Blast Design and Analysis

Blast Design and Analysis

National Business Park- Building 300 300 Sentinel Drive Annapolis Junction, Junction, MD, 20701 20701

Rebecca Dick Structural Option Dr. Memari Spring 2012 Senior Thesis April 4, 2012

Spring Thesis

| Structural Option Rebecca Dick |

Acknowledgments Acknowledgme nts A very special thank you Dr. John O. Hallquist, President, Livermore Software Technology Corporation (LSTC), and Marsha J. Victory, President, FEA Information Inc, for providing me access to the LS-DYNA finite element software for free for my senior thesis and to Gunther Blankenhorn and Todd P Slavik for the invaluable technical support and troubleshooting troubleshooting they provided. Additional thanks to Michael Baker, Jr., Inc. for the use of National Business Park- Building 300, especially Mr. Erik Spicker. Special Thanks to:

My Family: Thank you to my dad for always answering my questions for the last five years. Thank you to my mom for listening to me complain all the time and putting up with my breakdowns. breakdowns . Thank you to my brother for giving me academic competition. Penn State Faculty: Dr. Memari Dr. Hanagan Dr. Parfitt Ryan Solnosky My Friends: Brian Rose Dave Tran Mike Kostick Ryan Blatz Liz Kimble Rob Livorio RJ Fazio The past five years have been a blast and I couldn’t have made it through without without you guys. You’ve kept

me as close to sane as any AE can be through all the crazy projects, late night study sessions, and postexam celebrations. celebrations. I will always love love you guys!

Advisor: Dr. Memari

Page | 2

Spring Thesis


Acknowledgments Acknowledgme nts A very special thank you Dr. John O. Hallquist, President, Livermore Software Technology Corporation (LSTC), and Marsha J. Victory, President, FEA Information Inc, for providing me access to the LS-DYNA finite element software for free for my senior thesis and to Gunther Blankenhorn and Todd P Slavik for the invaluable technical support and troubleshooting troubleshooting they provided. Additional thanks to Michael Baker, Jr., Inc. for the use of National Business Park- Building 300, especially Mr. Erik Spicker. Special Thanks to:

My Family: Thank you to my dad for always answering my questions for the last five years. Thank you to my mom for listening to me complain all the time and putting up with my breakdowns. breakdowns . Thank you to my brother for giving me academic competition. Penn State Faculty: Dr. Memari Dr. Hanagan Dr. Parfitt Ryan Solnosky My Friends: Brian Rose Dave Tran Mike Kostick Ryan Blatz Liz Kimble Rob Livorio RJ Fazio The past five years have been a blast and I couldn’t have made it through without without you guys. You’ve kept

me as close to sane as any AE can be through all the crazy projects, late night study sessions, and postexam celebrations. celebrations. I will always love love you guys!

Advisor: Dr. Memari

Page | 2

Spring Thesis

Advisor: Dr. Memari


Page | 3

Spring Thesis

Rebecca Dick | Structural Option

Table of Contents Acknowledgments ......................................................................................................................................... 2 Special Thanks to: ................................................................................................................................. 2 Executive Summary ....................................................................................................................................... 6 National Business Park- Building 300 ............................................................................................................ 8 Structural Background .................................................................................................................................. 9 Foundations .............................................................................................................................................. 9 Floors....................................................................................................................................................... 12 Load Path ............................................................................................................................................ 12 Columns .................................................................................................................................................. 13 Lateral System ......................................................................................................................................... 14 Load Path ............................................................................................................................................ 15 Impact of Lateral System on Calculations ........................................................................................... 16

Design Codes ............................................................................................................................................... 17 Material Properties ..................................................................................................................................... 18 Original Design ........................................................................................................................................ 18 Reinforcement: ................................................................................................................................... 18 Structural Steel: .................................................................................................................................. 18 Metal Deck: ......................................................................................................................................... 18 Concrete: ............................................................................................................................................. 18 Blast Redesign ......................................................................................................................................... 19 Reinforcement: ................................................................................................................................... 19 Structural Steel: .................................................................................................................................. 19 Metal Deck: ......................................................................................................................................... 19 Concrete: ............................................................................................................................................. 19 Gravity Loads............................................................................................................................................... 20 Dead Load ............................................................................................................................................... 20 Live Load ................................................................................................................................................. 20 Snow Load ............................................................................................................................................... 20 Structural Proposal ..................................................................................................................................... 21 Blast Design ................................................................................................................................................. 22 Procedure .................................................................................................................................................... 23 Advisor: Dr. Memari

