Research On Steady Wind Fields Estimation Technology Of Deck By Multi-sensor Fusion Algorithm

As an important observation method of wind field,ship-borne wind sensors can provide continuous wind field observation record,which has an extremely important impact on marine operations,safe navigation of ships and combat operations.However,due to the structure of the ship and its movement,the airflow field around the wind sensors will emergence vortex because of the turbulence effect.It will lead to a big difference between the measured wind field and the true wind field at infinity,therefore this article will try to reduce the influence of turbulence effect by using some kinds of methods to improve the measurement accuracy of the wind field.It is significant to research the optimal arrangement scheme of ship-borne sensors,because the different location has a large extent influence in measurement accuracy.By the CFD calculation software FLUENT15.0 with a reasonably simplified physical model,and using the method of numerical simulation to calculate the airflow field of calculation model under different conditions.By analyzing the data of the flow field,the relationship between the measurement error of the wind sensor and the location of the wind sensor is studied,and the location of the wind sensor with relatively small turbulence effect is given.The calculation shows that the error characteristics of the existing two-sensor arrangement scheme are large,and then provides three new kinds of arrangement schemes.And using single parameter analyzing the wind error characteristics under the different conditions,finally the optimal arrangement scheme is determined.This paper provides an important basis for the optimal installation location of anemometer and improving its the measurement accuracy.However,the three dual-sensor arrangement all can be the best solution in some specific wind directions.According to this phenomenon,this paper proposes a four-sensor arrangement scheme,and hopes to improve the wind field measurement accuracy by the four-sensor data fusion algorithm.This article presents two kinds of four-sensor data fusion algorithms,which are the optimal weighted fusion algorithm and BP neural network algorithm respectively.When using the optimal weighted fusion algorithm for data fusion,firstly,the simulation data of each wind sensor and the true wind field at infinity are respectively orthogonally decomposed to obtain the wind speed components of the X-axis and the Y-axis.By calculating the mean square error of wind speed components to get the best estimate of the wind speed component,and finally the wind field information of four-sensorfusion is calculated by vector combination of the wind speed components of X-axis and Y-axis.When using BP neural network to fuse data,the structure model of neural network is first determined.Then,a total of 8 wind speed and wind direction data measured by four sensors are taken as input.Finally,the ideal model is trained and the trained Wind field information.Through the simulation verification of the two kinds of multi-sensor fusion algorithms,found that the accuracy of the wind field measured by the two algorithms is much higher than the wind field measured by the traditional two-sensor's algorithm,it proves that the two algorithms are effective.