Design Considerations For Pressure Swing Carbon Adsorption Vapor Recovery Systems
by W. Nicholson Tuttle, P.E. and Jon W. Young, P.E.
Contents
BASIC VAPOR RECOVERY SYSTEM MODULES
Vacuum regenerated activated carbon vapor recovery units typically consist of two major operational modules;
-
The activated carbon vessels, and
-
The regeneration skid.
Activated Carbon Bed Vessels
There are at least two identical activated carbon bed vessels as part of each vapor recovery unit. One bed is "on-line" adsorbing hydrocarbons vapors generated by the loading operations while the other carbon bed vessel is being vacuum regenerated. The carbon beds are cylindrical, vertical, carbon steel, ASME coded pressure vessels designed for full vacuum and are supported by either a skirt or legs. Except for the smaller VRUs, where the beds may be mounted on the skid, the carbon bed vessels are usually installed adjacent to the regeneration skid on a concrete pad. Since all vapor phase activated carbon is somewhat friable it can be damaged so the vessels are shipped empty and the activated carbon is loaded into the vessels after they have been installed at the job site.
Regeneration Skid
Mounted on the regeneration skid is all of the equipment necessary to regenerate each activated carbon bed vessels once each cycle during normal operation. This equipment consists of the sequencing valves, the vacuum system, the absorber column (or condenser), and product circulation pumps. The control system consisting of a PLC mounted in an enclosure, the motor starters and inlet power distribution box are usually mounted on the skid but can be installed at a remote location. Other items found on the skid are the local instruments, pressure and temperature indicators, etc., fittings, valves and piping. If the vacuum system utilizes a liquid ring vacuum pump then the following items are also included on the skid; the seal fluid pump, the seal fluid cooler, and a separator vessel.
DESIGN CONSIDERATIONS
Loading Profiles:
The loading profile of a specific location is used by the equipment manufacturer to size the two modules of a vapor recovery system. The manufacturer’s experience and understanding of the loading operation will determine how the values associated with the loading profile are utilized.
Different Types of Loading Operations
Truck Loading Terminals Without Vapor Holders
For a typical truck terminal the loading profile will include the following:
-
Qi -instantaneous loading rate
-
Qcycle -maximum loading which can occur during a cycle (to be defined)
-
Q1 -maximum one hour loading rate
-
Q4 -maximum four hour loading rate
-
Qdaily -maximum daily loading rate
The dimensional units for the instantaneous loading rate (Qi) can be gallons per minute, cubic meters per hour, etc., any volumetric rate per a given time interval. Make sure that what ever units are used is communicated to the manufacturer. For the other rates in the loading profile, only volumetric units are used, gallons, cubic feet, cubic meters, etc., since the time interval in inherent to the definition of that data point.
In the United States most truck’s cargo trailers range in capacity from around 8,000 to 11,000 gallons. The trailers are divided into smaller volumes, usually four to eight compartments per trailer. The two main reasons for the compartmentalization of the trailer is for safety in transporting liquids and the compartments allow each cargo trailer to carry multiple products such as regular, mid-grade, and premium grade gasolines, diesel fuel, or various chemicals without cross contamination.
When loading operations are underway the loader will usually connect no more than three loading hoses (also known as loading arms) to the cargo trailer. For truck loading operations (without vapor holders) the instantaneous vapor flow to the vapor recovery unit is a direct result of the liquid displacement of the cargo compartments and varies almost moment to moment. The actual flow of products into each of the trailer’s cargo compartments occur over a very short time interval and is directly related to the product flow rate per loading hose and the size of each compartment. Once a compartment is full the loader can move that hose to another compartment and continue loading until all compartments are loaded.
Qi - Instantaneous Flow Rate-
Typically the loading rate per hose ranges from 450 gpm up to 1000 gpm. In most truck loading operations three hoses are hooked-up to a cargo trailer at any given time, so the flow rate to a given trailer can range from 1350 gpm to 3000 gpm (3 times the instantaneous rate per hose). If more than one truck is being loaded at a given time, that is, there are multiple loading spots or bays, then this flow rate can be multiplied by the possible number of trucks that could be loaded simultaneously to yield the maximum instantaneous loading rate.
For example, if the a truck loading operation has four (4) loading bays, four trucks can be loaded simultaneously, and three (3) product hoses are connected at each bay at a rate of 500 gpm per hose, then the maximum instantaneous flow rate will be 4 X 3 X 500 = 6,000 gpm.
