Research On Shaft State Data Fusion Processing Based On Wavelet Analysis

Shaft power is an important parameter of ship condition monitoring,and the advancement of its data monitoring and processing technology is of great significance to the development of smart ships.The measurement of ship shaft power is achieved through the measurement of torque and speed.The technology for measuring speed is very mature with high accuracy.Real-time online monitoring of shaft torque is affected by many environmental factors,such as temperature,humidity,and severe vibration.Wait.In response to this problem,this article has carried out the following research around shafting condition monitoring,denoising,and real-time power prediction:(1)This article comprehensively considers the development difficulty of the monitoring system software design and the complexity of the system operation,and uses Lab VIEW software as the development platform of the shafting condition monitoring system to improve the stability and efficiency of the system on the premise of meeting the project requirements.The purpose is to formulate the overall design plan of the system,and finally realize the real-time measurement,display,processing,storage,fault alarm and other functions of multiple state parameters such as shaft torque,temperature,vibration acceleration and so on.(2)Aiming at the noise problem of shafting condition monitoring data,this paper uses wavelet transform to denoise the data.Analyzed the method of selecting the number of decomposition layers in wavelet denoising-the method based on signal characteristics,the physical and geometric meanings of the four commonly used traditional indicators,the root mean square error and the smoothness two indicators are selected,based on the improved entropy weight method A new composite evaluation index is constructed to determine the optimal decomposition level of the experimental data,and the optimal wavelet base is determined layer by layer according to the wavelet entropy of the low-frequency coefficients of the experimental data after wavelet decomposition.The results show that the decomposition level proposed in this paper And the wavelet threshold denoising scheme determined by the preferred method of wavelet base has better denoising effect.(3)In order to obtain reliable shaft power data,this paper selects neural network fusion algorithm to comprehensively utilize the monitoring data in the multi-sensor measurement system.Through the optimization of various hyperparameters in the BP neural network,the BP suitable for this project is determined The neural network model and the fusion result show that the fusion accuracy of the network model can meet the needs of the project.After comparing the original measurement data and the wavelet preprocessed data as the input data of the neural network model,the fusion accuracy of the network is compared,and it is concluded that the test data can be preprocessed to improve the accuracy of the fusion prediction model.(4)According to the requirements of this project for the functions,reliability and human-computer interaction of the monitoring system software,this article adopts a modular approach and a top-down approach to subdivide the functions of the shafting monitoring system.The functions of the system are distributed to each sub-module-system initialization module,user login module,parameter configuration module,data display module,data processing module,alarm record query module,etc.,and each sub-module is called by dynamic vi in Lab VIEW Connect to get a complete shafting condition monitoring system.Experimental tests show that the host computer software designed in this paper can run stably and reliably under the premise of meeting the needs of the project.