Feature recognition method for umbilical cable leakage fault

Abstract

he invention relates to a feature recognition method for a umbilical cable leakage fault, characterized by comprising the following steps of: 1) acquiring the state feature signal of an offshore platform underwater production system according to the hydraulic pressure of a umbilical cable and the state feature signal of a chemical agent pipeline monitored by a sensor during the operation of the umbilical cable; 2) according to the state feature signal of the offshore platform underwater production system, obtaining the operating state features of the umbilical cable; 3) based on a umbilical cable leakage simulation model, constructing a umbilical cable leakage fault sample feature database and using the same as a training sample; 4) studying and training the training sample to construct a data model of the training sample; 5) diagnosing the operating state features of the umbilical cable according to the data model of the training sample to obtain a diagnostic network describing a mapping relationship between feature variables and the leakage fault so as to realize the feature recognition of the umbilical cable leakage fault. The feature recognition method can be widely applied in the field of umbilical cable leakage fault diagnosis.

Description

technical field [0001] The invention relates to a feature recognition method for umbilical cord leakage faults, and belongs to the field of umbilical cord leakage fault diagnosis. Background technique [0002] As the key equipment of the underwater production system of the offshore platform, the deepwater umbilical cable provides hydraulic pressure, electricity, signal control and injection of chemicals for the underwater oil well. A difficult technical problem. Due to the limited test conditions, it is impossible to conduct physical tests on leakage faults on umbilical cables to extract characteristic signals, and the samples of leakage faults are very limited. It is almost impossible to use real umbilical cables to study the leakage of hydraulic and chemical agent pipelines. To this end, the researchers proposed an umbilical leakage fault feature identification method based on the combination of umbilical leakage fault sample feature library and big data analysis, and car...

AI Technical Summary

Problems solved by technology

Constrained by the underwater environmental conditions, the probability of successfully observing a leak is low. Even if a leak is detected, it is very time-consuming. Only a dozen meters to tens of meters long pipelines can be detected in one day, and underwater, especially deep water (the water depth is greater than 1500 meters) the length of the umbilical cable is thousands of meters long, the ROV detection cannot meet the engineering requirements, and at the same time cannot monitor in real time, the engineering practicability and efficiency are low

Images