A Computational Study on Adaptive Multiobjective Optimization of Blowout Preventer Valve System

Abstract A blowout preventer (BOP) is a large valve that acts as a critical well control unit against oil, gas, and water overflow and well blowout. The BOP valve body assembly contains the main pressure-containing parts. For example, in deepwater or coiled tubing drilling operations for oil and gas, the BOP can be subjected to sustained high pressure cycles. These cycles may cause the concentration of stresses in some regions, which could result in cracks or fatigue damage. Once the BOP fails during the emergency shut-in process, a blowout accident can occur, which may lead to casualties and major economic losses. To ensure the structural integrity under service loads, the BOP valve body should be designed to fulfill the stress requirements as defined by the American Petroleum Institute (API) Specification 16A, 17G and ASME (American Society of Mechanical Engineers) Boiler and Pressure Vessel Code Section VIII, Division 2. It is necessary to ensure that the BOP valve can safely withstand the working pressure. However, overengineering the BOP design is undesirable as well in terms of value engineering because it wastes both time and money. Therefore, there is a strong need to efficiently explore the full design space and identify the optimal solution that results in a compromise between the competing objectives under the constraints for the BOP valve design. In this paper, a methodology is presented for integrating computer-aided design (CAD) nonlinear finite element analysis (FEA), and optimization package to enable adaptive multiobjective optimization and virtual verification and validation (V&V) for the BOP valve system. The BOP valve assembly is parameterized before being imported into an FEA solver. The FEA workflow is automated such that once a set of geometric parameters are given, the preprocessing, solving, and postprocessing steps can be completed automatically. Based on the automated FEA workflow with advanced optimizer (adaptive metamodel of optimal prognosis (AMOP) in Ansys™ OptiSLang), the BOP valve body design space is efficiently explored. The Pareto-optimal design points with respect to two completing design objectives on structural integrity and weight reduction are discovered. This process had never been done in the past. The optimal design set recommended from the optimization algorithm is finally sent to the three-dimensional nonlinear FEA solver for virtual V&V per the API Spec and ASME code compliance. The automated design optimization approach demonstrated in this paper could considerably enhance the product performance and reliability and, in the meantime, minimize product development cost.