Qcycle - Max. Loading During A VRU Cycle-
The cycle time is the length of time that a carbon adsorber is left on-stream to receive vapors from the loading operation. It is also the length of time that the other carbon adsorber is under regeneration plus the time needed to switch from one carbon adsorber to the other. The cycle time of the VRU is usually chosen between 12 to 17 minutes.
The cycle time must also take into consideration how the loading operation at the terminal is actually conducted. Choosing too short a cycle time may not allow for the correct regeneration of the activated carbon. Too long a cycle time will lead to the need for more activated carbon.
A good rule of thumb is that the VRU’s cycle time for truck loading operations is around 15 minutes and the value of Qcycle (Q15) is equal to the maximum volume of product that is loaded in a quarter of an hour. In the United States Qcycle is equal to the cargo trailer volume times the number of trucks that can be loaded simultaneously. In the U.S. the time for truck "turn-around" in the terminal is approximately 15 minutes, so for many applications Qcycle is equal to Q15.
For most applications the Q15 is not equal to Qi times 15 minutes. This is because there are a number of start and stop loading operations occurring for each cargo trailer depending upon the number of compartments in the trailer and the number of product hoses connected to the trailer.
For example, if a truck terminal has four (4) loading bays (see above example), Qi equals 6,000 gpm, and each truck has a capacity of 9,500 gallons then the Q15 would equal 9,500 X 4= 38,000 gallons (NOT 6,000 gpm X 15 = 90,000 gallons).
Q1 - Max. One Hour Loading Rate-
This value is the maximum volume of product that is loaded on-board the cargo trailers for any one hour period. The maximum this value can be four (4) times the Q15 rate.
Q4 - Max. Four Hour Loading Rate-
At most truck loading facilities there is a peak four hour period when a maximum number of trucks are loaded. This four hour period is usually in the mornings when the terminal is opened for truck traffic. If the terminal is a 24 hour terminal then the peak four hour period is again in the morning starting around 6 A.M. The maximum value this can be is four (4) times the Q1 rate. However, for most operations Q4 is between two (2) and three and a half (3.5) times the Q1 value.
Qdaily - Max. Daily Loading Rate-
This value is the maximum volume of product that is loaded in a twenty-four (24) hour period. Typically for most truck terminals without vapor holders this value is between two (2) and four (4) times the four hour rate.
Continuous Loading: - Truck Terminals With Vapor Holders and Marine Terminals
For this type of terminal the loading operation is considered continuous since the loading rate does not vary widely with time. Once the loading has commenced the flow rate is relatively constant during the loading episode.
Qi - Instantaneous Flow Rate - Truck Loading with Vapor Holder
The Qi for a truck loading operation with a vapor holder is determine somewhat differently than the Qi for a truck loading operation without a vapor holder. Associated with the vapor holder will be a blower, either continuous or variable speed, this blower takes suction from the vapor holder and conveys the vapors to the vapor recovery unit. For this type of operation the Qi is the maximum displacement of the blower, the maximum feed rate to the vapor recovery unit.
Typical values range from 200 to 600 cubic feet per minute (cfm), depending upon the terminal throughput capacity and the vapor holder volume.
Qi - Instantaneous Flow Rate - Marine Loading Operations
The Qi for a marine loading operation is measured just as the Qi for a truck loading operation WITHOUT a vapor holder. Qi for marine operations is the maximum rate loading can occur per loading arm times the number of loading arms in operation simultaneously. Loading rates can vary from 350 barrels per hour for a single berth loading operation upwards to greater than 25,000 barrels per hour for large multiple berth operations.
Loading Profile Interactions on the Sizing of Vapor Recovery Units
Qcycle
First, the volume of product loaded or the volume of vapor flow to the activated carbon bed and the maximum inlet hydrocarbon vapor concentration is used to calculate the amount of activated carbon, per bed, required to meet the emission limits.
Qi
Second, the instantaneous flow rate is used to calculate the pressure drop through the vapor inlet piping, inlet vapor valves, carbon bed and the vent piping and valves. The geometry of the beds, diameter & height, and the piping & valve diameters are selected.
Q1
The one hour loading rate does not directly influence any of the sizing formulas used for the VRUs. This value is used primarily as a check.
Q4
Third, Q4 along with the carbon bed sizing and the Qdaily rates are used to determine the vacuum pump capacity required to regenerate the activated carbon so as to meet the emission requirements.
