Bed Seal System For Adsorber Vessel
Design Specifications For A Bed Seal System Of A Moisture and Carbon Dioxide Adsorber Vessel
SUBMITTED TO: [pic]
Air Products and Chemicals, Inc.
ADDRESS: 7201 Hamilton Boulevard
Allentown, PA 18195
CONTACT: Vincent D’Imperio, Jr.
PHONE: 610-481-6652
FAX: 610-481-2400
EMAIL: DIMPERVJ@
SUBMITTED BY: [pic]:
Amy Boyce
J.R. Daversa
Ben Gatto
Jake Mattern
Brian Smith
TEAM CONTACT: Brian Smith
ADDRESS: 120 Greary Hall
University Park, PA 16802
TELEPHONE: 814-862-2265
EMAIL: bps138@psu.edu
DATE: March 14, 2002
EXECUTIVE SUMMARY:
This report provides the details of the final design of the bed seal system and the layout for the working prototype. Calculations are included which back up the claims made, showing that the thermal expansion of the different materials should not go beyond the confines of the vessel, nor lead to significant gaps through which absorbent material can escape, and that the movement of the different components should not cause the bolts holding them together to shear. Slight alterations were made from the previous design, based on preliminary calculations, leading to this final design.
The governing equations upon which our test prototype is based are given, explaining how thermal cycling can be simulated mechanically using a motor attached to life-size prototype. The prototype will be tested as to whether it can endure a 20-year life with six thermal cycles per day.
Based upon the information contained in this report and the future testing plans, implementation of this system as a solution for the leakage problem incurred by Air Products is a feasible solution.
TABLE OF CONTENTS
Introduction……………………………………………………………………………………. 4
Objectives……………………………………………………………………………………… 6
Nomenclature…………………………………………………………………………………. 7
Design Calculations…………………………………………………………………………... 8
Conclusions………………………………………………………………………………….. 13
References……………………………………………………………………………………14
Appendicies
Appendix A………………………………………………………………………………15
Appendix B………………………………………………………………………………18
Appendix C………………………………………………………………………………20
Appendix D………………………………………………………………………………22
Appendix E………………………………………………………………………………24
INTRODUCTION:
Pure gases play a vital role in a great many production processes in commercial applications. For instance, in the petrochemical segment they serve to keep oxygen at bay, displace or dispense combustible liquids or gases, protect sensitive goods against oxidation, and purge pipelines. Gases and chemicals such as pure oxygen, nitrogen, and argon serve many uses and applications in industry (7).
Oxygen:
The discovery that oxygen promoted combustion has been the basis for most industrial uses of oxygen. In the glass industry, oxygen/fuel combustion is used to reduce particulate and NOx emissions in melting operations. For metal fabrication, oxygen is burned with acetylene, propane, and other gases used in welding and cutting torches to provide a secure joint, stable fixture and fastening. In the pulp and paper industry, oxygen in used for limekiln enrichment.
In addition to industrial applications, oxygen is also recognized for its respiratory aid attributes. Traditionally, used for life saving applications, oxygen has been supplied to hospitals and through bulk tank installations, or larger cylinders, where the patient is confined to the institution. More recent advances have been concerned just as much with the quality of life, rather than just the saving of life.
Such quality can be enhanced if the patients can be allowed to return to their own homes, and even a more normal way of life, rather than being dependent on oxygen only available through the hospital or fixed installations.
As pure oxygen contains 5 times the concentration of oxygen in air, it can be used very successfully to provide ample oxygen to living systems that may be short of oxygen under normal atmospheric conditions. Such systems include biological wastewater treatment plants, rivers, lakes and fish farms.
Nitrogen:
Gaseous nitrogen is used in the chemical and petroleum industries for storage tank blanketing and vessel inerting applications. It is also used extensively by the electronics and metals industries for its inert properties. Liquid nitrogen, produced by the cryogenic air separation process, finds wide use as a refrigerant in applications such as cryogenic grinding of plastics and food freezing.
Argon:
Argon is used primarily for its properties as an inert gas in applications such as arc welding, steel making, heat-treating, and electronics manufacturing.
