Oil and Gas Processing | |
---|---|
Fluid Flow NPSH Calculations Control Valves Gas Dehydration Gravity Separation Relief Valves Separator Sizing Fractionation Fired Equipment Heat Exchangers Pumps Compressors Refrigeration Compressors Gravity Separation Stripping Gas Curves Gas Dehydration Formulas Engineering Formulas Relief Valve Formulas TEG Vapor Pressure Air Cooler Program Compressor Performance Program Hydraulics Calculation Program TEG Dehydration Valve Program Vessel Volume Calculation Program Linking Fortran DLL to Excel Visual Basic Entrainment Woodworking Boating Gas Processors Association
|
Home | Subjects |Graphs and Tables | About Us | sidecell | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Separator Sizing | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Gas/Liquid Separation
This paper will discuss gas/liquid separation, and will for the most part be focused on vertical separators, where there is a large amount of gas and a relatively low amount of liquid. Much of the discussion which follows is based on rV2 and K. The following defines these parameters: K value This K value is used to define the maximum allowable velocity for a mist elimination system. This has been traditionally used in calculations for mesh pads. The mesh pads are generally designed with K values between 0.15 and 0.35 . The mesh pad is an impact device. The mist is suppose to hit a strand of wire and coalesce with other small droplets to form big droplets, with eventual drainage out of the mesh pad and falling down into the liquid section of the vessel. If the K is too low the mist can dodge the wire mesh strands. If the K becomes too high, then the mesh pad floods. The traditional limit for K was 0.35. However this simple limit is affected by operating pressure, type of fluids being handled and mesh pad construction (very dense mesh pads will be more efficient, but have lower capacity). The formulae for K value is: K=U/√[(rL-rV )/rV ] rV2 rV2 is generally used to size vane units. The allowable values as specified by manufactures has varied considerable. The values have ranged from 20 to 150. The following defines rV2: r = Vapor Density, lb/ft³
The advantage of calculating rV2 using lb/sec is that it makes it clear the dependence on gas density and on flowrate. Once rV2 has been calculated, it can be converted into a K value by the following formulae: K=U/√[(rL-rV )/rV ] Separation methods for gas and liquid can be separated into the following major areas: Gravity SeparationThis technique is generally used for flare ko drums where no separator internals are tolerated. It results in large vessels as the result of low allowable velocities. I don’t plan on discussing this method any further at this time. Centrifugal SeparationThis technique spins the gas in an attempt at causing the denser material (liquid) to separate. This technique results in relatively small vessels. The draw backs are relatively high pressure drop and low turndown capability. Also this technique is not good for small particles. I don’t recall the break point for particle diameter. The “Portatest Separator” is an example of this technique Mesh PadsThis technique utilizes mesh pad . Mesh pads are typically made up of 0.011 inch diameter wire, woven into pads. The pads are stacked on top of each other to a thickness of 6 inches. Grids are placed on both sides of the wire mesh pads and separated by wire rods. This provides some structural integrity to the pad. The pads are woven in different densities for different applications. In the oil and gas industry we typically use a 6 inch thick pad of 9 lb/ft3 (type 431). A high efficiency, lower capacity pad frequently used is a type 326 (8 lb/ft3, .006 dia wire). The more targets the mesh pad has the more efficient it is. Pads can be made thicker to get more efficiency without a decrease in capacity. I have attached several graphs depicting the effect of surface area and wire diameter. In glycol dehydration service we use 16 inch thick York-Reid Mesh pads. These are a coknit of 316SS and dacron. It works real well in this service. The coknit doesn’t work well in dirty oily services. This tends to plug the coknit and make it flood. Mesh pads are also referred to as wire mesh pads, demisters and mist elimination systems. It should be noted that demister is a registered trade mark of Otto York. Mesh pads are designed to work in a velocity range. Too low of a velocity allows the liquid particles to dodge the wires and escape. Too high of a velocity floods the mesh pad and allows the liquid particles to escape. The allowable velocity is calculated using a “K” value. The allowable K value at low pressure (100 psig) is between 0.15 to 0.35. The K value gets derated