity Factors (VSFs), pre-clearing voltage states, and transaction-level Negative Impact Indicators (NIIs) are combined to identify voltage-adverse trading directions. The Negative Sensitivity-Based Fee (NSBF) forms the low-stress limiting state, while a node-wise activation factor continuously introduces Voltage-Aware Adaptive Fee (VAAF) weighting as pre-clearing voltage deviations enter a prescribed transition band, thereby avoiding abrupt mode switching. Adverse and beneficial voltage movements are evaluated separately to prevent mutual offsetting. The proposed fee is embedded into an auction-compatible two-stage framework that coordinates day-ahead network scheduling with rolling sequential intra-day clearing. The day-ahead stage establishes a grid-feasible operating baseline, whereas the intra-day stage incorporates the nodal service fee into participants’ effective bids and asks without reformulating the bilateral auction as a network-constrained optimal power flow. Studies on a modified IEEE 33-bus feeder show that cumulative adverse voltage movement is reduced by 32.91% and 39.47% relative to fixed fees under normal low- and high-voltage conditions, respectively. Under the voltage-stressed condition, the smooth mechanism reduces adverse voltage movement by 20.13% relative to NSBF while preserving the same total P2P trading volume. Nonlinear AC power-flow validation shows that the maximum error in the incremental P2P-induced voltage movement remains below $7.36 \times {{10}^{ - 4}}$ p.u., with a minimum Pearson correlation of 0.999473 and preserved comparative ordering among the evaluated mechanisms.