Air Products:
Air Products and Chemicals, Inc. is a company that produces a wide array of gases and chemicals for commercial use. Oxygen, nitrogen, and argon are taken directly from the air and separated using advanced technology and materials.
A common way to acquire pure oxygen and other gases is to cryogenically distill them from air. Air is cooled to the point where it liquefies, and then it is separated through distillation into its major parts. Before the air can be distilled, the water and carbon dioxide present in it must be removed so that it does not freeze and contaminate the distillation equipment.
Air Products and Chemicals, Inc. uses Temperature Swing Adsorption (TSA) technology to remove the troublesome water and carbon dioxide. In this process, air is forced through a vertical vessel, where it passes through an adsorbent bed of granular material. The adsorbent granules are able to remove the unwanted water and carbon dioxide from the air, so that the air is able to leave the vessel without taking its fouling components with it. The problem with using the adsorbent granular material is that it becomes saturated with water and carbon dioxide and is no longer able to remover after a few hours. To undo this saturation of the granules, a clean dry regeneration gas, such as nitrogen, is forced through the adsorbent bed in the direction opposite to that which the air is passed. While the regeneration gas is passing through the bed, it is heated, which in turn heats the granules to more effectively remove the water and carbon dioxide from them. After the undesirable agents are adequately removed from the adsorbent granules, the gas temperature cools down, cooling the bed to its appropriate operating temperature. At this point, the operation begins again.
The granules themselves, about 2mm in diameter, rest in the vertical pressure vessel on top a flat support screen. This screen supports the weight of the adsorbent material as well as the forces due to the pressure differential as gases pass through it. The screen must be able to sustain the massive temperature swing associated with the cryogenic distillation and the regenerative process. This temperature swing produces thermal expansion between the vessel wall and the screen.
Problem:
Thermal expansion creates a way for the adsorbent material to move about and travel into places in the assembly where it should not be. As seen in the attached photographs, the current setup of the screen region consists of a grating surrounded by a packing material. The screen is then bolted over this grating. The packing material seals the screen against the vessel wall preventing the adsorbent granules from entering into the chamber on the other side of the grating.
Currently, brass wool is used as the packing material. The adsorbent particles are able to travel through the brass wool and leak past the support screen. Thermal expansion moves the screen in such a way that the particles are able to move past the screen more easily. Another factor involved in passing the adsorbent granules past the screen is the pressure difference on opposite sides of the screen while air is forced through it.
This failure of the packing material to retard the progress of the granules past the screen is a consistent and long-term problem present in the process. Adsorbent leakage is disruptive to the process, difficult to clean, and a source of process down time.
The problem of leaking adsorbents is a consistent problem that has been attempted to correct in the past with little or no satisfaction. Other research has been done on this problem by teams from other schools. According to the sponsor, while these teams have had some interesting ideas, there is no adequate solution yet available to merit the investment in modifying the current equipment.
OBJECTIVES:
1. Determine the feasibility of the seal system design by calculating:
• The thermal expansion of the different pieces of the system in relation to one another, keeping in mind the varying materials used
• The shear stress on the bolts holding the seal system together
2. Specify the design of the prototype that will:
• Simulate the thermal cycling to which the system will be subjected
• Use a motor to approximate the physical motions that the system components would undergo
• Calculate the motor size and speed necessary to produce similar conditions that would test for a 20-year, six cycle per day life
NOMENCLATURE:
α Coefficient of linear thermal expansion ºC (ºF)
c Damping coefficient N•s/m (lb•s/ft)
δ Deflection mm (in)
E Modulus of elasticity Pa (psi)
Fo Excitation amplitude N (lb)
I Moment of inertia m4 (in4)
k Spring coefficient N/m (lb/ft)
N Factor of safety
NAct Actual number of cycles needed cycles
NFail Number of cycles till failure cycles
σ Thermal stress Pa (psi)
σTot Total stress Pa (psi)
P Load N (lb.)
