Systems Analysis and Stochastic Optimization of the Impact of a Mining Cluster Located near a Nuclear Power Plant on Frequency Stabilization

Abstract

Systems-analysis study and a stochastic optimization model of the impact of deploying a managed cryptocurrency mining cluster at a nuclear power plant (NPP) site on frequency stabilization in the unified power system are developed. The maneuverability of nuclear units is confined to a narrow primary-control band (±2 % of nominal power) and does not allow participation in secondary reserves (FRR/aFRR) due to nuclear-physical constraints, while the growing share of renewable sources increases the demand for such resources. The proposed approach resolves this contradiction: an ASIC cluster operates as a controllable consumer that reduces the consumption within the ENTSO-E FCR deployment time (full activation ≤ 30 s) without affecting the thermal regime of the reactor.

A two-stage stochastic optimization model is developed: the first stage (day-ahead) determines the baseline cluster loading and the volume of FCR commitments, while the second stage (real-time) adjusts the plan under FCR activations across stochastic scenarios. Reliability of FCR obligations is guaranteed by chance constraints at a specified confidence level (α = 0.95). Uncertainty parameterization covers cryptocurrency price forecasting and FCR activation frequency, while stochastic scenarios of the power system state are described by autoregressive models of post-war recovery of Ukraine's energy system and renewable generation. Scenario generation uses Monte Carlo sampling with reduction to 200 representative trajectories.

A numerical example for a VVER-1000 unit demonstrates that the proposed "mining + FCR" strategy provides approximately 23 % higher daily revenue than pure mining and approximately eightfold economic advantage over a battery energy storage (BESS) alternative through the absence of CAPEX and cluster monetization in standby mode. Load localization yields an additional effect of approximately 284 million UAH annually through network loss reduction. The scientific novelty lies in formulating a two-stage stochastic model with chance constraints together with an uncertainty parameterization that integrates cryptocurrency price forecasting, FCR activation frequency, and autoregressive scenarios of the state of Ukraine's power system and renewable generation. The practical contribution is an adaptation of the U.S.-implemented behind-the-meter architecture of the Nautilus Cryptomine project to the conditions of Ukraine's power system, where a mining cluster colocated with an NPP serves as an instrument for primary frequency regulation that compensates for the limited maneuverability of VVER-1000 units.