Optimal sizing of energy storage system for marine photovoltaic systems with MP-DPC: A case study

• A novel MP-DPC framework integrated with a rolling ultra-short-term PV power forecasting model is proposed. • The strategy achieves a significant reduction in Energy Storage System (ESS) capacity while adhering to strict ramp-rate limits. • A composite metric balancing technical performance and economics is introduced. • Comprehensive validation across six weather scenarios confirms the method's robustness. . The integration of photovoltaic (PV) power system into ship power system represents a promising route for maritime decarbonization. However, the dynamic ship navigation attitude and the changeable weather conditions cause severe PV output variability, challenging system stability and control. To mitigate these fluctuations, integrating a large-capacity energy storage system (ESS) into PV power system is generally regarded as one of the most effective means. Although the application of variable-power grid-connected methods can reduce the size of ESS, it is still difficult to meet stringent ramp rate limits because of the time delays associated with measurement, communication, and control actions. To overcome these constraints, this study proposes a model predictive direct power control (MP-DPC) strategy enhanced by an ultra-short-term PV power rolling forecasting model. The neural network based forecasting module supplies one second ahead PV power predictions, enabling MP-DPC to issue proactive charge/discharge commands that uphold ramp rate constraints while minimizing ESS capacity. The approach is evaluated on a 102.6 kW marine diesel/PV/ESS hybrid power system using one second resolution data across six weather scenarios. Fluctuation thresholds from 1 %/min to 10 %/min of rated power are tested, and a composite metric combining normalized ESS cost and mean ramp rate guides the trade-off between smoothing performance and economics. Results show that, at a 3%/min fluctuation threshold, the proposed MP-DPC framework achieves a system-level compromise that remains robust across different weighting perspectives, requiring an ESS capacity of 11.0042 kWh with an associated investment cost of approximately USD 5,502.12. Furthermore, comparative benchmarking against representative PV–ESS hybrid ships reported in the literature indicates that the proposed MP-DPC strategy achieves a markedly lower battery-to-PV ratio of 0.1072, demonstrating superior ESS capacity utilization.