Qdaily
The Qdaily rate is used along with the Q4 rate and the carbon bed sizing to determine the vacuum pump capacity requirements to insure proper regeneration of the activated carbon.
Cycle Times
During the cycle time of an activated carbon VRU two major steps must be accomplished. First, the bed that is in service to the loading operation must receive and adsorb the vapors and at the same time the other bed must be regenerated to a point where it can receive vapors from the loading operation the next cycle. Secondly, the beds must switch. That is, the bed on-stream will be put under regeneration and the bed that was being regenerated must be put on-stream.
To better understand this we must consider the various steps or processes that occur during each cycle. These individual steps or processes must be "viewed" from the stand point of what is happening each cycle. There are three basic steps that occur during each cycle. These steps can be described as the operational modes and they are: 1) the on-stream or adsorption mode, 2) the regeneration or desorption mode, and 3) the equalization or switching mode.
To view this, please see Sketch #1, we will label one carbon adsorption bed "A" and the second bed "B". In this example we will assume that the system has been running for some period and that bed "A" is on-stream receiving vapors from the loading operation.
Adsorption Mode
In the adsorption mode the valves are set so that bed "A" is receiving vapors from the loading operation. The air/hydrocarbon mixture from the loading operation passes into bed "A" and the hydrocarbon is selectively adsorbed from this mixture by the activated carbon. The air that was brought into bed "A" in the mixed inlet stream is vented to the atmosphere. At the same time that this is happening, bed "B" is either in its regeneration mode or in the equalization mode.
Regeneration Mode
With bed "A" on-stream in the adsorption mode we will start with bed "B" in its regeneration mode. The valves are set so that Bed "B" is receiving no vapors from the loading operation and the valve connected to the vacuum system of the regeneration skid is open. The vacuum system evacuates bed "B" to an absolute pressure of 74mm mercury ( approximately 90.3% of absolute vacuum). The time required to evacuate the bed to this pressure is dependent upon the vessel volume relative to the vacuum pump capacity and the amount of hydrocarbon loading experienced by the bed during its previous on-stream time.
After a preset time a small purge air solenoid valve is automatically opened by the control logic allowing a small amount of air into bed "B" while it is still under the deep vacuum. This purge air aids in the regeneration of the activated carbon. The desorption of activated carbon will be discussed later in this paper. Once the regeneration timer has timed out it is time for bed "B" to start the process, Equalization Mode, which will place it on-stream and bed "A" in the Regeneration Mode.
Equalization Mode
With bed "A" still on-stream receiving vapors from the loading facilities the Regeneration Valve associated with bed "B" closes, isolating this bed. Since bed "B" was under vacuum when the Regeneration Valve closed it must be brought to atmospheric pressure before it is switched to the on-stream service. During the equalization mode the bed "B"is equalized to atmospheric pressure by opening a special Re-pressurization Valve or by slightly opening the Vent Valve, both of which allow for atmospheric air to gradually bleed into the bed "B" breaking the vacuum. After a preset time the Vent Valve is fully opened and bed "B" is equalized to the atmosphere. At that time the Inlet Valve opens placing bed "B" on-stream. When the Inlet Valve on bed "B" is opening the Inlet Valve on bed "A" is driving closed and the Equalization Mode is completed, bed "B" is now on-stream and bed "A" is beginning its Regeneration Mode. On bed "A", the Vent Valve is then driven closed and the Regeneration Valve associated with bed "A" is slowly driven open placing bed "A" fully in the Regeneration Mode.
Sizing the Activated Carbon Beds
When sizing the activated carbon bed there are a number of considerations that must be taken into account. These are: 1) the amount of carbon required to process the vapors generated from the loading operation during an Adsorption Cycle, 2) the proper geometry of the carbon vessels so as to create an acceptable back pressure on the vapor collection piping, and 3) the proper physical and mechanical design of the vessel so as to; (a) withstand the operating conditions, (b) support the activated carbon and prevent its loss due to abrasion, (c) provide means to inject the purge air with proper distribution, and (d) provide means of breaking the vacuum with minimum disturbance to the activated carbon.
Activated Carbon
What is Activated Carbon?
The activated carbon is the heart of this type of vapor recovery system. Vapor phase activated carbon, the type of carbon used in VRUs, is very interesting material and exhibits the following characteristics:
1. It is very porous material consisting of numerous small openings and channels throughout a particle of the material. It is approximately 60 to 70% void and one handful of this material has the total surface area almost equal to that of a football field. The size and distribution of the pores is critical to an activated carbon’s effectiveness in a given application.
