At the most recent National Petroleum Refiners Association question and answer session on refining and petrochemical technology, refiners discussed, among other things, general processing and support operations.
Emphasized subjects included corrosion, fouling, and foaming problems. Several processing areas were discussed, such as crude desalting and distillation, asphaltene stripping, and sulfur-removal operations.
The 1993 NPRA Q&A conference was held Oct. 20-22 in Dallas. For details on the proceedings see OGI, Mar. 28, p. 41.
CRUDE TANKS
What has been refiners' experience in treating crude tanks to improve subsequent operations in the desalters and the rest of the unit's equipment?
Laabs: One of our refineries typically uses a demulsifier in tankage on crudes that are difficult to desalt. To date, they have not had any problems with this chemical injection. Other sites have reported that using a demulsifier in crude tankage has improved desalter operations, but are in the process of quantifying that benefit.
Identification of crudes that cause desalting and dewatering problems is an ongoing process. Several crudes from Egypt and Venezuela have caused difficulties in the past.
Another of our affiliated refineries does not have a desalting system, and must rely on tank farm dehydration. There has been some success with their program, but their crude unit feed does have a higher chloride concentration than that of our other refineries. We have found that certain combinations of crudes will limit the effectiveness of the injected demulsifier at the location.
Murphy. Many refineries have had positive experiences in treating crude in tanks. The program is usually implemented by refiners with limited desalter capacity, or those who want to remove unusually high salt levels of the crude before it is sent to the crude unit.
This program is also used by refiners whose crude feed has high volumes of water and those who observe significant fluctuations in the volume of water in the raw crude. These programs can be effective at preventing upsets in the desalter and crude tower due to water slugs.
Tank dehydration should be considered when: The raw crude feed from barges, tankers, or pipelines is inconsistent; when the raw crude contains greater than 0.5 vol % brine; when the desired target for salt content of the desalted crude is unattainable with the current equipment; or when the refiner is experiencing poor oil/water separation in the tank farm.
To develop an effective program, the following items should be completed: conduct a total survey of the tank farm system; evaluate the various feedstocks to the tanks; establish specifications for the program; write procedures for the application; and thoroughly train the operators on the program.
We have seen successful treatments of numerous domestic and international crudes from 18 to API 40 API using this method. Chemical rates are 3-12 ppm in these applications, and times in tankage vary from 24 to 48 hr. Heavier crudes may require heat to speed the process. The technical representative can accurately predict the conditions that will be required for the treatment by performing a fairly simple bench-scale test.
Osborn: During periods of processing high-bs&w crude (above 0.5 wt %) we begin chemical treatment in the crude storage and piping system prior to the desalters. The process involves adding the chemicals to water-wet the solids and remove them with the water.
Removing water in tankage will also significantly reduce the salts in the raw crude. This has been very successful in minimizing periods of desalter upsets, thus preventing water carryover, solids carryover, oily effluent, and chlorides in the overhead systems.
Paules: We have found that treating a crude oil tank to remove entrained water is generally not very cost-effective. We have attempted this on occasion and have found that it cost as much as 21/times as much to treat a tank chemically to remove water as it does to process the material through the desalter.
Also, under the recent National emissions standards for hazardous air pollutants regulations, crude tank bottoms are potential benzene-emission sources and it is generally more cost effective to control the emissions at the desalter.
Generally, the amount of water actually entering crude tankage is relatively small. And if it can be handled on a ratable basis to the desalter, little, if any, detrimental effects will be seen.
Kenneth N. Abrahams (Star Enterprise): Mr. Paules said that you can handle the water ratably through your desalter. Is it still with the crude?
Usually, you run into problems if you open a tank and have higher concentrations at the bottom. You tend to get a slug of water through your system. Do you have a specific way of handling that?
Paules: At one refinery, we make a concerted effort to keep the water mixed in with the crude oil with mechanical mixers; and in another refinery, we have actually installed a system to pump out tank water bottoms ratably back into the crude.
