At the time the meeting was held-Oct. 11-13, 1994, in Washington D.C.-U.S. refiners were gearing up to produce the new fuel in time for the Dec. 1 deadline.
Under discussion were:
- Blending instrumentation and control systems
- Flow metering
- Complex and simple models
- Gasoline properties.
For details on the format of this renowned meeting, see Part 1 of this series (OGJ, April 24, p. 61).
What instrument/control system changes are refiners making to optimize reformulated gasoline blending?
TREESE:
Unocal is adding on-line analyzers for key blendstocks and for the final blend header, as well as on-line optimization systems, at each of our refineries. All of our blenders are now distributed control system (DCS) or computer-controlled. We have eight to nine specifications to meet in California, not all of them immediately, but within the next year.
For the blend header, we are planning to use Applied Automation gas chromatography for on-line analysis of Rvp, distillation points, benzene, and aromatics. On-line analyzers we considered, but rejected as not yet proven to our satisfaction, are: gas chromatography for olefins, on-line X-ray for sulfur, and near-infrared (NIR) for octane. Obviously, we will reconsider these analyzers in the future as we gain some confidence in them.
For key blend stocks, which comprise 5 out of somewhat over 20 stocks that we are going to be blending, we are going to time-share one set of analyzers monitoring the same properties as the blend. All our analyzers are going to have an on-line reference stock that is used to calibrate the analyzers roughly every hour to hour and a half. I did not mention it, but we do have on-line knock engines for the final blend octane rating.
PORTER:
We can blend reformulated gasoline on our existing equipment, but because there are more variables to control we cannot optimize it as well. At the Pascagoula refinery, we are strongly considering a project to integrate our current blending models and do blend property control within a new instrumentation system to achieve better optimization.
LAUX:
As mentioned in the previous question, we have several new on-line analyzers installed that we believe will help us optimize our reformulated gasoline (RFG) blending. We have also installed a new Honeywell blend package with these analyzers and have purchased an RFG blending linear program (LP) (Wright Killen's Wkblend).
But, we have to master the techniques needed to use the LP, especially at the blender's level. It takes experience and technical ability to master these complicated programs. Our blenders are not at that level yet.
It will take some time to develop the intuition necessary to make changes to a blend recipe with all the new specifications. Since the RFG laws are very restrictive and the penalties for any noncompliance are high, we do not intend to start using these techniques immediately when we start blending RFG.
ARMBRESTER:
As I mentioned previously, we are using on-line analyzers to determine octane, vapor pressure, ASTM D86 distillation, aromatics, olefins, benzene and oxygenate values. The NIR analyzer that we are using has been in service for about 4 years at St. Paul and 2 years at Catlettsburg.
In the Catlettsburg blending operation, we are using Honeywell's Blend Ratio Control software in conjunction with a multivariable critical property control program developed by Oil Systems.
The current properties of the gasoline blender's product are used to calculate the optimum blend recipe for the remainder of the blend, using another Oil Systems product known as Mgblend. We have also installed a higher level optimizer to perform linear analysis of the refinery's product demand and gasoline component availability, and then set blend schedules and recipes which will maximize profit.
The simple and complex RFG models are now being added to these software packages in preparation for RFG blending. In addition, when we get to complex model blending, we will need to add an on-line sulfur analyzer.
I might also mention that at St. Paul Park we are using the Honeywell Blend Ratio Control, but some of the higher level software packages in use at Catlettsburg are not being implemented at St. Paul Park since the blending operation is considerably less complex.
BOYCOTT:
Although we are in a conventional gasoline area, we have made significant efforts to improve the control and optimization of our blending. They are inclusive of the NIR analyzer already discussed, installation of a new blending system inclusive of hardware modifications to eliminate component bottlenecks, and increased emphasis on consistent unit rundown parameters.
The result has been as much as a 75% reduction in Rvp and octane giveaway.