Pores are classified into three (3) general categories; Macro, Meso, and Micro pores. Macropores are in the size range of greater than 500 Angstroms in diameter, Mesopores are greater than 20 but less than 500 Angstroms in diameter, and Micropores are less than 20 Angstroms in diameter. (Angstrom is a unit of length equal to 3.937 X 10-9 inches. A human hair has an average diameter of 500,000 Angstroms). For hydrocarbons normally found in gasoline vapors the distribution of pores should be in the meso and micro range, 15 to 50 Angstroms.
2. Vapor Phase Activated Carbon has little affinity for air, nitrogen or most inert gases. It has a slight affinity for CO2 and a higher affinity for hydrocarbon molecules. As the molecular weight and molecule size increase so does the affinity of the activated carbon.
3. Adsorption is a surface phenomenon. Activated carbon adsorbs hydrocarbon molecules because an imbalance of forces, Van der Waal’s forces, are present on the carbon surface. This imbalance is satisfied by the physical attraction of other molecules onto the carbon surface. These attractive forces are similar, in strength, to the forces which produce surface tension, they are rather weak forces. Hydrocarbon molecules are attracted to the activated carbon surface and in so doing are separated from the carrier gas.
4. Regeneration (desorption) of the activated carbon is accomplished by over coming the forces which attract the hydrocarbon molecules to the surface of the activated carbon. This can be done by three basic methods, they are:
-
Thermal Regeneration. Heat, usually in the form of a hot inert gas or steam, is used to "boil-off" the hydrocarbon molecules from the surface of the activated carbon.
-
Purge Regeneration. Relatively large volumes of air or inert gas are passed over the activated carbon and the hydrocarbon is "evaporated" into this purge gas stream.
-
Pressure swing Regeneration: The pressure of the activated carbon bed is reduced from the pressure at which adsorption occurred. The hydrocarbon is "flashed" off the surface of the activated carbon.
For the activated carbon based VRUs in service at petroleum product terminals the regeneration is accomplished by two of the above steps. First Pressure Swing Regeneration is used by pulling a deep vacuum on the activated carbon bed. Then the second step of the regeneration is utilized, Purge Regeneration. While the bed is at the deep vacuum a few standard volume of ambient air is injected into the vessel. At the deep vacuum this small amount of purge air expands to a relatively large volume of purge air.
5. Vapor phase carbons are manufactured from a variety of carbonaceous materials including various grades of coal, soft and hard woods, coconut shells, peach pits, etc. The variations between carbons made from these different raw materials are, generally, variations characteristic to the structure of the raw material itself. These differences usually determine whether or not an activated carbon made from a specific raw material will be acceptable for certain applications. For example, activated carbon made from coconut shells will perform quite well for gas masks and similar applications but for the type of service expected of it in petroleum product terminal VRUs, coconut shell carbon will fail miserably in a short time.
Some of the characteristic features that must be considered before choosing an activated carbon for a specific application are:
1. Retentivity; how easily the activated carbon releases the hydrocarbon molecules. Generally, coconut shell activated carbon has a higher retentivity than does wood or coal based activated carbons and this is one of the reasons that it works well in gas masks.
2. Working capacity; the mass of hydrocarbon the activated carbon can adsorb under operational conditions of temperature, pressure, and the conditions which the carbon will be regenerated. Working capacity is usually measured in terms of mass of hydrocarbon per mass of activated carbon, grams/gram, pounds/pound, etc. or mass of hydrocarbon per volume of carbon, pounds/cubic foot, etc.
3. Resistance to abrasion and dusting; Some activated carbons are very soft and easily break down during a unit’s normal cycling between atmospheric pressure and the vacuum required for regeneration. Other activated carbons are so hard that they are brittle and friable and they too break down. Experience with different types of activated carbon will differentiate between those equipment manufacturers which have the knowledge of the characteristics best suited for terminal operations.
How Much Carbon is Required?
The amount of activated carbon furnished with the VRU is a function of the vapor flow rate from the loading operation during a cycle of the VRU, the hydrocarbon concentration of the inlet vapor stream, and a characteristic of the activated carbon known as its working capacity. An algorithm must be devised to take this input data into account and arrive at a mass of activated carbon that will process the hydrocarbon vapors generated by the transfer operation at the terminal loading operations.
