Eventos Anais de eventos
DINAME 2017
XVII International Symposium on Dynamic Problems of Mechanics
ASSESSMENT OF UNCERTAINTIES AND PARAMETER ESTIMATION IN OFFSHORE GAS PIPELINE
Submission Author:
Elói Rotava , SP
Co-Authors:
Elói Rotava, Jorge Luis Baliño, Flávio Trigo
Presenter: Elói Rotava
doi://10.26678/ABCM.DINAME2017.DIN17-0154
Abstract
Oil production in Brazil has turned its attention to the Santos basin in recent years because it is the most promising national oil province. Despite the crisis facing the operating companies due to the current low commercial value of the barrel, it is expected output growth in this region. It is an offshore basin with a large number of different reservoirs, with a strategy to start production in a short period of time. For the flow of natural gas associated with the production of this oil are being built infrastructure consisting of pipelines, reaching a total of over 700 km long, and gas processing plants on land. The concept applied in the development of these fields, which prioritizes the anticipation of production fields, follows the concept of main pipelines where production units are connected and export natural gas. The operation of this gas pipeline is based on numerical models for calculation of intermediate properties, future behavior prediction and estimates the capacity of the integrated flow. These models are based on assumptions for parameters such as heat exchange and field measurements of boundary conditions such as pressure, flow, temperature and composition of the natural gas. This paper presents the development proposed for state and parameter estimation based on the implementation of an extended Kalman filter, in order to determine appropriate values the parameters of the flow and use of complementary measurements in the boundary conditions. The proposed flow model, necessary to the Kalman filter, is one-dimensional, transient, compositional and capable of solving complex networks, being based on the mass, momentum, species and energy conservation, and uses as closing equations the equation of state proposed by Peng-Robinson and single-phase friction factor. The partial derivative of the flow model, necessary for the construction of the Jacobian matrix state of propagation and measurement, are obtained numerically. The state vector consists of parameters to be estimated and boundary conditions where there is direct measuring are considered as control vector, which reduces the state vector size and mathematical complexity involved. The other flow properties are defined only by the flow model across the flow profile. Results are compared for three months field data in the pipeline, which is subject to frequent pigging operations. These results are compared to different state estimation. The results show reduced systematic error between simulated and measured at the Kalman filter application for state estimation and reduction of deviation between calculated and measured values when using Kalman filter to estimate parameters when compared to conventional methods. This best fit and smaller deviation leads to better accuracy in calculate intermediate properties and predicting behavior, possibility a reducing in tolerances applied to the operating limits. Future developments of this work are application in more complex gas flow network and evaluation of the concept for synthetic data with known values for the parameters to be estimated and leakage.
Keywords
Kalman Filter, gas pipeline, offshore technology, Uncertainties, estimation
