Identification of the probability of the park effect in a wave-to-power system using the analytical hierarchical process and a polynomial neural network model

The park effect occurs for various reasons in field applications of wave energy conversion. Not all possible causal factors are equally responsible for causing the park effect. The varying significances of the causes were approximated in an objective manner using the analytical hierarchical process multi-criteria decision-making technique and were assessed with an index function where the reasons were used as indicators and the function was the aggregated weighted contribution of each individual reason. A real-time monitoring system was also developed to monitor peak effects in a wave-to-power system in a continuous manner, such that loss due to the park effect could be prevented in real time, ensuring minimum loss of production capacity. Such a monitoring system would use the same aggregated weighted function to monitor the effects, and a network of sensors fixed at strategic points would be used to retrieve the magnitudes of the causal factors. An automatic framework to estimate the park effect was developed using a polynomial neural network architecture, and it was tested for three different locations. An experimental validation of the index values for these three locations was performed. The results indicated that offshore locations of wave-to-power systems were the most vulnerable to the park effect. It was also found that, both offshore and near shore, the distance between converters has the highest significance, whereas for onshore converters, the position of the WEC in relation to the incoming wave was found to be the most important indicator among the six investigated in the present study. Significant uncertainties appeared due to the lack of field-tested results for use in calculating the park effect on devices. However, the use of a physical model reduced these uncertainties, and a reliable framework was developed. Further real-time testing of the indicators will help develop practical implementations of the index for the formulation of policy.