Compression and pumping systems are constantly changing infrastructures, with many of the older compressor/pumping stations requiring updates, repairs and inspections to maintain safe and efficient operations. These stations operate over a wide range of pressures, flows, and working fluids under varying environmental conditions. Operating condition factors, as well as original design and materials, can significantly affect corrosion rates, structural integrity, and the flow capability of these compressor/pumping stations. Station equipment can be logically inter-related using failure trees and each critical sub-component be assigned a mean time between failure and failure probability using acceptable industry standards. These individual components are then allowed to interact to determine sub-system, system, and full station level failure probabilities. This type of analysis has historically not been utilized by the oil and gas industry but is common to other industries, such as the aerospace and nuclear power industries.
This paper presents a new, comprehensive, consistent, and effective process for predicting risk, integrity, and reliability of the compressor and pump stations as well as each major subsystem and component within these stations. The model considers predefined “threats” such as mechanical, materials, electrical, third party, environment and external forces, improper maintenance, and operation of all its components; thus, typical failures modes are included in these threats. A semi-quantitative methodology with factored risk indices is applied where weighting factors are used to adjust the model with operational data. These factors are generated from reliability data extracted from the station. Comparisons between the model predictions and the reliability data will allow tuning of the weighting factors. Weighting factors are defined for each of the identified threats. The probability of failure is computed at a component level; however, it can be obtained at any level in the system based upon the specified categorization. The probability of failure is represented as a function of three factors: exposure, resistance, and mitigation, while the consequence of failure is estimated using the same approach based on three factors: receptor, hazard, and reduction.
This predictive risk and failure model has been defined based on international specifications and is consistent with actual operating conditions, capacity planning, and remaining life expectations, while assuring that the stations meet the day-to-day operational demands of the system. The model also is able to predict each individual equipment failure probability within the station systems and provides for easy output of the data in graphical form for proper operating, maintenance, repair, and testing decisions.