Estimation and Forecasting of Excitation Force for Arrays of Wave Energy Devices

Abstract

To maximise energy conversion, real-time control of a Wave Energy Converter (WEC) requires knowledge of the present and future excitation force (Fex) acting on the device, which is a non-measurable quantity. The problem of estimation and forecasting of Fex becomes more challenging when arrays of WECs are considered, due to the hydrodynamic interactions in the array. In this paper, a global Fex estimator for a complete WEC array is developed and compared to a set of independent estimators which utilise information local only to each device. A significant question is whether the array of measurements is sufficient to compensate for the greater complexity of the wave field, compared to the isolated body case. The paper shows that the global estimator is always more accurate than the independent estimator, improving up to 45% the estimation accuracy of the independent estimator. Regarding prediction, two different Fex forecasters for a WEC array are compared: a global forecaster, utilising Fex estimates from the full set of array devices, and an independent forecaster, utilising only a local Fex estimate. We demonstrate that the global forecaster achieves more accurate results, not only compared to the independent forecaster, but also compared to the isolated body case.

Robust Excitation Force Estimation and Prediction for Wave Energy Converter M4 Based on Adaptive Sliding-Mode Observer

“…wheref (Q) e k|k−1 is the predicted value of f (Q) e k at instant t k−1 and p is the order of the AR prediction model, φ i are the AR coefficients with i = 1, 2, ..., p. The h-step-ahead prediction is obtained by using an iterative combination of 1-step-ahead predictions with h > 1 as a integer. The effectiveness of the AR model to predict the wave excitation force has been verified in [26], where the coefficients φ i are identified by an off-line training. An improved AR model is developed whose coefficients are online trained by the latest data at each v step.…”

“…In this section, the future information of the wave excitation force is predicted by an improved Autoregressive (AR) model, whose coefficients are updated by online training. Compared with the conventional AR model, which has been adopted for WEC wave excitation force prediction in [26], the improved AR model enhances the prediction accuracy with different sea states. Based on the improved AR model and the proposed ASMO, the boundaries of the excitation force prediction error at each future instant can also be explicitly formulated, which can provide guaranteed control performance for the non-causal controllers.…”

Towards realistic non-linear receding-horizon spectral control of wave energy converters

“…9 and 10 does not consider the case where the observed excitation force values are noisy. It should be pointed out that, like the SBP, the AR-based WF technique (Brekken, 2011;Fusco & Ringwood, 2010;Peña-Sanchez et al, 2018;Tona et al, 2015) is also consistent with the assumption of Gaussian, linear waves, but differs in two respects from the SBP:…”

Adaptive Model Predictive Control of Wave Energy Converters

“…To achieve non-casual control, t p seconds of wave elevation prediction are assumed to be available, which can be obtained using a short term wave forecasting technique, e.g. autoregressive (AR) [33]. Here, n p satisfies t p = n p t s for sampling interval t s .…”

“…is a n p -by-1 column vector with each element as 1; M k and C k are defined in (33); C v and C z are defined in (31).…”