Dec-2021

Optimising critical level control in desalters by using multiphase level measurements

Refineries are a complex set of various operating units designed to process crude oil into refined products such as gasoline, diesel, jet fuels and feedstocks for thousands of different products that people use every day.

Sabrina Nees and David Williams
Berthold Technologies GmbH & Co. KG

Article Summary

In order to maximise operating margins, refiners rely on pushing the units to maximum throughput. To do this safely and effectively, they must have very good control of the process. Some processes in the refinery industry are a challenge for even the best and most experienced engineers. As Lord Kelvin once said, “If you cannot measure it, you cannot improve it.” and H. James Harrington expanding that thought to say “Measurement is the first step that leads to control and eventually to improvement. If you can’t measure something, you can’t understand it. If you can’t understand it, you can’t control it. If you can’t control it, you can’t improve it.” These statements apply directly to today’s refinery industry where in order to maximise margins, a plant must be able to produce more with existing equipment. To safely and efficiently control your process, you need to be able to reliably measure various parameters throughout the refinery. One of these processes that can greatly influence the refineries profit margins is the desalter.

With all the other processes in the refinery, the desalter is one that is often overlooked since it is a relatively simple process. Its main purpose is to separate the salts and minerals from the oil, so they do not get carried over into the other operating parts of the refinery. You do not need to be a corrosion expert to understand that salts are very harmful to metals in the processing units. These salts and minerals can cause other issues besides corrosion such as fouling of heat exchangers and furnace tubes, deactivation of downstream catalysts and decrease furnace efficiencies. When trying to operate a desalter, one must constantly evaluate different parameters, each with a specific effect on the performance of the desalter. These parameters are, for example, mix value setting, wash water (rate and quality), chemical feed (type and rate), operating temperatures and electrostatic grid voltages. While trying to control these parameters to operate the desalter in such a manner as to maximise the effectiveness of the vessel, one must be able to maintain control of the level so that the level does not get to high or too low. In the meantime, countries and local regulatory bodies are also applying strict demands to the quality of the brine/water leaving the desalters, giving operators more challenges.

Optimise existing desalters
How can we get the most out of our existing desalter? We can ensure that the grids are working to their fullest and that the maximum volume is being utilised in the vessel to increase residence time, ensuring maxi- mum desalting of the oil (Lobo, Kremer, & Cornelius, 2010, slide 23). To ensure that we can achieve this greatest efficiency, control of the interface level becomes crucial. When we talk about interface level, this is a misnomer, since there isn’t really a defined level between the crude oil and water/brine. There is instead a transition zone where the fluid slowly changes from crude to water/brine. This transition zone is both undefined and under certain circumstances can vary greatly in height (emulsion thickness, see Fig. 2).

When interface level is checked by sampling, you will see a variation in amount of oil and water mixed together at different elevations. Many refiners use sampling as their primary method of interface level control in the desalter, due to the lack of confidence in the current technology used. Various technologies have different issues with reliably measuring and thereby controlling the interface level. As refiners try to meet various national, local, and corporate standards, the sampling method itself causes safety risks and environmental impact concerns. What is done with the fluid that must be flushed from the sample lines? With environmental regulations, this oily water has to be collected and treated. How long do you flush the lines? The time it takes to flush line can lead to more fluids that must be collected and treated as well as errors if the line is not flushed long enough. How to limit human error? One person might determine the sample to be all oil whereas another person might determine the sample to be an emulsion. Then, there are safety concerns, these include raising the temperature to 300°F/150°C to assist with the separation of the water and salts from the oil. Therefore, sample lines must have coolers to prevent the potential burns to the person taking the sample and with refiners integrating petrochemicals, more benzene can be present causing an inhalation issue.

In the market today with all the opportunity crudes that are available, it is important to know how the different crude oils react when mixed together to ensure compatibility. If oils are mixed that are not compatible, they could form emulsions that are very difficult to break (Garrett, Rattanakhambay, Robbins, Wunder, & Yeung, 2016, p. 1). When this happens in the desalter, knowing not only the brine water level is important, but also the top of the emulsion layer to ensure proper desalter operation. There are several technologies that are used to measure the water level in the desalter:

­  differential pressure (DP)
­  RF absorption probes
­  nuclear density profile systems

These can be placed into two basic categories: direct and indirect measurement technologies, with all of them falling into the direct measurement category except the nuclear density profile systems. The direct measurement devices measure a physical characteristic of the fluids (dielectric constant (dK), capacitance, weight, etc.) by directly contacting process fluids. Each one of these technologies have their advantages and disadvantages. These technologies work very well when there is a defined interface between the two fluids, but due to the operational characteristics of the desalter, there might not be a defined interface but a graduation of densities as the fluids separate. When this graduation of densities occur, the nuclear density profile system is the best option. The nuclear option will provide the user with a density profile of the desalter so that the levels of water and emulsion can be determined, monitored and controlled. In this paper we will briefly explain how DP, RF absorption probes and nuclear density profiler systems measure the inter- face level in the desalter and some of the issues that each technology may have.

Differential Pressure
DP is an economical way of measuring levels in tanks and vessels and it is the second most common use of pressure measurement behind flow measurements. DP levels use Pascal law to determine level by relating the level, pressure and density of the fluid, simply stated, P = ρ * h, where P = pressure (pascals, Pa), ρ = density of fluid (kg/m³) and h = height of the column of fluid (meters, m). You can solve for h by rearranging Pascal law, h = P / ρ (Meribout, Al Naamany & Al Busaidi, 2011, p. 2). For example, a tank with 2 m of water will exert a pressure of 2000 mmH2O or 0.196 bar. If the fluid was kerosene with a density of 0.82 spg, the pressure exerted will be 1640 mmH2O or 0.161 bar. As you can see the density of the fluid has a great influence on the pressure. If this was a level application calibrated on water to measure a 2 m span and the process fluid changed to kerosene with a density of 0.82 spg, the error of the level output at 50% level would be 9%, or the output would read 41% but actual level would be 50%.

The errors are more complicated for DP levels when trying to measure interface levels in vessels especially if emulsions are present. The difference in pressure being measured is the difference in pressure exerted by the vessel full of the higher density fluid and the vessel full of the lower density fluid. Since two fluids are being measured, the error can be compounded if both fluid densities change. Another point that can cause issue with the DP level measurement is the height of the emulsion. Figure 2 (top) shows level with a relatively small transition zone and Figure 2 (below) shows a level with a larger transition zone. Emulsion is not the true word to use in the case of the desalter since emulsion implies a fluid of consistent density, but here we use this term to describe the transition zone from clear oil to clear brine. This zone is actually a gradient change in density from the lighter density fluid to the heavier density fluid. The DP level will only measure the average density across the full span of measurement and cannot tell the operator if the level output is the bottom of this emulsion, the middle or top of the emulsion height.