Activated Carbon Calculations:
The following is used to calculate the theoretical mass of activated carbon required (In this calculation standard U.S. engineering units will be used.):
Required Data:
-
Loading rate per cycle (gallons)
-
Inlet vapor concentration as a fraction. For most motor fuel applications this value is around 0.40 ( 40% hydrocarbon concentration).
-
Molecular weight of the hydrocarbon fraction of the inlet vapor. For gasoline vapors this value is between 60 and 70 (pounds /pound mol).
-
Working capacity of the specific activated carbon to be used in the VRU. The dimensions are mass per mass (grams of hydrocarbon / gram of activated carbon, or pound/pound, etc.)
Formula:
Theoretical Mass of Activated Carbon Required (ACtheoretical) =
Discussion:
This formula is used to calculate the theoretical amount of activated carbon required for a VRU application. In actual practice there are a number of other "FACTORS" which must be included in this calculation to determine the actual amount of activated carbon supplied. These factors include the growth factor, the recycle factor, a safety factor and other factors.
Definitions:
The Growth Factor takes into consideration the fact that a gallon of liquid hydrocarbon loaded into a transport vehicle actually displaces more than a gallon of vapor.
The Recycle Factor takes into consideration the fact that there are some non-absorbed or condensed vapors coming from the Absorber Column that must be adsorbed onto the activated carbon.
The Safety Factor takes into consideration the fact that in the real world perfect distribution of the vapors through the activated carbon beds does not exist.
Other Factors are applied depending upon the level of emissions required or allowed to be vented from the system. Another factor which must be accounted for is the fact that activated carbon is shipped in discrete sized containers and the purchase of whole containers is the normal practice. The amount of activated carbon purchased is rounded up to the nearest whole number of containers.
The total value for these Factors ranges between 1.2 and 1.4 and includes allowances for those factors listed above.
"K" is the value for the number of cubic feet of vapor per pound-mol of ideal gas. At sea level and 60°F, "K" = 379.
The value of the variable labeled "Working Capacity" is dependent upon the type of activated carbon being used and the hydrocarbon concentration of the inlet vapor. Typically vapor phase activated carbon has similar volumetric working capacities that is, approximately the same mass of hydrocarbon can be adsorbed per cubic foot of activated carbon.
Working capacity is normally calculated in terms of mass of hydrocarbon adsorbed per mass of activated carbon (pounds/pound) so the density of each specific type of activated carbon must be taken into consideration when calculating its Volumetric Working Capacity.
Example:
For an example, one 8,500 gallon truck cargo trailer is being loaded with gasoline, the MW of the vapor is 65 and the hydrocarbon concentration is 40%. The emission limit is 35 milligrams of HC in the vent per liter of product loaded (mg/l). Ambient temperature is 60° F and the inlet vapor temperature is also 60° F.
To calculate the amount of activated carbon required this example will use two different types of acceptable vapour phase activated carbon:
A. Carbon I has a density of 23 pounds/ft3. The mass working capacity is 0.051( lb of HC / lb of AC).
B. Carbon II has a density of 15 pounds/ft3. The mass working capacity is 0.080 ( lb of HC / lb of AC).
Replacing values in the formula we obtain:
Carbon I : [(8,500 / 7.48) * 0.4 * (65 / 379) * 1.2 * 2]/0.051 = 3,669 pounds AC
or Volume Required: 3,669 lbs / 23 lb/ft3= 160 ft3 Activated Carbon I
Carbon II: [(8,500/ 7.48) * 0.4 * (65 / 379) * 1.2 * 2]/0.080 = 2,338 pounds AC
or Volume Required: 2,338 lbs / 15 lb/ft3 = 156 ft3 Activated Carbon II
As can be seen, the volumetric amounts of the two types of vapor phase activated carbon is approximately equal.
Carbon Bed Vessel Design
The carbon bed vessel design and sizing is a function of the amount of activated carbon specified by the Carbon Required Formula, for a given application and the maximum allowable pressure drop through the bed at the loading facilities peak instantaneous loading rate. The maximum allowable pressure drop through the entire vapor gathering circuit, the VRU being part of that circuit, is dependent upon the maximum permissible back pressure the transport vehicle can withstand prior to activation of its over pressure system or the activation of the pressure safety valve located on the vapor collection piping.