Scott Bieber (BPCI/Chem-Link): We have successfully treated crude oil in storage tanks in two U.S. refineries processing heavy crudes. Our main goal in these cases was to reduce the amount of desalter oil carry under due to entrainment of oily solids from the desalter interface.
We were able to significantly reduce the amount of oil carry under in both of these refineries. One plant saw their tailwater oil content fall to less than 0.5%, compared to 2 5% being experienced before the crude oil pretreatment program was initiated. The second refinery saw their oil carry-under problem disappear altogether.
Additional benefits measured at the second plant included improved atmospheric tower overhead system pH control and improved corrosion control throughout the crude unit.
Muhammad T. Uthman (Saudi Arabian Oil Co.): In one of our refineries, we let the crude settle in the tanks after desalting. We experienced good results just settling the crudes for 24 hr.
ASPHALT PUMPS
Asphalt product pumps are difficult to specify due to the presence of dissolved or entrained solvent and/or seal flush oil. What type of pumps are best for this application?
T. Williams: In our residual oil solvent extraction (ROSE) unit, we use Viking gear pumps, Model LQ225. We pipe these in parallel with common spares and put as many pumps in service as needed to match the production volume. These pumps are packed instead of having flush seals. The packing usually has a 6-month to 1 year life.
In other services we have also used Worthington pumps, Model 2GRJ.
Sloan: Rotary positive displacement, or "screw type," pumps have been used successfully in asphalt product service. This type of pump can handle a wide range of high viscosity and dirty fluids. However, this type of pump may experience capacity limitations from reduction in viscosity if seal flush is not accounted for in the specification. Double mechanical seals may be used to eliminate this problem. If double mechanical seals are not used, an estimate of the minimum fluid viscosity accounting for seal flush should be made for the specification. Small amounts of dissolved gases or solvent should have little effect on the performance of the pump if adequate NPSH is available.
Proper product stripper design will minimize the amount of solvent present in the asphaltene product.
Laabs: We are currently using screw pumps for this application. The pumps are run at slow speeds to avoid problems. You can also use gear pumps, also run at slow speeds.
O'Brien: Our asphalt product pumps are centrifugal pumps. On a start up there is frequently foam in the asphalt, so the pumps have difficulty taking suction. The pumps must then be bled down, making a considerable mess. Replacing them with gear pumps (positive displacement) was considered, but could not be justified.
ASPHALTENE STRIPPING
What is your experience with the use of a cutter stock in the feed to the asphaltene product stripper of solvent deasphalting units to solve foaming problems? Can foaming in deasphalted oil (DAO), resin, and asphaltene strippers be reduced or eliminated by controlling operating conditions such as temperature, pressure, and/or stripping steam rate? What antifoaming agents have been successfully employed for foaming control in these strippers and what typical dosages are required?
Juno: At one of our refineries we have never experienced foaming problems in the resin stripper. The resin recovery consists of sending the resin propane mix of the contactors through a furnace to a pressure flash tower and then on to a resin stripper.
We have experienced foaming problems in the flash tower. This is detectable by using the blotter test at a vent valve at the inlet of the condenser for the resin flash tower vapors.
The flash tower contains a demister pad and it is found that the demister pad loses its ability to entrain liquid if the viscosity of the entrained liquid is too high. Therefore, we maintain a minimum furnace outlet temperature.
Murphy: Foaming severity is related to temperature and pressure. For a given pressure, a lower temperature will make the oil more likely to foam. We have employed antifoaming agents to control foaming very successfully at several locations.
At one refinery, we injected an antifoaming agent into the asphaltene stripper feed. We observed a dramatic drop in foaming, demonstrated by significantly reduced plugging of the overhead fin fans due to decreased asphaltene carryover. This program was then repeated at the deasphalted oil tower and sin War results were observed. Unit upsets due to foaming were eliminated after these programs were started.
Silicon based antifoams diluted with the gas oil carrier normally will be effective at controlling foaming. The dosages are dependent upon the severity of the foaming and the method of application.
O'Brien: We have had problems with the level indication in the asphalt tower, probably caused by a foam layer in the tower. We have found the best thing to do is to allow the tower to go empty, thus yielding the foam to the asphalt stripper, and then to reestablish the level.
