A pipeline optimization system combines modeling techniques, economic analysis, optimization methods and data management into a package which can improve the operating economics and effective management of a pipeline system. Pipeline modeling techniques provide the foundation; the model used must be able to accurately predict operating costs and limitations imposed by operating constraints over the entire feasible range of pipeline operation. Economic factors provide objectives or goals that the operation should strive to meet. Optimization methods provide a means to integrate economic goals and operating constraints with a mathematical model of a pipeline to provide enhanced insight for effective pipeline operation.
A pipeline model is a set of coherent mathematical equations or correlations which can be used to predict pressure profiles, flowing capacities, pump performance and power requirements given boundary values for injection/delivery flows and pressures, fluid descriptions, pump status, and parameters which describe the physical pipeline. Pumps are usually modeled using a quadratic or cubic polynomial representation of the manufacturer's published performance curves for head and efficiency versus flow. Such programs have been used extensively for the To provide meaningful insight into pipeline operations, a pipeline model must be calibrated to match observed pipeline performance and validated against measured data where available. Calibration can be accomplished by adjusting pipe roughness values until predicted hydraulic profiles match the observed data. Once calibrated, the model must be reliable over the feasible range of pipeline operation, which means it should take account of all variables which materially affect flowing capacity and economics. Most time-transient phenomena (such as pressure surges due to pump starts and stops) are automatically handled by pipeline instrumentation and controls, and do not significantly impact on the overall operating economics. Where longer term transient effects are important (such as gradual packing and unpacking of compressible liquids, or shifting batch locations), assumption of "quasi-steady-state" behavior over some finite time period permits the steady-state techniques to be adapted to the modeling of the actual pipeline operation.
Before optimization can be applied to a given pipeline operation, it is necessary to understand the pipeline's operating economics. Thus a good starting point for implementation of an optimization system is to examine the feasibility of reducing operating costs while maintaining typical operating constraints for pressure limits and volumes shipped. The major cost item (besides personnel) for most liquid pipelines is the electric power consumed for pumping. Power rates vary widely, and each utility company has its own unique contract structure. However, power costs can be broken down into two general categories:
unit energy costs for electricity consumed in kilowatt-hours, and
demand costs or "ratchets" which are tied only to the maximum rate of consumption over any 15 or 3Ø minute time period.
Because of the wide variation in utility power contract terms, it is important to draw the distinction between economic savings and "power" savings -- in many cases operation at minimum power consumption does not result in the minimum operating cost.