Research On Acoustic Emission Based Offshore Platform Pipeline Leakage Signal Identification

Offshore oil and gas platform pipelines play an extremely important role in the process of oil and gas extraction,transportation and processing,but due to its harsh working environment,in the long-term service process will inevitably produce aging damage problems,once the media leakage and not detected in time,it is very easy to cause economic losses and casualties,so the platform oil and gas pipeline leak detection is very necessary.Due to the complex working environment and structure of offshore platform oil and gas pipelines,some traditional detection methods are difficult to apply effectively,while acoustic emission technology has performed well in the field of pipeline leakage detection in recent years,with its simple arrangement,high positioning accuracy,high detection success rate and the advantages of being able to detect in the operating condition of the equipment also providing support for its application in the direction of detecting oil and gas pipelines on offshore platforms.Therefore,this paper carried out a study on the identification and location of oil and gas pipeline leakage on offshore platforms based on acoustic emission.The main work contents are as follows:First,the pipeline leakage acoustic emission model experiment and the field noise acquisition experiment of offshore platform were designed and carried out.A large number of acoustic emission signal samples were collected,and the characteristics of leakage signals in the time domain and frequency domain were compared and analyzed.Then the two were integrated to simulate the acoustic emission signals in the actual pipeline leakage.Secondly,for the problem of large noise interference in the actual platform pipeline,wavelet packet decomposition combined with PNN probabilistic neural network method is used to extract and identify the features of the pipeline leakage signal.By determining the optimal number of wavelet packet decomposition layers and the optimal wavelet bases,the method extracts features according to the different frequency distribution intervals of the leakage signal and the noise signal,and then combines the PNN probabilistic neural network for recognition,finally achieving good recognition results and certain anti-noise interference capability.Then,in order to further improve the universality of the detection method,a pipeline leakage signal identification method based on modal decomposition combined with PNN probabilistic neural network is investigated.The feature extraction results of EMD,VMD and CEEMDAN decomposition of pipeline leakage signals and the recognition results combined with PNN probabilistic neural network are compared,and the results show that the feature extraction and recognition method of CEEMDAN decomposition combined with PNN probabilistic neural network is the most effective and has the best resistance to noise interference.Finally,to address the problem that leak source localisation is prone to errors caused by noise interference,the CEEMDAN decomposition combined with HB-weighted generalised intercorrelation method is used for pipeline leak localisation.By screening the effective CEEMDAN components to reduce noise interference and selecting the HB weighting with strong resistance to noise interference,the localization accuracy is higher,and afterwards the BP neural network is used to correct the localization results,and the localization accuracy can be further improved after optimization.The research work of this paper will provide some theoretical and data support for the acoustic emission detection technology of oil and gas pipelines on offshore platforms.