Page | 4

Spring Thesis


Blast Load Determination ........................................................................................................................... 24 Column Design ............................................................................................................................................ 28 Beam Design ............................................................................................................................................... 30 Girder Design .............................................................................................................................................. 32 Slab Design .................................................................................................................................................. 33 Moment Connection Design ....................................................................................................................... 36 Final Design ................................................................................................................................................. 39 LS-DYNA Modeling ...................................................................................................................................... 40 Comparison of Results ................................................................................................................................ 44 Additional Comments and Conclusion ........................................................................................................ 48 Lateral System Implications .................................................................................................................... 48 Effects on Foundations ........................................................................................................................... 48 MAE Requirements ..................................................................................................................................... 49 Breadth Topic I: Site Redesign .................................................................................................................... 50 Breadth Topic II: Façade Redesign and Heat Transfer ................................................................................ 54 Façade Design ......................................................................................................................................... 55 Heat Transfer .......................................................................................................................................... 57 Cost Estimates ............................................................................................................................................. 61 Conclusion ................................................................................................................................................... 63

Advisor: Dr. Memari

Page | 5

Spring Thesis


Executive Summary National Business Park- Building 300 is a seven story office building located in Annapolis Junction, Maryland. It was designed in 2007 and construction was completed in 2009. The structure of NBP-300 is composed of a composite steel system and utilizes four eccentrically braced frames for the lateral system located at the core of the building. National Business Park- Building 300 was not designed to resist blast, and therefore, were an attack made on the building, heavy structural damage would likely result. For this reason, it will be assumed that NBP-300 will be redesigned as a high risk building for terroristic threats, and as such would be required to meet certain criteria for blast loading. Designing for interior blast has many challenges, as well as many solutions. There is no clear “cookie-cutter” method to designing for blast, which makes it one of the most intriguing and interesting new topics of study. Blast design isn’t new per say, but in the realm of structural design, it has only be in consideration for the last 50 years or so. Although progressive collapse is a major issue when designing for blast, the scope of this redesign will be limited to strength factors only due to time restraints, and redundancy of design will be ignored. A specific situation has been created for the design. It can be assumed that the greatest threat in terms of explosives will be a briefcase sizes device detonated in the interior of the lobby on the second floor. The UFC -340-02 documents outlining blast requirements were used to identify threat levels, design site security, and find blast loads. The blast load from this situation was calculated and a typical bay located at the lobby location was designed to withstand the blast loading found. Additionally, the façade of NBP-300 was redesigned to allow venting of the interior during the blast. The large percentage of glass currently on the façade could potentially be harmful to occupants when the façade fails. LS-DYNA Blast Modeling software was used to analyze a critical portion of the original and redesigned structure. A comparative study was completed where the assumed explosive for this situation was detonated in the original LS-DYNA model and the redesigned LS-DYNA model. The intent was to minimize structural damage as much as economically possible through this redesign. By comparing the original structure to the redesign it became apparent that the redesign was highly conservative, but overall, withstood the blast load and incurred very little damage. In addition to the blast design and analysis, a site redesign was completed. Also, a façade breadth was completed, including a façade redesign with anchored “blowout panel” and a heat transfer study of the new façade. MAE coursework was incorporated into several aspects of the redesign for NBP-300, including building modeling techniques (AE 597A), a moment connection design (AE534), and a façade design (AE542).

Advisor: Dr. Memari

Page | 6

Spring Thesis


Going into this thesis design, the following goals were set forth and achieved: 

Design a typical beam, girder, column, moment connection, and floor system



Use a finite element analysis software (LS-DYNA Blast) to model the effects of blast in the original structure and redesigned structure, and compare the results



Show that designs using hand calculations are overly conservative



Redesign the site of NBP-300 to mitigate large exterior blasts



Redesign the façade of NBP-300 to allow venting of the interior during an interior explosion



Calculate heat transfer through the new façade and determine if the new design is acceptable

Advisor: Dr. Memari

Page | 7

Spring Thesis


National Business Park- Building 300 National Business Park- Building 300 is an office building located in an industrial park in Annapolis Junction, Maryland. Owned by COPT, the seven story office building, which has an additional mechanical penthouse on the roof, was designed in 2007 and construction was completed in 2009. As part of an industrial park, the architecture was intended to complement the existing offices in the surrounding area. Shown below in Figures 1 and 2 is the building footprint of National Business ParkBuilding 300 and a satellite view of the building, respectively. NBP-300 is a 212,019 gross square foot, 116.92’ tall, composite steel framed building that utilizes braced frames for its lateral system. Foundations consist of strip and spread footings, but since the first story is only a partial floor plan, and is sub-grade at a few locations, footings are located below both the first and second stories at their respective locations. The façade system is composed of a glass and precast concrete curtain wall that is tied into the structural system at each floor level.