as the operating pressure increases, for every 100 psi over 100 psi subtract .01 from the allowable maximum K value. K max=0.35-0.01*(Operating Press-100)/100. If two liquid phases are present the lowest density liquid should be used when calculating the allowable velocity. Gas distribution is an important consideration in designing then mist elimination system. The gas inlet and outlet nozzles cannot be located close to the mesh pad, or localized high velocity will result. If you have no chose and have to locate the mesh pad close to the outlet nozzle, some of the maledistribution can be over come by installing a perforated plate above the mesh pad. The perforated plate must be above the mesh pad to avoid problems with maldistribution (it will take several hole diameters downstream of the hole to get uniform flow, while upstream is not effected by the local hole velocity, and the high velocities in the perforated holes will shatter droplets into smaller diameters – downstream of the mesh pad this doesn’t matter). The perforated plate causes pressure drop, which help maintain even flow across the mesh pad. Typical problems with mesh pads are that they fail as a result of not being carefully held into place. I prefer to have a support ring top and bottom for them and to wire them into place. With the support ring above the pad, it would be very difficult for the pad to get lose, even if the tie down wires were to get lose, or break.
Separation efficiencySeparation efficiency for mesh pads can be estimated by: Calculating the target efficiency, by using the attached graph. The x axis consists of K=D1²Vel(rL-rV )/18mDWire
Using the attached graph and the separation number, the target efficiency can be read. The target (impact) efficiency is then used with the following formulae to predict the overall mesh pad efficiency: Separation Efficiency=100-100/e 0.21STE Check out the separation efficiency graph and the impact of mesh pad thickness
VanesVanes are another method used to separate gas and liquid. While they can be oriented in either the horizontal or vertical (or even in a V-Bank, frequently used in horizontal vessels), the preferred orientation is vertical, with the gas flow horizontal. The vanes operate at higher velocities than the mesh pads. They also are generally used in combination with the mesh pads. The mesh pads precede the vanes and coalesce the small particles into large particles. The mesh pads are generally operated in a flooded condition, with the liquid exiting the mesh pads being removed by the downstream vane unit. The liquid impacts the vane and flows along it, until it hits a pocket. The liquid then flows down the pocket to a trough and is routed to a drain pipe and into the liquid portion of the separator. The drains are either routed to below the low level shut down point, or have pee traps. The drains must have some form of liquid seal, or entrainment laden gas will flow up the drain. I prefer to have both a pee trap and route the drain to below the low liquid level. Also note the pee traps should be liquid filled before the vessel is started up, if possible. The traps should eventually fill with liquid, but you are better off started with them full. The operating staff should be made aware that a liquid level must be maintained in the vessel for it to function correctly. The vanes operate with a pressure drop. The pressure drop for the vane is about 1 inch water column, the associated mesh pad will have a much higher pressure drop (3 to 10 inch water column). If the pressure drop is too high, liquid will be sucked up the drains and drawn off with the exiting gas. This will result in your separator failing. The vanes systems are good, but they are not idiot proof. Many of the problems with this type of separator can be in the drain system, so it is worthwhile to carefully review the drains and seals during the design phase. Also you should calculate the pressure drop that the liquid will incur in the drain system. If the pressure drop is too high, the liquid is again carried out with the exiting gas. Other problems are the drain lines are not well restrained, and are in the path of the incoming gas. This can result in setting up a harmonic vibration, which can eventually cause the drain system to fail. The mesh pad is generally 4 inches thick and a type 431. I have used a thicker pad for more difficult (small particles) separations. Vanes are sized based on a rV2. I use a rV2 of 32 (normal flow, 40 for design rate) for sizing Peerless P8X vanes.
Typical Vane Installation
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
body2 | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
footer | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||