Prediction Of The Noise Of Offshore Platform's Cabin Based On Intelligence Algorithm

Prediction of the noise of the Offshore Platform's cabin are important aspect in improveing Structure-Acoustic optimization design and reducing the noise.With the growing attention to enhances safety for operators involved in the offshore oil production.Both International Marine Organization(IMO),International Organization for Standardization(ISO)or many classification societies can formulate trestrictive rules for noise control of Offshore platform's carbin.However the offshore platform is an integrating system of production and life,which has complex internal structure and is difficult to accurately predict cabin noise.So this paper decided to adopt artificial intelligence method to solve this problem.The corresponding structural parameters are selected as input variable and compartment noise octave target act as output variable of intelligent arithmetic model.All these new steps make cabin's band-wise noise can be predicted quickly and easily.The research work are summarized as follows:(1)For the offshore platform's on-site test of the vibration and noise.These paper elaborates the principle of working conditions,measuring-point arrangements,test methods and presents some targeted methods of data handling.Which wil provided underlying data for follow-up numerical simulation and prediction algorithm.(2)Numerical analysis of the 3-D offshore platform acoustic model is simulated by using ststistical energy analysis method with the experimental load spectrum.Meanwhile I compare simulation results and experimental values to discover the distribution of cabin's noise.Then I research some critical influences about load noise,load transfer characteristics,construction features by the way of gray correlation analysis.And the databases are ultimately established though structural factors load factors and outfitting factors of the offshore platform.(3)There are five algorithms are briefly described from principle,algorithm idea,characteristics Including K-means clustering algorithm,gradient descending algorithm,orthogonal least square algorithm,particle swarm optimization and differential evolution algorithm.Meanwhile the corresponding MATLAB programs are also written which will be used to test algorithm performance with simulation data and finishing the optimization of algorithm parameters.The results indicate that differential evolution algorithm has best consequence of prediction algorithm of the cabin noise level,which has well-balanced propagation accuracy,excellent global optimization ability and good convergence.The rest show very large errors(4)Taking the actual data,joint data of simulation and actual,and actual data of another offshore platform into the prediction studies.The results indicate that algorithms have very large errors in the case of the band-wise noise prediction with a single actual data due to the imitation of sample.For the joint data of simulation databases and experimental databases.All algorithms have great propagation accuracy,high-stability and generalization capacity in larger training data set that can help improve forecast accuracy.For the another offshore platform's noise databases prodected process.All algorithms show bad predicted performance and don't have the predicted ability for another offshore platform's band-wise noise.