Abstract：The high proportion of new energy grid connection increases the frequency fluctuation of the power

grid，and grid-forming energy storage is regarded as a key resource to improve frequency stability due to its fast

response and inertia support capability. To address the underestimation of fast-response resource value and

insufficient fairness in revenue allocation within existing frequency modulation auxiliary service mechanisms，a

method for optimizing the capacity allocation and revenue of grid-forming energy storage was proposed based on

the improved Shapley value. A cooperative scheduling framework was constructed，in which grid-forming energy

storage focuses on frequency modulation while grid-following energy storage cooperates with peaking. The inertia

support and dynamic response characteristics were quantified through a dual-track compensation mechanism. A

two-level game model was established to minimize grid power purchase costs and maximize energy storage

revenue for global optimization. The Shapley value method was improved by introducing a three-phase correction

involving technological contribution，capacity availability，and dynamic response capability，thereby constructing a

multidimensional contribution evaluation system. Simulation results demonstrated that the proposed method

increased the system frequency modulation demand satisfaction rate to 98.7%，enhanced the frequency modulation

capacity availability of grid-forming energy storage by 35%，raised the revenue of grid-forming energy storage by

15.5%，and boosted the total revenue by 59.7%. These improvements effectively enhanced resource allocation

efficiency and system economy. The study provides a theoretical paradigm and practical pathway for the marketbased operation of frequency modulation resources in high-proportion renewable energy grids.

      Key words：grid-forming energy storage；frequency modulation ancillary services；improved Shapley value

method；two-level game；capacity allocation；revenue optimisation