D'AURIA:
With the advent of advanced on-line analytical techniques as well as the availability of powerful control schemes, refiners are revamping their blenders with the latest technology. As regulatory demands increase, in both fuels production and quality documentation requirements, refiners have turned to the NIR analytical technique for its capabilities, and to advanced blend ratio and blend property control systems for real time blending optimization.
DAN HARTLEY (HAVERLY SYSTEMS INC.):
The blending systems that I see going in vary from refinery to refinery, but mostly involve software programs which optimize the recipe off-line. Several of these programs are available in the marketplace.
Some of the software programs that optimize the recipe will also optimize the blend schedule. With all the specifications and regulations that are coming along, it is almost impossible to do these things by hand, so I think this is where everybody is heading.
FRANK J. KLEINSCHRODT (SETPOINT INC.):
Major changes include two areas: new analyzers to measure RFG and regulated conventional properties, and reallocation of tankage and piping in off-site areas to isolate conventional and RFG. A major change in control system configuration and operation is required to optimally segregate reformatted RFG and conventional gasolines.
For an in-line gasoline blending system, is there a flowmeter which can measure flow independent of stream composition so that several different fluids can be blended and measured through one meter without recalibration?
CUNEO:
For in-line gasoline blending, we recommend the use of turbine meters for the total flow after blending of all of the components. We are also installing turbine meters on the oxygenates and specific gravity analyzers on the total gasoline blended to calculate the weight percent oxygen.
ARMBRESTER:
Both turbine meters and vortex meters work well for in-line gasoline blending service. The most important requirement for an in-line blender meter is the ability to provide accurate measurement over a wide range of flow rates. Since the different materials being measured are all gasoline streams, switching components from one meter to another is not a problem.
Modern control software, such as Honeywell's Blend Ratio Control, has the sophistication to allow the operator to assign a specific blending component and its associated properties to a given flow meter and then perform temperature correction based on the API gravity of the material flowing through the meter.
TREESE:
Our experience has been primarily using turbine flow meters, which have good "rangeability." They are relatively insensitive to stream composition, or at least the composition differences among the stocks that we are blending is consistent enough that it does not make a significant difference to the meter.
FRANK J. KLEINSCHRODT (SETPOINT INC.):
The question needs to be separated into two subjects: component flowmeters and custody transfer meters.
Almost all the refineries in the world we have studied and in which we have done projects run many different components through a flowmeter without any recalibration. Most motor gasoline blending installations use turbines as primary elements. They are viscosity-sensitive, but the viscosity differences between motor gasoline components are negligible for measurement accuracy.
Vortexes are gaining popularity as primary elements; they have to be operated above a threshold Reynolds number. This is usually accomplished with minimum fluid velocity, not recalibrating the primary element.
The meters used to deliver motor gasoline (mogas), or any finished product, for custody purposes is a different matter altogether. In this case, any significant stream composition changes would have to be accounted for if conventional meters are used.
Some of the new coriales-technology-based meters should have a wider range of applicability without recalibration, but we have not had commercial experience with those meters at this time.
What changes in unit operations will result from the reformulated gasoline complex model vs. operations under the EPA simple model? Arc any refiners planning to use the Complex Model prior to 1998 and if so, how will fungibility be handled?
D'AURIA:
The simple model correlates gasoline emissions with Rvp, oxygen, benzene, and total aromatics. In addition, there are caps placed on olefins, sulfur, and T90, based on the refinery's baseline levels. For the complex model, olefins, sulfur, E200, and E300 are added to the previous parameters in the simple model equations.
When you compare the simple and the complex models, the two parameters that stand out are sulfur and olefins. Sulfur plays a significant role in toxics and NOx emissions under the complex model. Olefins also play a significant role in NOx emissions, especially at olefin concentrations above 15 vol % in the gasoline.
Changes in unit operations resulting from going from the simple model to the complex model clearly will be those that control gasoline sulfur and olefin content. In the case of sulfur, this means fluid catalytic cracking (FCC) feedstock hydrotreating, or heavy FCC gasoline sulfur reduction. Reducing gasoline olefin content can be achieved by removing the C5 olefins via alkylation or tertiary amyl methyl ether (TAME) production.
