Machine Learning-based Window Period Analysis Of Offshore Wind Power Construction Ship

In the depletion of traditional energy fuels such as oil and gas and the aggravation of the global climate environment,inexhaustible offshore wind energy has become the battlefield for the development of new energy in various countries,floating gondola as one of the indispensable operating ships in offshore wind power hoisting construction,its operation window directly affects the installation and construction costs in offshore wind power projects,and the prediction of the hull operation window of floating gondolas is very important for the construction of offshore wind power projects.With the reform and innovation of computer performance,big data technology has also entered the track of rapid development,and now the intelligent equipment data on the ship can realize real-time monitoring and storage of motion data in hull operations,and analyze these data through machine learning technology,which can predict the operation window of offshore construction ships,which is of farreaching significance for the construction of China’s future offshore wind power projects.In this thesis,the motion influencing factor of floating crane system is analyzed by numerical simulation calculation method,and the neural network technology in machine learning is used to study the operation window period in floating gondola operation,and the main research content is as follows:(1)According to the project,the floating crane system coupling model of crane shiptransport ship-fan impeller was constructed by using SESAM hydrodynamic software,and different lifting operating conditions were constructed by combining different external environmental factors and floating lifting factors,and the motion characteristics of the floating crane system coupling model under lifting operation conditions were studied,and the influencing factors affecting the stability of the floating crane system were obtained through comparative analysis.(2)On the basis of the motion law of the floating crane system under lifting conditions,the data sample database of the motion response of the floating gondola and the hanging object impeller under different working conditions is generated by combining different wind and wave current environmental parameters and floating lifting factors,and a BP neural network floating crane system motion response extreme value prediction model is built by using MATLAB,taking the average absolute error,root mean square error and average symmetric absolute error as the evaluation indicators of the network model performance,and determining the number of hidden layer nodes,the number of neurons and the amount of sample data in the network structure model.Combined with the model training fit and the error comparison between the final model prediction output value and the original data value,the effectiveness and feasibility of the BP neural network prediction model are verified.(3)Taking the motion amplitude of hull and hanging objects in the floating crane system as the research object,by reviewing the literature and combining the data observed by Chinese stations in the marine hydrology of the China Ocean Data Platform,the environmental characteristics of a wind farm in the East China Sea were analyzed,and 1728 sets of working condition data were generated based on environmental feature simulation,and the corresponding hull and hanging object movement response values predicted according to the BP neural network prediction model.Combined with the lifting equipment specifications of ships and offshore facilities and the operation specifications of the Nob Damton floating crane vessel,the operating window conditions of the floating gondola were studied and analyzed.(4)Based on the operation window conditions of floating gondolas,the decision tree theory is used to select the CART algorithm with the smallest cross-validation error and standard error to construct a decision tree model in MATLAB software,and the hull motion values are classified and processed,and the cost composition of offshore wind turbine lifting under different operating conditions is obtained,and the offshore wind power lifting cost under different sea conditions is obtained by combining the classification results with the prediction of labor,materials and ship leasing costs in offshore wind power construction.