The float column used for level indication was also replaced with a differential-pressure level indicator, which is calibrated by finding the actual level with tri-cocks.
Paules: We have never added cutter stock in the feed to our asphalt stripper. However, we have occasionally used a silicone based antifoam. The dosage rate was approximately 0.5 gal/1,000 bbl of asphalt product.
We have the capability of lowering the asphalt stripper pressure to less than 5 psig, which helps reduce the surface tension of the solvent, thereby reducing foaming.
Sloan: According to Kerr-McGee Corp., no one is injecting cutter stock into the feed to a ROSE asphaltene stripper to reduce foaming. One ROSE unit fed heavy cycle oil (HCO) to the top tray of the asphaltene stripper in an attempt to reduce asphaltene carry over to the solvent condenser. They reported that the HCO helped reduce carry over but the light ends of the HCO was stripped and contaminated the solvent.
Low stripper operating temperatures can promote the formation of foam. Product forming within the stripper can sometimes be reduced by increasing the stripper temperature. At the higher temperature, the viscosity, of the product is reduced, and lower viscosities mean that any foam that forms will break much easier.
Increasing the stripper pressure or reducing the stripping steam rate lowers the amount of solvent removed from the product. Although solvent losses are higher, the tendency for foaming is reduced because less solvent will flash from the products. In addition, steam or hot oil coals could be added below the bottom tray to help break foaming.
A two product ROSE unit is successfully using an antifoam agent in both their product strippers. They use a low silicone antifoam agent, such as Nalco 5700, at the rate of 1 2 gal of antifoam per 10,000 bbl of product.
CORROSION MONITORING
What new tools are available to monitor corrosion in a crude overhead system? Can underdeposit corrosion be monitored?
Murphy: The trend in corrosion monitoring is toward continuous real time feedback. Corrosion rarely occurs in an even rate and continuous monitoring is the only way to capture all spikes as they occur, thereby assuring a diagnosis of the root cause of the corrosion. The automation of conventional monitoring techniques, as well as the improvement or introduction of some more sophisticated equipment, has been observed.
For example, electrical resistance (ER) probes have been used successfully for many years. However, the frequency of probe readings is often limited due to the inconvenience of manually collecting the data, especially for those probes located in difficult locations such as the top of the crude distillation tower.
To combat this, recently introduced remote data loggers have become popular. These units are designed to attach to ER probes in the field, can be programmed to take readings at various intervals, and can download the information to a computer for analysis.
Combining this technology with a slipstream overhead simulation exchanger provides very good insight into the around the clock factors affecting corrosion at the dew point of water. Additionally, flexible corrosion probes, which are inserted directly into the tower and sit on the tray surface, give excellent localized corrosion information.
Another trend is to select more predictive variables to monitor, including tracking such items as the toal acid content of the stream, the dew point pH, key upstream operating parameters such as salt out of the desalter, and any suspected corrosion cofactors such as the sulfur content.
To the extent that these proactive measurements can be captured continuously, the better the potential corrosion control. A recently introduced on fine strong acid analyzer can measure the actual amount of acids present and adjust the neutralizer pump accordingly.
Considering the question of underdeposit corrosion, we are aware of no direct means to precisely determine the extent of corrosion taking place under all the various deposits which may exist in an overhead system while in operation. Most monitoring techniques only give general information. The corrosion rate under a deposit can be as much as 100 times the general corrosion rate.
There are on line tests or tools which can help to characterize the problem. The slipstream corrosion simulators allow one to visually examine any salts or deposits which might form on any probe. Still the only true indication of the extent of underdeposit corrosion taking place in a unit is the information obtained during turnaround inspections.
Osborn: We have used several of the Nalco products that Mr. Murphy was mentioning. We have used a "flex" probe, installed on the second tray of a crude tower, with good results in monitoring for a potential corrosion problem. Routinely the sen,ice representative uses the strong acid analyzer and we also use in line pH meters on our overhead system.