Pressure Drop Considerations
In order to keep the pressure drop through the vapor circuit to an acceptable level and compatible with the vapor collection system and the pressure relief devices on the cargo trailers and the vapor collection piping, the equipment designer has only three variables that can be manipulated. The carbon bed vessel diameter, the vapor piping & valve diameter, and the particle size of the activated carbon. The first two are rather self explanatory, with a given mass of activated carbon and a maximum instantaneous flow rate from the loading facility, the larger the diameter of the vessel and the piping the lower the pressure drop. A given mass of activated carbon it can be contained in a small diameter tall vessel, which would have a relatively high pressure drop or a larger diameter vessel that would be some what shorter and would have a lesser pressure drop. The piping will offer less resistance to flow a given flow rate if the diameter is larger.
For a given amount and depth of activated carbon, the larger the particulate size the lower the pressure drop through the bed. If there is only one particulate size available to the designer then they are limited to only two variables that they can manipulated for pressure drop constraints and are locked into larger bed and piping sizes for many flow rates.
Flow Pattern through the Carbon Bed (down-flow vs. up-flow)
At the very low flow velocities (superficial velocities of less than 20 feet per minute) found in these applications any advantages claimed for having the vapor flow in a downward direction rather than in an upward direction is non-existent. The potential problems claimed to be addressed by having the vapor flow in the downward direction are bed fluidization (fluffing), compaction of the bed or mechanical attrition (dusting of the carbon). Any real advantage in flow direction is dependent upon how the carbon is supported in the vessel and how the air is directed from the vessel, in other words, the flow pattern through the vessel. This will be discussed under Carbon Bed Vessel support design, the next topic.
Bed fluidization or "fluffing" problems occur when the bed superficial velocity is too high and if cycled too many times manifests itself in attrition of the carbon particles into dust. Dusting then leads to a higher than expected pressure drop through the bed. Bed fluidization or "fluffing"; occurs when the solid particles, activated carbon in this case, are suspended by the passage of a flowing gaseous stream. If the superficial velocity through the bed is great enough then the bed may become fluidized. The superficial velocity at which bed fluidization starts to occur is dependent upon the solid particle size, the density of the solid, and the amount of bed height above a specific particle, the deeper into a bed the less likely that fluidization will take place. For vapor phase activated carbons the velocity value is within the range of 30 to 120 feet per minute depending upon the activated carbon utilized. At the low superficial velocities found in properly designed carbon beds dusting due to fluidization or "fluffing" is not a problem.
There is one potential problem that exists with a down flow bed and that is a phenomenon known as diffusion. If a bed is idle for several hours or longer, even though it has been regenerated, there is the potential for some of the hydrocarbons to diffuse from the activated carbon. Since the hydrocarbons are heavier than air gravity will pull them down and they will tend to settle in the bottom of the vessel. With the flow pattern during adsorption through the bed in the downward direction these hydrocarbons may settle near the outlet piping. When that bed is again put into service these hydrocarbons may be vented.
Carbon Bed Vessel Support Design
In a packed bed system which processes vapors, such as activated carbon system, catalyst beds, etc., the bed support device serves not only to support the bed, as its name implies, but to aid in the distribution of vapors across the cross sectional surface of the bed. A number of devices have been utilized throughout the years to support packed beds and include loose gravel, specially sized alumina packing material, perforated plate, fine screen wire mesh, wedge wire and other devices. Most of these devices have served the one intent of supporting the bed but few have also successfully aided in the distribution of the vapors. Of the devices listed only the fine screen wire mesh when coupled with an "open" vessel head design have successfully served as both a support of the media and a distribution device.
A properly designed bed support and distribution system will insure a minimum of vapor channeling through the media in the bed and maximum contact utilization of the material in the bed. A poorly designed bed support and distribution system, on the other hand, will insure that there will be channeling through the bed and less than maximum contact utilization of the material in the bed.
On the opposite end of the packed bed is the outlet, and this too must be properly designed for good flow distribution. In a down flow system the media is usually packed into the vessel head and around the outlet nozzle which is in the vessel bottom, see Sketch #2. In this case the vapors flowing downward are channeled towards the outlet nozzle, they take the path of least resistance, leaving a sizable quantity of media under utilized. In an upward flow system the top head is usually left "open" and as the vapor leaves the media there is a smooth transition to the vapor outlet nozzle.
