Industrial pipelines must be inspected to detect typical failures, such as obstructions and deformations, during their lifetime. In the petroleum industry, the most used non-destructive technique to inspect buried pipelines is pigging. This technique consists of launching a Pipeline Inspection Gauge (PIG) inside the pipeline, which is driven by the pressure differential produced by fluid flow. The purpose of this work is to study the application of artificial neural networks to calculate the PIG’s velocity based on the pressure differential. We launch a prototype PIG inside a testing pipeline, where this PIG gathers velocity data from an odometer-based system, while a supervisory system gathers pressure data from the testing pipeline. Then we train a Multilayer Perceptron (MLP) and a Nonlinear Autoregressive Network with eXogenous Inputs (NARX) network with the gathered data to predict velocity. The results suggest it is possible to use a neural network to model the PIG’s velocity from pressure differential measurements. Our method is a new approach to the typical speed measurements based only on odometer, since the odometer is prone to fail and present poor results under some circumstances. Moreover, it can be used to provide redundancy, improving reliability of data obtained during the test.

The pipeline inspection using a device called Pipeline Inspection Gauge (PIG) is safe and reliable when the PIG is at low speeds during inspection. We built a Testing Laboratory, containing a testing loop and supervisory system to study speed control techniques for PIGs. The objective of this work is to present and validate the Testing Laboratory, which will allow development of a speed controller for PIGs and solve an existing problem in the oil industry. The experimental methodology used throughout the project is also presented. We installed pressure transducers on pipeline outer walls to detect the PIG’s movement and, with data from supervisory, calculated an average speed of 0.43 m/s. At the same time, the electronic board inside the PIG received data from odometer and calculated an average speed of 0.45 m/s. We found an error of 4.44%, which is experimentally acceptable. The results showed that it is possible to successfully build a Testing Laboratory to detect the PIG’s passage and estimate its speed. The validation of the Testing Laboratory using data from the odometer and its auxiliary electronic was very successful. Lastly, we hope to develop more research in the oil industry area using this Testing Laboratory.

Pipelines are a key component of an oil and gas supply system, so their maintenance is essential. Among the maintenance techniques, the use of PIGs has been successfully applied in many situations, such as cleaning, product separation and integrity inspection. Generally, the velocity a PIG travels inside the pipeline must be constant for higher efficiency, so we propose a system that aims to simulate the conditions of a gas pipeline in order to develop speed controllers. The system consists of a 6-inch pipeline with pressure sensors to detect PIG's motion in order to provide information about its dynamics.

A device known as a pipeline inspection gauge (PIG) runs through oil and gas pipelines which performs various maintenance operations in the oil and gas industry. The PIG velocity, which plays a role in the efficiency of these operations, is usually determined indirectly from odometers installed in it. Although this is a relatively simple technique, the loss of contact between the odometer wheel and the pipeline results in measurement errors. To help reduce these errors, this investigation employed neural networks to estimate the speed of a prototype PIG, using the pressure difference that acts on the device inside the pipeline and its acceleration instead of using odometers. Static networks (e.g., multilayer perceptron) and recurrent networks (e.g., long short-term memory) were built, and in addition, a prototype PIG was developed with an embedded system based on Raspberry Pi 3 to collect speed, acceleration and pressure data for the model training. The implementation of the supervised neural networks used the Python library TensorFlow package. To train and evaluate the models, we used the PIG testing pipeline facilities available at the Petroleum Evaluation and Measurement Laboratory of the Federal University of Rio Grande do Norte (LAMP/UFRN). The results showed that the models were able to learn the relationship among the differential pressure, acceleration and speed of the PIG. The proposed approach can complement odometer-based systems, increasing the reliability of speed measurements.

No abstract