Brierley: Syncrude uses both standard and flush-mounted electrical resistance probes with electronic data loggers to collect corrosion-rate data. There is also an inexpensive, externally mounted, hydrogen patch-type monitor that is mounted using an epoxy adhesive.
While it does not provide a quantitative corrosion rate, it is sensitive to process changes that affect the corrosion rate. The fact that it is externally mounted without welding or extensive surface preparation makes it very versatile. Neither of these tools can differentiate the type of corrosion.
A costly method of monitoring electrochemical noise by real time, advanced linear polarization system has recently become available as a field tool, but we have yet to have field experience with it. It will differentiate the types of corrosion and would be able to detect underdeposit corrosion.
Pedersen: Regardless of the systems available on the market, we still believe that a good pH control and follow up on the iron content in the sour water will continue to be essential.
We have tried on line corrosion monitoring at our installations with a varied degree of success. Of course, as has been mentioned earlier, one of the problems with these systems is to locate the probes where the corrosion rates are highest, or at least representative, e.g., in the condensing zone in the overhead condensers.
Another problem is underdeposit corrosion, which we monitor on stream by radiographic or ultrasonic testing. We inspect tubes,. for instance, in air fin condensers, by fiber optics off stream.
Rajguru: Just to add to the panel's comments, the flexible corrosion probes are the ones that we are very familiar with and those have been added recently.
Roy: One other tool that we are familiar with is the Betz Process Chemicals Inc. variable temperature corrosion Probe. It employs a hollow coupon that can be cooled by an appropriate fluid.
You install the coupon well above the suspected condensation point that you are trying to monitor. The coupon can then be cooled to a downstream temperature region by a fluid of choice.
Maintaining the temperature for a period of time will allow for a deposit to be collected and analyzed. This method tries to simulate the point of condensation to measure maximum corrosion rate.
CLAUS DEBOTTLENECKING
An increase in operating pressure has been investigated as a possible alternatie to debottlenecking a Claus unit. What modifications are needed? What is the experience with compressors in H2S service? What limits the maximum operating pressure of a Claus unit, provided that the liquid seals have been modified to withstand higher pressure?
Laabs: Recently, one of our affiliated plants modified its sulfur recovery unit (SRU) trains to operate at a higher pressure. The modifications required them to increase the length of the seal leg by 5 ft, giving it the ability to withstand a pressure of 12 psig. The need to keep the amine unit regenerator bottoms temperature below its corrosion limit of 260' F. required the replacement of the acid gas fine. This new line provided a lower pressure drop.
The waste heat reboilers were changed from 600 psig to 250 psig steam production to avoid problems with temperature constraints. An additional quench cooler was installed in the tailgas treating unit to further cool the feed to the booster blower. The lower temperature resulted in a lower actual volume of gas.
The modifications to these units were made so that they could be run at 8 psig, but the closest approach since start up has been 7.2 psig, due to problems at the waste water stripper. The problem is being successfully minimized by using pressure control at the stripper, so the limiting factor is expected to be the combustion air discharge pressure.
Osborn: Assuming properly designed diplegs, probably the only thing stopping us from running at higher pressure is that our blower capacity is limited, and in fact we are using oxygen enrichment. Also, the temperature of the reboiler vapor line on the still increases to a point that corrosion in the amine unit is a problem.
It is important to keep this below 265' F., above which corrosion starts to increase significantly. Our seal legs allow us to operate up to roughly 15 psig. The condenser capacity would probably be the next bottleneck. I do know there are totally steam jacketed blowers, and we do use a blower in our tail gas unit.
Sloan: Operating the Claus unit at higher pressures enables the product water, as well as the sulfur, to be condensed, thereby driving the Claus reaction further to completion. Equipment sizes within the SRU and downstream tail gas treating unit are also reduced. The increase in recovery exists despite the fact that higher pressure adversely affects the reaction furnace and the catalytic beds.
The reasons why high-pressure operation is still being investigated and is not yet commercialized are: That reaction furnace temperatures tend to be at or above the maximum allowable for existing refractory materials; the catalyst modifications, either in design or in use, to address the problem of increased sulfur vapor pressures are still being explored; and issues related to corrosion, safety, and the use of enriched air or oxygen in high pressure operation are still being researched.