Se Endurance limit Pa (psi)
Sut Ultimate tensile strength Pa (psi)
Sys Yield strength Pa (psi)
τ Shear Pa (psi)
ω Excitation frequency rad/s
[pic] Bolt Force lbs
[pic] Screen Plate Force lbs
[pic] Bolt Torque lb-ft
[pic] Summation of Moments about pt 2 lb-ft
[pic] Max Deflection in y-direction in
[pic] Second moment of area about y-axis in4
DESIGN CALCULATIONS:
Bending Calculations
The group’s assumption as to how the beads are leaking through the current assembly was that they were passing underneath the hold down ring. The beads passed through gaps created by the ring’s bending/buckling between the bolts that fasten it down. The bending/buckling was caused by a bending moment created by torquing the bolts down onto the ring, which is supported by an uneven surface as seen in drawing 3 and 4 in Appendix A, and the free body diagram seen in drawing 2 Appendix E. The prove that this was the weak link in the assembly and the reason why it was leaking, the group would have to create a design that would eliminate or drastically reduce the bending/ buckling created in the ring by the grate and torquing down the bolts. Therefore, before further action could be taken the group must first verify this assumption. To determine if there was significant bending in the ring a preliminary “hand” calculations were used. After proving a significant bending in the ring, a computer simulation and analysis of the system were then done to verify the “hand” calculation and to gain more exact results. This computer simulation was done in Pro Mechanica using Finite Element Analysis (FEA), Drawing 3 in Appendix E.
The “hand” calculation was done using a bolt torque of 115 ft lb, and a bolt diameter of 0.75 in. The distance between bolts is 12.2 in as seen in Drawing 2 in the Appendix. This figure was obtained using the pressure vessel dimensions provided by the sponsor. The distance between bolts is 18(, which on a 7 ft pressure vessel is 12.2 in. On a larger pressure vessel the distance between bolts would obviously be larger, and the larger distance between bolts would result in a greater deflection in the area of the ring between the bolts. The result of the calculation for the 7 ft vessel was that the deflection in the middle of the ring was 0.0212 in. The mean diameter of the adsorption beads is only 2mm (0.0787 in), which indicates that some of the beads will be smaller, or may be broken into smaller pieces that will fit through the gap created between the bed and the ring. Therefore, the deflection in the ring was significant and was further analyzed using FEA as seen in Drawing 3 in the Appendix E. The tabulation of the “hand” calculations can be seen in the Appendix B.
Thermal Expansion
From the given graph (Appendix), the temperature change for the shell is 60 degrees Celsius. From the same graph, the temperature change for the screen plate is 150 degrees Celsius. This is for maximum temperature change.
From the proposed prototype, a motor would be placed at one specific point, and used to simulate the linear thermal expansion. In order to determine what deflection the motor must simulate, the linear deflection of the system due to thermal expansion must be calculated. This can be calculated using the following equation:
[pic]=[pic] (1)
The deflection of the pressure vessel’s shell due to thermal expansion was figured out using equation 1. The range of given values for the thickness of the shell was from ½ in to 1 in. The maximum thickness for the shell was used to obtain the maximum deflection of the shell. The thickness of 1 in was converted to the metric measure of 25.4mm for use in the calculation. Using the linear coefficient of thermal expansion for carbon steel (1) and equation 1, which can be found in the Appendix, the deflection of the shell due to thermal expansion equals:
(shell=0.01646mm = 6.48e-4 in
For the proposed redesign of the screen plate, the diameter is 84 in. To determine the best conditions, two calculations were made for the screen plate. One calculation was made assuming the screen plate was a solid disc. This would result in the maximum possible deflection by using the entire diameter for point linear thermal elongation. The diameter of 84 in was converted to the metric measure of 2133.6mm for the calculation. Using the linear coefficient of thermal expansion for carbon steel and equation 1, which can be found in the Appendix, the deflection of the screen plate as a disc due to thermal expansion equals:
(screen plate (disc) = 3.456mm = 0.131 in