Regeneration Skid
When designing the Regeneration Skid there are a number of considerations which must be taken into account. Some of these are: 1) the loading pattern experienced by the petroleum products loading facility and the amount of hydrocarbon that is to be adsorbed by the carbon bed, 2) the size of the activated carbon beds and the removal of hydrocarbon from the bed during the limited regeneration time, and 3) a safe and adequate means of removal of the hydrocarbon for the activated carbon allowing that bed to return on-stream and be within acceptable emission limits.
It must be remembered by the designer that the sizing of almost every major component on the Regeneration Skid is directly dependent on the vacuum pump capacity. That is, once a Regeneration Skid is selected for say a vacuum pump with a capacity of 500 acfm, one cannot simply increase the vacuum pump capacity to handle a larger loading profile. Everything down stream of the vacuum pump should be sized to handle the 500 acfm flow from the vacuum pump. Seal fluid flow to the vacuum pump varies with vacuum pump size, the heat exchange equipment is based on the seal fluid flow and the duty associated with this flow. The separator is designed to allow for the proper separation time of the seal fluid from the condensed hydrocarbon with minimum back pressure to the vacuum pump. The absorber column is sized to process the hydrocarbon vapor flow and to insure proper absorption with a minimum recycle at a relatively low pressure drop.
Vacuum Pump Considerations
The regeneration of the activated carbon in this type of system is a two step process. The first step consists of placing the carbon to be regenerated under a deep vacuum. For gasoline vapor recovery the vacuum level will be 90.3% of absolute vacuum for both the 35mg/l and 10mg/l systems. (For higher efficiency systems the vacuum level approaches 97% of absolute vacuum). The second step is the injection of a small amount of purge air at the very deep vacuum. The time at which the purge is injected into the vessel is determined either by an elapsed time, a preset timer, or after a preset pressure has been reached.
The deep vacuum level is accomplished with a mechanical vacuum pump and the most commonly used type of vacuum pump for the Vapor Recovery Systems currently in operation today is the Liquid Ring Vacuum Pump. The selection of the capacity of this vacuum pump is dependent upon a number of operational conditions and is complicated by the fact that for this type of system the rate of evacuation of a vessel filled with activated carbon saturated with mixed hydrocarbon vapors is not a " static " process, it is very " dynamic. " That is, when evacuating a void vessel containing only air or when evacuating a vessel filled with virgin activated carbon, the rate of evacuation is fairly predictable and follows the pump down curve provided by the pump manufacturers. However, when the variable of mixed hydrocarbon vapors adsorbed on to activated carbon enters into this equation, then the problem becomes very dynamic. The various species of hydrocarbon molecules are released from the activated carbon at different rates and at different vacuum levels. A rule of thumb is that the lower molecular weight compounds are release at a vacuum level closer to atmospheric pressure while the heavier molecular weight compounds are released at deeper vacuum levels. For gasoline vapor recovery systems the adsorbed hydrocarbons start coming off of the activated carbon some where around 65% of absolute vacuum. As the vacuum level continues to get deeper more and heavier hydrocarbon start being released from the carbon.
The rate and vacuum level at which the various species of hydrocarbons are released from the activated carbon is dependent upon a number of variables, the most crucial being the type of activated carbon being utilized. The other variables include: 1) adsorbate vapor pressures, 2) adsorbate inlet temperature and concentration, 3) pressure during adsorption, and 4) heat of adsorption. Considerable time must be invested in research to obtain ample empirical data to produce a viable algorithm for vacuum pump sizing.
A liquid ring vacuum pump utilizes a liquid sealing fluid inside of the pump case to allow the pump to achieve the moderately deep vacuum levels required in this service. The maximum achievable vacuum level is limited by the vapor pressure of the liquid sealant and the solubility of the hydrocarbon in the seal liquid.
The liquid normally utilized in the liquid ring vacuum pumps is an aqueous mixture of a specially blended ethylene glycol based industrial coolant and water, NOT ENGINE COOLANT or ANTIFREEZE. Since this material is aqueous it provides a level of safety when evacuating air and hydrocarbon mixtures from the activated carbon. The internal surfaces of the liquid ring vacuum pump are constantly being bathed in the seal fluid. If the remote chance of having metal to metal contact due to some catastrophic failure of the vacuum pump should occur then there is little likelihood that there would be a spark.
