No abstract

The mooring systems give stability to the floating platforms against environmental conditions, stabilizing the platform with mooring lines attached to the seabed. The mooring systems are among the main components that guarantee the safety of the staff and the various operations carried out on the platforms. The current approaches used to monitor mooring lines are inefficient as line tension sensors are expensive to install, maintain, and have durability problems. This article presents the development of two neural networkbased machine learning systems: a Multilayer Perceptron (MLP) and a Long Short-Term Memory (LSTM). They are able to detect mooring line failure in near real-time based on the comparison between measured and predicted motion. The implemented systems were trained and evaluated with simulated motion data generated using real environmental conditions measured in the Campos Basin, in Rio de Janeiro, Brazil. The results showed the MLP and LSTM models were able to detect a failure in the mooring lines, with increasing difference between the predicted and the measured motions when there is a line breakage. A comparison between the two machine learning models revealed the LSTM model performed better at predicting the motions of the platform. INDEX TERMSMooring line failure, failure detection, machine learning, neural networks, floating production storage and offloading.

The focus of this work is the extension of nonlinear state estimation methods to gas-lifted systems. The extended Kalman filter (EKF), unscented Kalman filter (UKF) and particle filter (PF) were used to estimate the nonlinear states. Brief descriptions of the filters were first presented starting from the linear Kalman filter. Hypothesis tests on the expectation of the residuals were performed to show how close to optimal the estimation methods are and it showed the UKF estimates to be slightly better than EKF while PF performs the worst. The PF has poor accuracy using residual visualisation, hypothesis test and the root mean squared error (RMSE) values of the residuals. The gas-lifted system exhibits casing heading instability where the states show oscillatory behaviour depending on the value of the input but the results here do not change in a known way for each filter as the input is changed from the non-oscillatory region to the oscillatory region. Therefore, for this noise distribution and model assumption, either the EKF or UKF can be used for nonlinear state estimation with UKF better preferred if computational cost is not considered when control solutions are used in gas-lifted system.

A failure in the mooring line of a platform, if not detected quickly, can cause a riser system failure, extended production downtime, or even environmental damages. Therefore, integrity management and timely detection of mooring failure for floating platforms are critical. In this paper, we propose a new model for an ANN-based mooring failure detection system. The proposal’s idea is to train a Multilayer Perceptron (MLP) to estimate the platform’s future motion based on its motion’s temporal data without failure. A classifier then indicates whether or not there is a failure in the mooring system based on the difference between the predicted and the measured motion. The results with several tests of the implemented system show that our proposal can correctly predict the motion of the platform in most environmental conditions. The system shows a precision, accuracy and F1-score of 99.88%, 99.99% and 99.94%, respectively, for detecting changes in platform motion in near real-time, quickly signaling a possible breakage of mooring lines.

This paper proposes a solution based on Multi-Layer Perceptron (MLP) to predict the offset of the center of gravity of an offshore platform. It also performs a comparative study with three optimization algorithms – Random Search, Simulated Annealing, and Bayesian Optimization (BO) – to find the best MLP architecture. Although BO obtained the best architecture in the shortest time, ablation studies developed in this paper with hyperparameters of the optimization process showed that the result is sensitive to them and deserves attention in the Neural Architecture Search process.