Integrated Marine Monitoring Systems (IMMS) and Riser Monitoring Systems (RMS) are designed to help operators reduce operational risk by providing information about the environment and platform responses and riser bending and fatigue to assess degradation, respectively.
In spite of efforts to keep monitoring systems in working condition by following planned maintenance and upgrades, some sensors may fail intermittently or may generate spurious data. Quite often, intervention to repair or to replace a faulty sensor is either difficult or not feasible.
Methods such as Artificial Neural Network and Extended Kalman Filters have been employed and specifically tuned to particular challenges.
Artificial Neural Network
The Artificial Neural Network (ANN) tool estimates parameters that have been confidently measured when all sensors were fully functional. The ANN is trained using representative data of inputs and the expected output. To simulate the failure of each strain sensor, several networks are generated. This produces ranges of data that train the network to test data in these various failure modes. ANN stands out as a robust method for estimating riser tension and bending in the face of sensor failures.
Adaptive Kalman Filter
BMT has deployed and fielded an innovative GPS/INS fusion system which applies adaptive Kalman filtering to the problem of combining GPS and INS data. The Adaptive Kalman Filtering system combines the output of multiple GPS receivers, gyros and an INS into a best estimate of platform position and attitude. In order to estimate false alarms caused by sensor errors, the BMT system adaptively estimates sensor noise levels on a sample-by-sample basis, allowing the Kalman filter to discount measurements from faulty sensors - even those whose errors are not yet high enough to cause a rejection by conventional means. The system has demonstrated significant improvements in rejecting sensor errors and reducing false indications that the platform is off desired position.
BMT's forecasting services include focused predictions for a platform's near-future behavior based exclusively on the prior established relationship between environmental forces and the platform-specific responses. Forecasting spans the intervals from a few hours to a week in advance from the present time.
Vortex Induced Motions (VIM) and Vibrations (VIV)
Vortex induced motions and vibrations are events of cyclic motions of offshore platforms, mooring lines, and risers exposed to a specific ocean current speed and direction. The phenomena cause relatively large lateral motions which have negative impact on production. On a long run, repetitive VIM and VIV events have pronounced impact on the fatigue life of offshore structures. BMT has significant experience in identifying and analyzing VIM/VIV events based on measured data.