Cenovus Energy has deployed a rigorous multiphase flow assurance online solution to detect leaks and monitor hydrate formation conditions at the White Rose Field and satellite extensions 350 km east of St. John's, Newfoundland and Labrador, Canada. Twenty-five production wells are connected via subsea manifolds to the SeaRose floating production storage and offloading (FPSO) system, through four flexible flowlines and risers.
An online subsea advisor has been developed to provide control-room operators with enhanced monitoring/visibility in detecting potential leaks and hydrate formation within the subsea system, including the mechanical flowline connectors. The online solution makes use of a commercial multiphase flow simulator. An online real-time mode (RTM) was developed to simulate the production loops connecting the manifolds to the SeaRose FPSO. The wells are equipped with multiphase flowmeters, which are calibrated at regular intervals during well test campaigns. Reconciled flow rates from the meters are used as inlet boundary conditions to the real-time multiphase model. This RTM acts as a digital twin of the production network.
As part of the online subsea advisor leak detection system (LDS), Schlumberger has delivered improved algorithms for leak detection in multiphase production networks. The solution makes use of 14 signatures indicating leaks, which form the basis for a generalized multivariable LDS. Artificial intelligence and data clustering are used to determine whether the signature vector indicates a leak. By making use of multiple leak signatures, the system becomes more robust with respect to sensor faults and drift. Multiple signatures also reduce the number of false alarms and make the LDS less dependent on model calibration. The use of signatures, artificial intelligence and data clustering is new compared to traditional mass balance model-based LDS. The theory is described with results from four of these 14 signatures in the paper.
The advisor system monitors the potential for hydrate formation conditions and calculates the hydrate margin at the flowline connectors, which have been identified as potential "cold spots." A rigorous flowline connector model has been implemented at positions along the flow path where they exist in the field. This model is fine-tuned to estimate mechanical flowline connector wall temperatures. This gives the control-room operators a realistic estimate of reaction time to manage an emergency shutdown and initiates an alarm when hydrate conditions will be reached, prompting immediate action of predefined safeguard measures.