Infometrics - Consulting Engineers Infometrics - Consulting Engineers Infometrics - Consulting Engineers
WiND FoRECASTING SERVICE USING ADAPTIVE NON-LINEAR KALMAN FILTERS

infoMETRICS proposal WiFoSAK has been evaluated, ranked 3rd among 155, and selected for funding. This project aims to develop a real-time, recursive and non-linear technological framework for accurate short and long term wind forecast that are derived from numerical and physical weather prediction models. The proposed model extends the traditional non-linear Kalman filters (Extended Kalman Filters) in a direction that allows real-time forecasting and recursive estimation of the unknown, non-linear function that relates the spatial-temporal distribution of wind flow over complex terrain. In particular, the traditional Kalman filters are enhanced with adaptive and non-linear capabilities algorithms, yielding to non-linear Kalman models which accurately estimate the unknown covariances of the measurements through a continuous train of the models with up-to-date measurements and under a real-time computational framework.

The main concept of this project is to develop a real-time, recursive and non-linear technological framework for accurate short and long term wind forecast that are derived from numerical and physical weather prediction models. The proposed model extends the traditional non-linear Kalman filters (Extended Kalman Filters) towards a direction that allows real-time forecasting and recursive estimation of the unknown non-linear function of spatial-temporal distribution of wind flow over complex terrain. Therefore, the traditional Kalman filters are enhanced with adaptive algorithms leading to accurate estimations of the unknown covariance of the measurements by continuously train the non-linear Kalman filter with up-to-date measurements under a real-time computational framework. The proposed methodology incorporates recursive non-linear models into the framework of Extended Kalman filters, resulting into a new smart, adaptive, non-linear prediction solution, capable of real-time estimating the unknown non-linear wind power model.

In addition, we inter-weave a Service Oriented Architecture (SOA) in order to communicate, in real time, with the SCADA wind turbines and retrieve wind flow operational characteristics, associating all this information with the adaptive, recursive, non-linear filter.

...More

Funded by







Tel: 0030 211 770 0998 · Fax: 0030 211 780 9899 · Email: info_at_infometrics.gr