The second calculation was done assuming the screen plate as a ring. This was a more realistic calculation. It was assumed that the grating thermal expansion would be negligible compared to the expansion of the banding and ring on the outside of the screen plate. The length used was 4 in, which was then converted to the metric measure of 101.6mm for the calculation. Using the linear coefficient of thermal expansion for carbon steel and equation 1, which can be found in the Appendix, the deflection of the screen plate as a ring due to thermal expansion equals:
(screen plate (ring) = 0.1646mm = 0.00648 in
To determine the thermal expansion of the gasket, the change in temperature of the gasket was assumed to equal the change in temperature of the screen plate. The given data for the Teflon® used can be seen in the Appendix. Since the gasket is a ring, the difference between the inner and the outer radius was determined to be 4 in which was then converted to the metric measure of 101.6mm. Using the linear coefficient of thermal expansion for Teflon® (4) and equation 1, which can be found in Appendix, the deflection of the gasket due to thermal expansion equals:
(gasket=0.762mm = 0.03 in
Stress Analysis
The downward force that the bolts placed on the gasket was 8000lb. This equates to a pressure of 175psi on the gasket. It was known that the minimum tensile strength for Teflon® is 4000psi (4). To confirm that the maximum stress due to temperature and torque does not exceed the minimum tensile strength using equation 2:
Eq. 2 (=E((T (3)
Using equation 2, the coefficient of linear thermal expansion and Modulus of elasticity for Teflon® (4), the stress caused by the change in the temperature of the gasket was determined to be:
(=870psi
The total stress placed on the gasket was the sum of the two determined stresses from the bolts and the thermal stresses. This gives a total stress of:
(Tot=1045psi
This total stress is well below the minimum tensile strength of 4000psi.
Cyclic Fatigue
One of the desired outcomes was for the new seal system to endure for a certain number of cycles. The new seal system needs to be able to endure 6 cycles a day for 20 years. The total number of cycles desired is:
NAct. =4.368*104 cycles
The theoretical number of cycles before failure occurs due to fatigue is determined using equations 3, 4, and 5 found in the Appendix. The total number of cycles before failure occurs on the gasket is:
NFail=2.052*109 cycles
Since the desired number of cycles is less than the number of cycles till failure, this gasket will not fail during to course of operation of a 20-year period of time.
Bolt Shear
From the calculations of the thermal expansions, the total displacement of the hold down ring relative to the pressure vessel wall and flange were determined. This total relative displacement, (T, is 3.44 mm for the disk approximation and 0.1481 mm for the ring approximation of the assembly. From these displacements, calculations were performed to determine if the bolts would shear under this displacement [5]. First, the material spring coefficient, K, for a bolt in shear was found. The spring coefficient was determined to be 45e6 lb/in or 7.88e6 kg/mm. From this value and the total relative thermal displacement, the force on all the bolts, P, was found to be 13.55e6 N and .583e6 N, for the disk and ring approximations respectively. Assuming that 20 bolts are used to secure the hold down ring, the force per bolt can be found simply dividing by 20. The force per bolt, FB, was found to be 667.7 kN and 29.2 kN, for the disk and ring approximations respectively.
To determine the shear force on each bolt the equation below is used.
[pic] (5)
The cross-sectional area of the bolt, AB, can be found using the equation, [pic]. For bolts that are 0.75 inches in diameter, the cross-sectional area is 2.85e-4 m2. From this, the shear force in the bolts, (, was found to be 2.378 GPa and 0.102 GPa, for the disk and ring approximations respectively. Using an approximate yield strength for the bolts of 50 ksi, the lowest yield strength for carbon steel, the factor of safety was able to be determined with the following equation.
[pic] (5)
The factor of safety for a solid disk, a clear overestimate of the worst-case scenario, was found to be 0.145. For the approximation of ring, which is the more accurate of the two approaches, the factor of safety was found to be 3.367.
This is a sufficient factor of safety so that the bolts in the actual vessel would not shear. Even though these are just approximations, the physical redesigned grate would be most closely approximated as a ring as opposed to a solid disk. Also, bolts of the exact same dimensions are being used in the design currently and there is no problem with shearing the bolts, as has been investigated.
