Access Official Technical PDF

Share Summary:LinkedInEmail

Electric ship propulsion can create substantial power fluctuations as propellers interact with waves and changing operating conditions. These variations can affect the shipboard electrical network and contribute to mechanical stress in propulsion equipment. Battery-flywheel energy storage offers a way to buffer those changes: the storage system responds to fluctuating demand while generator sets supply average propulsion power.

For engineers and system integrators, the central question is not simply whether storage can smooth a load. It is which combination of technologies and control strategies can do so effectively under different operating conditions. A comparison of battery-flywheel (B/FW) and battery-ultra-capacitor (B/UC) hybrid energy storage systems shows why sea state, system losses and energy management all matter.

Why buffer propulsion load fluctuations?

Propulsion-load variations arise from propeller rotation and wave encounters. They can include components at different frequencies, from slower changes associated with wave effects to faster fluctuations linked to propeller blades. In an all-electric propulsion system, these variations appear as changing electrical power demand.

An energy storage system can absorb or supply power to reduce the fluctuations seen by the shipboard network. Using a hybrid system combines two storage technologies so they can share the work. This can help avoid relying on a single technology to meet every power and energy requirement, although the best pairing depends on the application.

The B/FW configuration combines lithium-ion batteries with flywheels. The battery state of charge (SOC) represents the available charge relative to capacity; flywheel SOC is tied to rotational speed. In a B/UC system, the second storage element is an ultra-capacitor, whose SOC is represented through its voltage. These different operating characteristics shape how each configuration responds to load changes.

Battery-flywheel energy storage: balancing response and losses

The control problem involves two competing goals: reducing power-tracking error and limiting losses in the storage system. More aggressive compensation can smooth the power demand more closely, but it can also increase energy-storage losses. The analysis therefore treats these goals as a multi-objective optimization problem and examines the trade-off between them rather than assuming that one operating point is best for every condition.

Dynamic programming (DP) is used to calculate global solutions and map the trade-offs as Pareto fronts. Because propulsion-load behavior is periodic, a periodic DP method can reuse information across repeated load cycles. This reduces the computational effort and memory required compared with treating every step across the full operating period as independent.

The comparison considers three sea states, representing smooth, moderate and severe conditions. Its results show that the relative value of each hybrid configuration changes with operating conditions:

  • At low and moderate sea states, the B/UC system generally achieves better performance than B/FW, although the difference varies with the operating point.
  • At the high sea state, B/FW can provide better load-fluctuation compensation with lower overall storage losses than B/UC.
  • Battery loading also differs: the B/FW configuration requires less battery current in the representative comparisons, including lower peak current and less time in high-current operation.

These findings are specific to the modeled configurations and conditions; they are not a universal ranking of flywheels and ultra-capacitors. They do, however, show why storage selection should account for the expected operating profile rather than only nominal power or energy requirements.

Understanding the flywheel trade-off

A key consideration for B/FW systems is flywheel spinning loss. At low sea states, when load fluctuations are smaller, this loss can become a significant part of the system’s total losses. The battery may need to supply energy to offset the flywheel’s spinning losses and maintain its operating speed. In that regime, the B/UC arrangement can have an advantage.

At high sea states, larger power demands over longer intervals can change the balance. In the comparison, the flywheel’s ability to sustain power at high speed supports stronger compensation, while its spinning loss is less dominant relative to the demanding load. The results also identify flywheel drag as an important design consideration, particularly for high-speed systems.

Battery condition matters as well. The sensitivity analysis found that increased battery resistance had a more noticeable impact on performance than reduced capacity in the examined cases. The effect of resistance varied with sea state and storage configuration, reinforcing the value of including battery condition in system-level assessment.

Making energy management suitable for real-time control

DP is useful for evaluating achievable performance, but its computational and memory requirements make it unsuitable for direct real-time control over long operating horizons. Model predictive control (MPC) provides an alternative: it repeatedly optimizes over a shorter prediction window, applies the next control action and then recalculates as new measurements become available.

A short MPC horizon can overlook longer-term operating needs. For example, the flywheel SOC may drift away from an efficient operating range, increasing the torque needed to deliver power and potentially worsening losses and tracking performance. The proposed approach addresses this limitation by adding an SOC reference penalty to the MPC objective. This encourages the storage element to remain near a selected operating range without requiring a long prediction horizon.

The study reports that the resulting B/FW MPC performance was close to the DP benchmark in the examined comparison. The approach was also evaluated on a laboratory-scaled test bed using a B/UC system to compensate load fluctuations and regulate a DC bus. This distinction is important: the reported laboratory validation demonstrates the MPC approach on that test configuration, while the B/FW comparison is based on the study’s modeled performance evaluation.

Practical implications for system design

For shipboard power-system designers, the results support a condition-based approach to hybrid storage selection. B/FW may be attractive where severe propulsion-load fluctuations are a key concern, while B/UC may offer an advantage under lower sea-state conditions. Flywheel standby losses, battery resistance, storage operating limits and control strategy should all be considered together.

More broadly, the comparison illustrates a principle relevant to industrial energy management: the storage technology alone does not determine system performance. A suitable architecture, realistic loss models and an implementable controller are all needed to translate load buffering into effective operation. For electric ship propulsion, pairing storage selection with an MPC strategy that accounts for SOC can help address both immediate power fluctuations and longer-term operating efficiency.

Access Official Technical PDF

About The Author