Acid gas compressors have been used in sulfur-recovery unit revamps to increase capacity, and when a tail gas unit has been added. The service is difficult because of potential problems with wet acid gas corrosion sour water condensation and sealing requirements The performance of these compressors has been mixed, and if high pressure Claus operation is to be successful, it will help if the reliability of the acid gas compressors can be improved.
The maximum pressure in the Claus unit is limited by the design temperatures of refractory materials and the need to maximize the use of carbon steel equipment. At a given pressure, operating temperatures depend on acid gas feed composition.
Present research indicates that plants employing both oxygen enrichment and high pressure Claus operation at pressures as low as 70 psig can result in dramatic increases in recovery and reduction in equipment sizes. From an economic standpoint, we are probably looking at pressures of less than 100 psig to realize the benefits of high pressure Claus operation.
PREHEATER FOULING
What experience is there with plugging and corrosion in air preheaters or furnaces? How are they being treated?
Bonelli: Chloride present in the refinery fuel gas in our operation caused corrosive atmospheres in the furnaces. One method to reduce corrosion in the furnace preheaters, burners, process tube exteriors, and the hangers in the fires boxes is to eliminate the chloride in the hydrogen from the reforming units. We intend to install solid bed activated alumina scrubbers.
A significant problem for us is corrosion that occurs in the stacks, preheater elements, and duct work assemblies, especially in our largest duty heaters. Usually, it is a condensation of sulfur compounds in the flue gas. We have tried various coatings in the refractory appliances and have not been successful in achieving a full run length without some corrosive attack. We have considered operating the furnace inefficiently to keep the stack temperature high enough to prevent condensation, but that obviously is an uneconomical thing to do.
Laabs: All of our refineries have had some sort of plug-gage problems in one or more of their combustion air preheaters. Fouling and corrosion are typically controlled by maintaining an outlet flue gas temperature above a certain target. The target varies from site to site, anywhere from 275 to 350 F.
The air preheater on the atmospheric furnace at one of our refineries uses finned cast iron tubes and Pyrex tubes near the flue gas outlet. Recently, the tube sheets for the glass tubes were changed to titanium to avoid corrosion problems. Sour waste gas is burned in the heater, and they have had bowling ball sized chunks of corrosion products falling off the cast iron tubes and breaking large numbers of glass tubes. Careful monitoring of the flue gas outlet temperature and metallurgy changes have successfully minimized this problem.
Another site has not had plugging problems with their regenerative preheaters. They have seen fouling in their recuperative preheaters when they have had tube failures. They monitor the outlet flue gas temperature to avoid fouling problems and, for existing fouling, they recommend putting the heater into natural draft, blinding the preheater, and washing the flue gas side with hot water. They have also installed a light in the bottom of one preheater to allow visual inspection for plugging. Our third refinery has experienced preheater fouling due to corrosion products in both their regenerative and their recuperative air preheaters. They burn sour waste gas as one of their fuels, They also depend on maintaining the outlet flue gas temperature to avoid plugging.
Osborn: Farmland has a variety of air preheaters in its furnaces and boilers. We have used a Ljungstrom regenerative preheater with a rotating heat wheel, the DEKA gas exchanger containing glass tubes, and more recently the OCAP unit supplied by North Atlantic Technologies, which is a plate type exchanger. We have had minimal fouling in recent years while operating with a cleaner fuel gas. Before that, we did have some fouling problems because of sulfur in the fuel gas. The preheaters are usually equipped with a water or steam washing device. We prefer steam cleaning since it does not make deposits as wet and as corrosive.
The key to good results is to do the washing on a regular basis instead of waiting until there is an excessive pressure drop or heat loss. Once again, stack temperatures are important to be maintained above dew point levels.
Pedersen: We have air preheaters, both rotary and static, on furnaces which primarily are gas fired and have not experienced severe plugging problems. We control corrosion by preheating the air upstream of the air preheater to 175 to 180* F. by other heat sources than flu gas, for example hot oil, steam condensate, or hot water.