Prototype Design:
A mechanical approximation will be used to simulate the thermal expansion that the seal system experiences. In the mechanical simulation, the thermal expansion will be simplified to the grate moving with respect to the wall flange and the hold down ring. A motor will be used to drive a crank rocker mechanism, which in turn moves the grate. The movement of the grate will be a total distance of 0.25 inches. The prototype will be a linear section of the total vessel. The vital parts of the prototype will be the same scale as those of the real system. Due to the large diameter of the vessel, the curvature of the fourteen inch section will be neglected.
The spring constant of the wall of the vessel can be approximated by,
[pic]. (6)
where keq is the equivalent spring constant of a beam, E is modulus of elasticity of the material, I is the moment of inertia, and l is the length of the beam.
The motion of the motor will create a movement in the vessel similar to the changes created by thermal cycling. The equations governing the motion of the motor on the system are that of harmonic forcing on a damped system,
[pic]. (6)
Where meq is the equivalent mass of the motor-vessel system, which can be determined by weighing the system; ceq is the damping constant of the system, which can be determined experimentally; keq is the previously mentioned spring constant; and Focosωt is the forcing function of the motor, where ω is found from the speed of the motor.
The resulting amplitude of movement of the wall of the vessel can be found from,
[pic]. (6)
CONCLUSIONS:
Calculations were done to examine the possibility that the hold down ring was bending upward due to torque as it was bolted to the rest of the bed seal assembly. This was found to be the case, and a majority of the redesign was done to correct this possibility. Aspects of the hold down ring, grating, and packing material were eliminated or redesigned. A cross sectional view of the entire redesigned and dimensioned assembly can be seen in Drawing 1 in the Appendix.
The hold down ring was thickened to 1 in. to decrease its susceptibility to bending. The grate was banded to retard beads from passing through the gaps in the grate itself. Then a 4 in. solid ring was placed around the outside to support the grate on the flange of the vessel. This ring will provide an even surface for the hold down ring to be bolted down onto, minimizing the deflection between bolts. The ring will have holes drilled in it for the bolts. The holes will be oversized to allow for motion during thermal expansions. This way the bolts will not be constricted, reducing the shear force on them. Teflon will be used as gaskets between the metal on metal interfaces. This design will securely seal the bed assembly such that no beads will be able to pass through. The low coefficient of friction of Teflon will also reduce the wear of the gasket during thermal expansion.
A design for a prototype was developed. The prototype will simulate the thermal expansions using a mechanical system. All pieces of the assembly will be replicated as close as possible to the actual pieces to more closely simulate the actual system.
REFERENCES
(1) Mischke and Shigley. Mechanical Engineering Design 5th Ed. McGraw-Hill Inc., NY, 1989
(2) Hibbeler. Mechanics of Materials 3rd Ed. Prentice Hall, Inc., Upper Saddle River, NJ, 1994
(3) Queeney and Conway Jr., Mechanical Response of Engineering Materials. Penn State University, State College, PA, 1997
(4) “Properties of Gore Wire Insulation Materials” Gore Electronics, (Viewed 2/02).
(5) Norton, Robert L. Machine Design: An Integrated Approach 2nd Ed. Prentice-Hall, NJ, 1998.
(6) Rao, Singiresu S., Mechanical Vibrations 3rd Ed. Addison-Wesley Publishing Company. Reading, Massachusetts, 1995.
(7) Air Products. (Viewed 3/12).