Seal Fluid Circuit
The major components of the seal fluid circuit consists of the following:
-
The Seal Fluid Cooler
-
The Seal Fluid Pump
-
The Separator
The Separator is on the discharge side of the vacuum pump and is usually part of the absorber vessel. In the separator the liquid seal fluid is allowed to gravity separate from the hydrocarbon liquid which condensed from the vapor in the vacuum pump. The separated seal fluid is then pumped by the Seal Fluid Pump through the Seal Fluid Cooler where the seal fluid is cooled prior to re-injection back into the Liquid Ring Vacuum Pump. The Seal Fluid Cooler is most often a shell and tube type heat exchanger and the seal fluid is pumped through the shell side of the cooler. In some cases the seal fluid cooler may be an air/fan type of cooler, either dry or evaporative.
Other Regeneration Skid Components
Sequencing Valves
The sequencing valves are mounted on the regeneration skid and are part of the control system. Their purpose is the operational switching of the carbon beds. For a two bed system there are three valves required for each carbon bed and they sequence the operations of the carbon vessels from being on-stream receiving vapors to regeneration. Typically these valves are electric motor actuated because there is rarely a source of instrument air at most petroleum product terminals. Air operated valves can be used if a source of clean and dry instrument air or bottled nitrogen is available. Soft seated butterfly valves, to insure a good seal in vacuum service, are used to minimize pressure drop. During the normal operation of a typical Vapor Recovery System these valves will sequence from open to closed between eight and ten times per hour, therefore high quality valves and actuators are essential.
Absorber Column
After separation from the liquid the vapor discharge from the Liquid Ring Vacuum Pump passes into the Absorber Column where it is contacted, counter current, with fresh product liquid from storage which is circulated to the absorber column by the Product Supply Pump. As the liquid cascades down over the packing in the column the vapors are absorbed into this "lean" product. It is in the absorber column that the "recovery" actually takes place and the hydrocarbon vapors are converted back into a liquid for return to product storage. The recovered hydrocarbon and the fresh product which was circulated to the absorber are combined and pumped back to storage, on level control, by the Product Return Pump. The discharge of the Product Return Pump directs the liquid through the tube side of the shell and tube Seal Fluid Cooler. If an air/fan type cooler is used then the liquid is returned directly to storage.
In some applications this Column is utilized as a Direct Contact Condenser rather than an absorber. In these cases the liquid product is circulated through a cooler or chiller prior to injection into the column. As the liquid cascades down through the packing it condenses the vapors rising counter flow. The liquid is collected and pumped on level control to storage as outlined above. This configuration eliminates the need for two liquid lines, to and from storage, since fresh absorbent is not needed.
System's Controls
The major control system components are:
-
The Programmable Logic Controller (PLC)
-
Motor Starters
-
Locally mounted instruments
-
Sequencing Valves (discussed earlier)
As required by code, all control items should be designed for use in a hazardous area, mounted inside an enclosure rated for the hazardous area, or remotely installed in an unclassified area. Hazardous area ratings of Class I, Gp. C or D, Div 1 or 2 are the common area classifications for this service. The PLC and the Motor Starters are usually mounted on a common control panel located on the Regeneration Skid. The PLC is programmed to stage the operation of the Carbon Bed by correctly opening and/or closing the Sequencing Valves on each of the carbon beds. A PLC is particularly well suited for this type of operation and is highly reliable for the safe operation of the system.
The locally mounted instrumentation consist of local pressure and temperature indicators and the following "first out" shutdown devices:
-
High/Low Separator Level
-
High Carbon Bed Temperature
-
Low Seal Fluid Flow
-
Supply/Return Pump Failure
-
Vacuum Pump Failure
-
Emergency Shutdown
-
Sequencing Valve Fault
-
Other "shutdown" devices may be required for specific applications.
PUTTING THE MODULES TOGETHER
Final Design Selection:
If the activated carbon beds and the vacuum regeneration skid have been correctly sized then it is a relatively simple matter of putting these two modules together to form a complete PRESSURE SWING CARBON ADSORPTION VAPOR RECOVERY UNIT. The final steps in completing the design is to insure that the following are met:
-
Adequate Supply and Return Pump NPSHr or discharge requirements to meet the local needs of the specific terminal.
-
Any special design features or requirements of the client. These special features may include API-610 pumps rather than ANSI designed pumps, special electrical requirements, special code requirements for the pressure vessels, special paint requirements, etc.
-
Any optional items that the end user wants to include such as, local area light over the control panel, flame arrestor in the vent piping, inlet knock-out tank, etc.
Once all of the special features and optional equipment items have been selected and included the design and selection process for the PRESSURE SWING CARBON ADSORPTION VAPOR RECOVERY UNIT is complete.
Return to Technical Paper List
