Real time monitoring of a fluid transportation system is still a challenging matter, due to the complexity of the asset and a continuous demand of sustainability. The current frontier is a re-design of the information management during the asset lifecycle, with the digitalization and collection of large datasets which are used to infer "data driven" solutions. This paper presents a monitoring strategy based on the medium/long term detection and tracking of "smart indicators" of the oil & gas transportation system operations. The indicators exploit the digital vibroacoustic signals recorded by proprietary stations located on the conduit at an inter-distance of around 20 km. The collected data are analysed on a long-term basis, together with the measurements coming from other instruments (i.e. temperature, density, flow rate), in order to highlight small variations in the pipeline, which move the operational parameters to new states. As an example, sound speed, and attenuation are inverted for the fluid composition and the inner pipe condition. The correlation analysis of the signals recorded in consecutive pipeline segments, interpreted as an equivalent acoustic channel, reveals pipe deformations and/or flow anomalies.

We show a case history of an oil trunkline in Nigeria, conveying a variable ratio water-oil mixture, with an inter-distance between the stations along the pipeline at about 17 km. We derive from the long-term and medium-term database some key parameters, which are the actual input for automatic digital monitoring algorithms.


Big data technology is permeating oil & gas industries for information management and predictive analytics. The assets are becoming "digital", i.e. a network of interconnecting and interacting data, software and hardware, making it possible for the design of "data smart" operations. A fluid filled pipeline transportation system is a complex system whose dynamic parameters (i.e. pressure, temperature, flow, etc.) vary over a wide range of values (Glen, 2005). For a given time, these parameters can be represented as a "working" point in a multidimensional space, and a variation of the operational status (i.e. pumps switch, flow regulation, pigging campaigns) defines a "trajectory" in the same space. Therefore, managing the system's integrity means to define the safe regions in the parameters space, the safe trajectories to go from one to the other, and to implement operative procedures so as to keep and/or to move the working point within these boundaries. The complexity increases when dealing with multiphase fluids, due to the variability of flow regimes and of the mixture composition, which makes it almost impossible to obtain a full theoretical framework.

Real time monitoring of oil & gas scenarios is then a challenging issue: very promising results are envisaged today by the "digital" view of the asset, feeding machine learning procedures with high rate multi domain data (e.g. pressure, temperature, vibrations, flow rate, density, superficial velocity, etc.). On one hand, machine learning techniques are able to extract, from a pre-classified set of training data, constitutive relations that serve to estimate the "movements" of the working point. On the other hand, the training process efficacy and the overall classification performance can be enhanced when providing combinations of data, or "smart" indicators, which are more directly connected with some system parameter than the raw data itself (Mohamed, 2015).

This paper defines some "smart" indicators derived from vibroacoustic measurements collected on oil transportation pipelines in Nigeria. The proprietary integrity monitoring system is operative on downstream and upstream pipelines, around 1200 km in Italy and Nigeria (Bernasconi et al., 2014; Giunta et al. 2014, Giunta et al. 2016).

This content is only available via PDF.
You can access this article if you purchase or spend a download.