Proper design of separators for surface production facilities is essential in order to maintain the quality and quantity of produced hydrocarbon fluids and avoid major operational problems in downstream equipment. Separators that are not properly sized or built will invariably be involved in operational mishaps encountered in natural gas surface operations and processing. This study addresses the basis of two-phase separator design and selection and explores the applicability of Artificial Intelligence techniques, such as Artificial Neural Networks (ANNs), for the creation of intelligent systems capable of predicting proper two-phase separator dimensions which can guide their selection. The expert system is able to unveil the most accurate mapping among input parameters and output sizing parameters and greatly facilitates identifying which parameters have the most influence and/or govern separator design and selection, while quantifying their relative influence. The proposed system is robust, fast, dependable, and unambiguously quantifies the relevance and impact that each fluid property and process conditions has on the correct selection of separation devices needed for natural gas processing applications.
All Science Journal Classification (ASJC) codes
- Energy Engineering and Power Technology