Research on the Risk Assessment Model of Prefabricated Wind Power Hybrid Tower Based on Big Data and Machine Learning Algorithm

The heightened use of wind energy triggers the enhancement of risk evaluation methods that will support the reliability of prefabricated hybrid wind power towers. The evaluation tool in this piece of research is a data driven risk analysis model incorporating big data and machine learning techniques suitable for prefabricated hybrid towers. Analysing a dataset with more than a billion records that include data from 150 active towers from 2013 to 2023, the present work points to the most prominent risks, including material fatigue, vibrations, weather conditions, and problems with maintenance. The main intelligent techniques used in the framework of this work included Gradient Boosting Machines (GBMs) and Long Short-Term Memory (LSTM) networks for the assessment of risk levels and the identification of the best time for maintenance. The effectiveness of the proposed GBM model can be evidenced by experimental results which yields an overall accuracy of 93%, and the root mean square error (RMSE) of 0.12 and precision rate of 91%. Thus, the ratio of feature importance regarding the identified risks was consistent with the physical conditions affecting wind tower operation, where vibration levels and material fatigue dominated (55%). The proposed model has application to greatly minimize the intensity of manual inspection, improves the precision of predicted figures, and enables maintenance planning well in advance, thereby likely to help organizations to cut down operational expenses by as much as 15%.