Equipment sizing decisions in the oil and gas industry often have to be made based on incomplete data. Often, the exact process conditions are based on numerous assumptions about well performance, market conditions, environmental conditions, and others. Since the ultimate goal is to meet production commitments, the traditional method of addressing this is to use worst case conditions and often adding margins onto these. This will invariably lead to plants that are oversized, in some instances, by large margins. In reality, the operating conditions are very rarely the assumed worst case conditions, however, they are usually more benign most of the time. Plants designed based on worst case conditions, once in operation, will, therefore, usually not operate under optimum conditions, have reduced flexibility, and therefore cause both higher capital and operating expenses.
The authors outline a new probabilistic methodology that provides a framework for more intelligent process-machine designs. A standardized framework using a Monte Carlo simulation and risk analysis is presented that more accurately defines process uncertainty and its impact on machine performance. Case studies are presented that highlight the methodology as applied to critical turbomachinery.