APPENDIX A
Pictures of existing support and seal system
|[pic] | Vessel Wall |
| | |
| | |
| |Wire Mesh |
| | |
| | |
| |Hold Down Ring |
| |
|View of Completed Bed Support Assembly from the Top Side |
|Figure 1 |
|[pic] | Vessel Wall |
| | |
| | |
| |Hold Down Bolt |
| | |
| | |
| |Brass Wool |
| | |
| | |
| | |
| | |
| |Hold-Down Ring |
| | |
| | |
| | |
| | |
| |Wire Mesh |
| |
|Closeup View of Completed Bed Support Assembly from the Top Side |
|Figure 2 |
| Vessel Wall Hold Down Bolts |
|[pic] | |
| |Brass Wool |
| | |
| |Support Ledge |
| | |
| | |
| | |
| | |
| | |
| |Support Grating |
| | |
| | |
| | |
| |Support Beam |
| |(Note: Running in wrong direction…should be |
| |rotated 90°) |
| |
|View of Bed Support Assembly from Top Side (Assembly Stage) |
|Figure 3 |
|[pic] | Vessel Wall |
| | |
| |Support Ledge |
| | |
| | |
| |Brass Wool |
| | |
| |Hold Down Bolts |
| | |
| | |
| | |
| | |
| | |
| | |
| |Support Grating |
| |
|Support Beam (see Note Fig.) |
| |
|Closeup of Brass Wool Packing |
|Figure 4 |
APPENDIX B
Bending Calculations
Bending Calculations
Bolt Torque
(Specified by ASME standards) = 115 ft lb
Diameter of Bolts = [pic]ft
[pic] (5)
[pic]
[pic]
*Deflection will be greatest in the middle of the ring (i.e. x = 6.1 in)
[pic]
[pic] (5)
[pic] (5)
Using Hold Down Ring Cross Section:
[pic] [pic]
(5)
[pic]
APPENDIX C
Thermal Expansion and Cyclic Fatigue
Coefficients of Linear Thermal Expansion
αcarbon steel = 10.8*10-6/(C
αTeflon = 10*10-5 /(C
Modulus of Elasticity
ETeflon = 0.58*105psi
Thermal Expansion
[pic]=[pic]
δshell =10.8*10-6 /(C * 60(C *25.4mm
δshell = 0.01646mm = 6.48*10-4in
δscreen plate (disc) = 10.8x10-6 /(C * 150(C * 2133.6mm
δscreen plate (disc) = 3.456mm = 0.131in
δscreen plate (ring) = 10.8*10-6 /(C * 150(C * 101.6mm
δscreen plate (ring) = 0.1646mm = 0.00648in
δgasket = 10*10-5 /(C * 150(C * 50.8mm
δgasket = 0.762mm = 0.03in
Stress Analysis
σ = EαΔT
σ = 0.58*105psi * 10*10-5/(C * 150(C
σ = 870psi
σTot = σ + 175psi
σTot = 1045psi
Cyclic Fatigue (3)
Sut = 4000psi
Se = 0.5Sut
a = (0.9Sut)2/Se
a = 6480psi
b = -1/3 log (-0.9Sut/Se)
b = -0.0851
NFail = (σ/a)^1/b
NFai l= 2.052*109 cycles
APPENDIX D
Bolt Shear Calculations
Bolt Shear:
Total Thermal Displacements:
(T = (screenplate-(shell
Disk: (T = 3.456 - 0.01646 = 3.44mm
Ring: (T = 0.1646 - 0.01646 = 0.1481mm
Expansion on either side of the bolt will be half the total:
Disk: (T = 1.72mm
Ring: (T = 0.07405mm
Load and Shear on the bolt:
k = EA/l = 30e6psi * 1.5in * .75in / .75in = 45e6 lb/in = 7.88e6 kg/mm
P=k*(T
Disk: P=7.88e6 kg/mm*1.72mm = 13.55e6 N
Ring: P=7.88e6 kg/mm*0.07405mm = 0.583e6 N
Per Bolt:
FB = P/20
Disk: FB = 13.55e6 N/20 = 667.7kN
Ring: FB = 0.583e6 N/20 = 29.2kN
Area:
AB = (r2= (*(0.01905/2)2 = 2.85e-4 m2
( = FB/AB
Disk: ( = 667.7kN/2.85e-4 m2 = 2.378 GPa
Ring: ( = 29.2kN/2.85e-4m2 = 0.102 GPa
Factor of Safety
N = SYS/(
SYS = 50,000psi = 0.3447 GPa
Disk: N = 0.3447GPa/2.378GPa = 0.145
Ring: N = 0.3447GPa/0.102GPa = 3.367
APPENDIX E
Hold Down Drawings and Diagrams
................
................
In order to avoid copyright disputes, this page is only a partial summary.
To fulfill the demand for quickly locating and searching documents.
It is intelligent file search solution for home and business.