Font Size: a A A

Research On Key Technologies Of Leak Detection In Ship Pipeline Network Based On Transient Flow Characteristics

Posted on:2022-02-02Degree:MasterType:Thesis
Country:ChinaCandidate:Y WangFull Text:PDF
GTID:2492306566470474Subject:Master of Engineering
Abstract/Summary:PDF Full Text Request
Ship safety accidents caused by leakage and damage to the ship’s pipe network have occurred from time to time.If the ship’s pipe network is leaked,timely maintenance and overhaul will greatly reduce environmental pollution and ensure the personal safety of crew and passengers.Therefore,ship management is carried out.The research of network leakage detection technology is of great significance.Based on transient flow theory and computational fluid dynamics(Computational Fluid Dynamics,CFD),this paper takes the ship cooling piping system as an example to carry out leak detection technology research.Several typical pipe fittings in the cooling system are selected,and the transient process of fluid under different working conditions such as steady-state operation,leakage,pump adjustment and valve adjustment is simulated and calculated;the fluid behavior of different pipe fittings under different working conditions is analyzed and compared.Based on the characteristics of transient flow,an optimization plan for the layout of ship pipe network sensors is proposed;a pipeline experiment platform is built,and the collected signals are processed with wavelet noise reduction;a support vector machine(SVM)is established.Leak identification model,and optimized it;use the attenuation characteristics of negative pressure wave to conduct preliminary research on the location of the leak.The research results of this paper show that the transient flow characteristics of different pipe fittings under different working conditions are very different.The signal quality obtained when the sensor is arranged in the transition section of T-shaped pipe branch pipe,elbow pipe and Z-shaped pipe is higher.When denoising the ship’s pipe network signal,the wavelet denoising method using the default threshold of db4 wavelet is better.The SVM-based pipeline leak identification model can effectively identify leaks,and the use of Genetic Algorithm(GA)can increase the accuracy of the model to 97.72%.The pressure difference attenuation method is used to conduct a preliminary study on the location of short-distance straight pipe leakage,and the relative error is 8.93%.
Keywords/Search Tags:ship pipe network, leak detection, numerical simulation, wavelet analysis, support vector machine
PDF Full Text Request
Related items