Looking further down the line, control of sulfur and olefins will become even more important when you get into the Phase 2 regulations, in the year 2000.
Another difference between the simple and complex models is that the trade off between benzene and aromatics is about 50% greater for the complex model than it is for the simple model. So as you move into the complex model, you have the opportunity to go to higher total aromatics levels in the gasoline, at a given benzene concentration.
We would not expect most refiners to use the complex model prior to 1998, primarily due to the cost and the complexity of segregating complex model gasoline from simple model gasoline.
ARMBRESTER:
We agree with Mr. D'Auria's comments regarding both sulfur and olefin content being the major concerns. On sulfur level of the gasoline, we also believe that we will either need to reduce the FCC gasoline end point or hydrotreat the FCC gasoline stream for sulfur removal.
The olefin content of the gasoline will also be an issue, and as Mr. D'Auria mentioned, we are looking at reducing the Rvp of the FCC gasoline stream and using the C5s as feedstock to a TAME unit. That has the added benefit of increasing alkylate and/or methyl tertiary butyl ether (MTBE) production, both of which are desirable blending components for RFG.
We do not plan to use the complex model prior to 1998 due to the fungibility issue. If we were to go ahead and implement complex model blending, that would require our exchange partners to do so also.
PORTER:
We are not aware of any refiner who is going to be able to use the complex model because of what was just mentioned, even though some refiners may find it easier to meet the complex model than the simple model requirements.
It really depends on what your baseline was back in 1990 vs. the industry baseline on these key parameters. Some refiners may not have to do anything. However, if your refinery has changed considerably since 1990 and sulfur and olefin levels are higher, you may have to make some changes.
SAYLES:
I agree with the other presenters that the simple model will be used at least for the first 2 years; the complex model is required in 1998 unless something changes.
The key difference here, for those of us who are conventional gasoline producers, is what we have to do to make gasoline that will allow us to take advantage of some of the breaks built into the complex model. Some of the options are adding ether production, maybe C5 alkylation, and benzene reduction.
TREESE:
Echoing the comments of the other panelists, we are going to use the EPA simple model for RFG compliance. In our case, we have a few different factors perhaps that weigh in the decision. A primary issue for us is that common carrier pipelines in California will not accept complex model gasoline until 1998. Again, there are also the exchange issues.
Also, we will be subject to the California Air Resources Board (CARB) specifications in addition to the EPA'S. The CARB regulations are more strict, setting specific values for around eight to nine properties, as opposed to EPA's more performance-based specifications.
We cannot take advantage of the breaks that might be inherent in the complex model. The CARB specifications are going to control our production.
FRANK J. KLEINSCHRODT (SETPOINT INC.):
I would like to extend the question just a little bit and ask about impacts in the off site areas as a result of the reformulated fuels requirements.
It is clear that there are going to be some additional grades of gasoline and additional components. There is a new RFG before oxygenate blendstock (RBOB) component coming along into the common carrier system. I know that there are going to be some tankage impacts, but I have not heard a lot of talk about big investments being made yet.
Are people able to make the necessary corrections without spending a lot of money at this point? If so, is it going to affect their storage and product disposition flexibility?
LAUX:
One thing we are doing is adding some blenders at our marketing terminals so we can blend "plus" from the supreme and the unleaded gasoline, and that frees up a tank. We are avoiding building tanks that way.
ARMBRESTER:
I think tankage will be an issue, but perhaps a greater issue will be logistics. If we are trying to increase the number of grades of gasoline without significantly increasing tankage, then it is going to require much closer coordination of gasoline shipments and sales.
With the advent of new fuels mandated by EPA and some states, does anyone have tangible evidence concerning effects this will have on gasoline properties such as corrosivity, additive compatibility, polymerization, etc.?
