Saudi Aramco operates several electrostatic coalescers for bulk emulsion separation and crude desalting. One of the major challenges in operating electrostatic coalescers is the potential buildup of tight emulsions and a rag layer at the interface layer, which causes short-circuiting of the electrostatic grids which increases the risk excessive carryover of water with the crude. Conventional liquid level instrumentation cannot measure the thickness of emulsion layers since the level taps are at the clean oil and water layers. Consequently, the buildup of emulsions is normally not detected by operators. A capacitance-based emulsion detection system was installed at one of the electrostatic coalescers of a Saudi Aramco facility. The system is comprised of multiple probes installed at various elevations in the vessel. Each probe measures the capacitance of the liquid in which it is immersed in. The data is then transmitted to the DCS, where an algorithm computes the oil/water content. Saudi Aramco developed an enhanced predictive alarm logic and advisory tool using the measured capacitance data so that operations may take preemptive measures to prevent upsets from occurring. The alarm system was tested over an extended period of time and it has shown that it can accurately detect the buildup of emulsions prior to an upset in the electrostatic coalescer. What is unique about the system is that it utilizes a combination of absolute capacitance measurements and capacitance variations in the algorithm. Emulsion buildups are detected by the alarm system hours before a potential upset, providing operators ample time to take preemptive measures such as increasing the demulsifier injection rate, desludging the vessel or lowering the interface level. The system significantly reduced the number of electrostatic coalescer upsets at the facility and crude quality was enhanced. Upon inspection of the probes during shutdowns, no buildup of deposits, which impacts capacitance readings, were found on the probes since a flushing system was installed. The alarm system has been utilized for four years with no major issues. Utilizing the capacitance probes to develop an algorithm for an alarm system is a novel technique to detect emulsion layer buildup hours prior to a potential electrostatic grid upset. Large-scale deployment is more economical as it is more cost-effective than radioactive profilers and is logistically easier to manage.

This paper describes a sophisticated control method (known as "Smart Demulsifier Control") for automatic adjustment of optimum demulsifier chemical injection, which was adopted at several Gas Oil Separation Plants in Saudi Aramco. Demulsifier chemical is a key additive used for oil and water separation. Excessive dosing would lead to increased operational costs while demulsifier under-dosing will impact crude quality due to separation upsets such as desalter grid overload. The system utilizes models that were developed using data analytical tools to optimize demulsifier chemical injection in real time. In a Gas Oil Separation Plant, formation water is first separated in a horizontal three phase separator. The rest of water is removed in electrostatic coalescers to achieve the required crude specification. The amount of demulsifier added can balance the separation within certain limits. There are several operational parameters that affect the separation. Most are not controlled (disturbances) such as crude type, emulsion tightness and temperature. The method uses basic instrumentation measurements to measure the online separation balance, compares the actual data to expected data based on a process model and adjusts demulsifier concentration within an adequate dynamic window using a combination of PID algorithm control loops. The implementation of Smart Demulsifier Control scheme at several GOSPs was proven to be a reliable and robust method to fully automate the addition of this important chemical. The online separation measurement provides a clear visualization of the overall separation balance and it is used to adjust dosages accordingly with PID loops. PID algorithm can be easily adjusted at each plant to ensure the speed of response is adequate. Usual plant disturbances that can affect separation balance such as well alignment changes, feed rate and water cut variations are tackled by the control scheme through live separation monitoring. Operators can also influence the separation with some margin parameters provided in the scheme. The dynamic operating window ensures that the amount of demulsifier is within reasonable limits at all times. The design is parametrized to allow for easy modifications in case of new crude feeds, demulsifier effectiveness formula changes or seasonal changes (summer/winter). This novel scheme has been implemented at nine Gas Oil Separation plants processing Arab Light crude. Based on a complete year of evaluation, a very significant reduction in electrostatic coalescer upsets related to inadequate water separation in the three-phase separator has been achieved along with an average reduction of demulsifier consumption of 15%. A patent for the control scheme has been granted by the United States Patent and Trademark Office. It is unique in its simple approach to optimizing demulsifier dosage while ensuring a smooth process of crude oil emulsion treatment and separation.