Real-time implementation, validation and physical constraints often determine whether a wave energy control concept will hold outside the laboratory. Here, we explain how.
To better understand what separates simulation from operational reality, we sat down with John V. Ringwood, Professor of Electronic Engineering at Maynooth University and Director of the Centre for Ocean Energy Research (COER). In INFINITY, he leads the development of the control system, from mathematical modelling and simulation to near real-time implementation together with the other project partners.
Ringwood describes a structured path from model to controller: first a validated mathematical description of the system, followed by an energy-maximising control philosophy, and finally a device-specific control algorithm. The model can also be used to construct an estimator for the wave excitation force (WEF) — a directly unmeasurable quantity required by some wave energy controllers. He points to WEF as a key variable in determining the optimum velocity trajectory for the wave energy converter (WEC).
How accurate must the model be for the control design to remain valid in real operation?
– This question does not have an exact answer! No model is perfect, Ringwood says.
The control algorithm must be computationally implementable in real time. That requirement limits the complexity — and therefore the level of detail — of the model that can be employed.
So where does the practical limit lie?
– In practice, it manifests itself in the achievable sampling rate. The required sampling rate, in turn, is determined by the speed of the WEC system dynamics.
As more validation data from real-world conditions become available, across increasingly broad operating ranges, the accuracy of the model can be improved. Validation data may come from tank testing or from open-ocean trials. As the operational envelope expands, the model can be progressively refined. The validation process follows established procedures used to test key system components. However, Ringwood highlights one aspect in particular:
– The design of the excitation signals for waves and the power take-off (PTO) is important in providing broad validation of models across a wide range of operating conditions.
Which modelling assumptions most often cause problems when moving from simulation to a physical wave energy system?
– Models are often assumed to be linear, which is generally only valid within a limited operating range.
Moving from linear to nonlinear modelling — and adapting the control design accordingly — involves a significant increase in complexity, analysis and computational requirements. For that reason, many practitioners continue to rely on linear models.
Physical constraints must also be addressed explicitly. Typical constraints in a WEC relate to power take-off (PTO) parameters, such as displacement, velocity and force. In some systems, bi-directional power flow is not permitted, introducing an additional restriction on the direction of power flow. These constraints must be integrated into the control design.
Real-time control depends on measurement
Ringwood identifies WEC displacement and velocity, PTO force, and the wave excitation force as the most important variables for the controller. As noted earlier, WEF is directly unmeasurable and must therefore be estimated.
How can control remain effective when sea states, parameters or components deviate from the model?
– A branch of control, known as robust control, can be used to reduce sensitivity to unknown or changing system parameters, Ringwood explains, and continues:
– Robust control is not a magic wand – a more accurate model, with reduced uncertainty, will result in better performance. The reality is that robust control is no substitute for good modelling!
As the conversation draws to a close, we ask what he considers the most important lesson in bridging the gap between simulation and physical systems.
– There are few substitutes for real-world experience, Ringwood concludes.
He points not only to the importance of testing a theoretical design, but also to incorporating operational experience already at the modelling and control design stage. Such experience provides a more realistic foundation for the assumptions underlying both the model and the controller.
Taken together, his answers describe a tightly interconnected process in which modelling, validation and implementation must work in concert. The step from model to ocean is not defined by a single breakthrough, but by the consistent alignment of mathematical description, computational feasibility and operational reality.