When transporting multiple products in a pipeline, it is important to always understand the location of the head and tail of each batch. The operator will know where batch interfaces are in real-time and be ready to swing the valve at the exact time a batch arrives at a station to deliver product to the right storage tank or an end-customer with minimal contamination. It is relatively easy to track multiple batches in a pipeline with no elevation changes, and a fixed internal diameter. However, it is far more complex to track multiple batches in a pipeline with drastic elevation changes and different sizes in diameter. Column separation of the liquid, also known as slack, occurs when the pipeline pressure drops below the vapor pressure calculated at the liquid temperature in a specific area of the pipeline. This effect reduces the amount of liquid volume contained within the pipeline region, changing the actual physical position of the batch head and tail interfaces, and reducing the accuracy of the volume within the region and the corresponding Estimated Times of Arrival at stations. Draining or filling a pipeline section by delivering product at a different rate from what injected causes the same effect.
A scientific approach to calculate the areas of slack and volume contained within a pipeline should provide sufficient information to track batches with a high degree of accuracy. However, it is neither simple nor straightforward to simulate this phenomenon offline, and it is much more challenging in an online, real-time environment. Online, additional complexities affect how batches and their interfaces are tracked, causing a big discrepancy between Estimated and Actual Time of Arrivals. This paper discusses an empirical approach developed to calculate the volume contained within a pipeline region by tracking the volume entering and leaving the region.
Estimated and Actual Times of Arrival are within a 15-minute time window after a batch has traveled a total distance of 1,200km with drastic elevation changes along the route. This method has proven that batch tracking can be highly accurate and reliable with less of the theoretical assumptions used in a hydraulic simulation package, with no need to model every single characteristic of the pipeline in detail. This removed the uncertainties that attend those assumptions and allowed this system to perform well on a pipeline with severe slack flow and draining/filling operations.