CUNEO:
We have made test blends of EPA gasoline and future CARB gasoline. They passed all the required tests. We are in the process of developing California additive certification for 1995 for RFG.
D'AURIA:
We have not seen any published data on the changes in properties such as corrosivity or additive compatibility.
The earlier corrosion problems associated with use of alcohols have been mostly eliminated, both by improvements in the alcohol production processes and by upgrading the vehicle fuel system materials. The proposed reduction in olefin content could help with respect to gum or sediment formation.
BARLOW:
We have some preliminary data that support a possible increase in corrosivity in several oxygenated gasoline blends. We are not prepared to say it is conclusive, and there have been a lot of changes made, so we are basically monitoring and continuing testing.
As far as diesel fuels are concerned, we think the low-sulfur fuel will exhibit better stability properties based on the hydrotreating efforts required to reduce the sulfur content.
We are also taking a look at the corrosivity effect of those fluids due to the removal of some trace polar compounds that act as natural corrosion inhibitors. There may be an increase, but, again, we are still doing a lot of test work.
FELTROP:
Use of alcohols, such as ethanol or methanol, as oxygenates tends to increase the corrosivity of the gasoline. This problem has been effectively addressed by treating fuel-grade alcohol with the corrosion inhibitor selected for this purpose prior to gasoline blending.
Because some conventional fuel additive formulations are not compatible with the lower level alcohols, additives for direct use in the alcohol-based alternative fuels like M85 and E85 [blends of 85% alcohol and 15% gasoline] must be specially formulated.
At high oxygenate levels, fewer compatibility problems arise since most deposit control additives are soluble in the blended fuel, and alcohol-soluble corrosion inhibitors have been developed for use directly in the alcohol.
SAYLES:
I agree.
What precautions are being taken to prevent water contamination of MTBE during marine shipment? How do you deal with wet MTBE to prevent resulting haze in finished gasoline?
PORTER:
We have not had any problems that have resulted from actual marine shipment. At Pascagoula, to avoid problems, we have added dome roofs to our MTBE floating roof tanks and we inspect barges to ensure they are water free before loading.
As far as dealing with wet MTBE, we have limited the percentage in our gasoline blends to avoid a haze in the final product. We are aware of a distillation system that will remove soluble water from wet MTBE, but we have not used it.
TREESE:
I do not have any comments necessarily on the marine shipment. Marine shipment has not been a big problem for us.
We do get our MTBE wetter by collecting water from rain. Wet MTBE can be blended off in gasoline and still be bright and clear up to reasonable levels of water in the MTBE, under maybe 2,500 ppm. We now store all our neat MTBE in covered tanks.
FELTROP:
Wet MTBE presents a unique haze problem in gasoline blending. Up to 15,000 ppm of water is soluble in MTBE, and MTBE with water content of 2,000 ppm or greater is likely to produce a hazy gasoline blend.
The resulting hazy blends usually require several hours to clarify by static settling. Dehazers can be effective in clarifying these hazy blends within minutes to a few hours.
A more effective approach has been found where the MTBE is known to be wet. In this case, the dehazer is added to and mixed with the MTBE before blending with the gasoline.
This approach either prevents haze in the first instance or greatly increases the rate of clarification. In addition to faster responses, lower doses of dehazer are typically required in the prevention method.
BARLOW:
just to reiterate what has been said, we think the primary problem is poor housekeeping. The resolution to the primary problem is through improved housekeeping procedures.
If you have a problem after that, chemical dehazers have been proven to be effective. We have used 50 ppm of a dehazer to eliminate a persistent water haze (around 0.1 vol %) from an 85% gasoline/15% MTBE mixture,
DENNIS GRANDLE (AGAR CORP.):
We now have several installations where we are able to measure the concentration of water in MTBE using energy absorption microwave technology. We are able to measure 0.1% now, even though the water is soluble in the MTBE at low concentrations.
Copyright 1995 Oil & Gas Journal. All Rights Reserved.