Figure 1: NBP-300 Footprint on Site

Figure 2: Satellite View of NBP-300 provided by

Courtesy of Baker and Associates

©Google Maps, 2011

Advisor: Dr. Memari

Page | 8

Spring Thesis


Structural Background Foundations Since the first floor is only a partial floor and the grading on the site is such that the second floor is at ground level at the main entrance, foundations occur below both the first and second floors at the respective areas where no preceding floor exists below. Figure 3, on the following page, outlines the area of the partial first floor. Figure 4, also on the following page, provides more detail by showing a section cut in the North-South direction, which shows the partial first floor and the slab on grade and foundations on the second floor. Additionally, two side of the first floor will be supported by basement walls. These walls are reinforced with #7 bars at 10” on center, each face, vertically, and #5 bars at 12” on center, each face, horizontally. These walls will see lateral pressure from soil, which was noted to be 60 psf per foot of depth in the structural notes. The foundations of NBP-300 consist of strip and spread footings with a 5” slab -on-grade on the partial first floor and on the second floor where it does not sit over the first floor, and were designed to meet the suggestions set forth in the geotechnical report prepared by Hillis-Carnes Engineering Associates, Inc. 6x6 W2.0xW2.0 Welded Wire Fabric was used in all slabs-on-grade. A bearing pressure of 3.5 ksi was required for all foundations. 4” of compacted backfill meeting AASHTO 57 course aggregate was also a requirement below the slabs-on-grade. The strip and spread footings are located around the perimeter of the partial first floor and at the perimeter locations on the second floor that contain the slab on grade. Additionally, interior columns are supported by spread footi ngs. These range in size from 8’ square to 19’ square, depending on the location and the load. The depths of the spread foundations also vary between 24” to 46”. A 5’-0”x5’-0” footing at piers is typical, reinforced with #5@12” on center, each way. Additionally, the mechanical room on the first floor is sunken 3’ below the standard floor level which

required the foundations to be stepped down at the transition locations.

Advisor: Dr. Memari

Page | 9

Spring Thesis


Figure 3: Outline of Partial First Floor Plan on Foundations, Courtesy of Baker and Associates

Figure 4: Building section showing partial first floor and slab on grade at second floor, Courtesy of Baker and Associates

Advisor: Dr. Memari

Page | 10

Spring Thesis

Advisor: Dr. Memari


Page | 11

Spring Thesis


Floors NBP-300 was constructed using a composite steel system. For the floors, this entailed the use of cambered composite beams and girders. A 3 ½” lightweight concrete topping (117 pcf) on 3”x20 gage metal floor deck, bonding type, was used on each elevated floor for the composite slab, reinforced with 6x6 W2.0xW2.0 welded wire mesh. Additional reinforcement was required over filler beams and girders along column lines. This additional reinforcement consists of #3 bars at 18” on center, 8’-0” long. Composite action is obtained through the use of ¾” diameter, 5 ½” long shear studs spaced equally along beams. See Figure 5, below, for a typical section cut through the composite floor system. Floor beams vary in size and vary per floor, but are cambered 0”, ¾”, or 1”, depending on the location. The exterior girders also vary in size, but are not cambered. Beams typically span 35’ at 10’ on-center, while girders typically span 30” at 35’ on-center.

Load Path Gravity loads on each floor are transferred from the composite slab to the beams. The load is

then transferred to the girders, which then transfer the load to the columns. The gravity loads terminate at the foundations at the second level or at the partial first level.

Figure 5: Typical Composite Floor System, Courtesy of Baker and Associates

Advisor: Dr. Memari

Page | 12

Spring Thesis


Columns Wide flange columns are used throughout NBP-300. Although the weight of the columns differs throughout the building and across floor levels, most wide flanges are W14 and all conform to ASTM A992 Grade 50 steel requirements. Splices occur at levels 4 and 6 for most columns. This information is presented in a column schedule, of which a partial view can be seen in Figure 6, below. At the penthouse level, square HSS posts are used around the perimeter of the enclosure to support the penthouse cladding. Base plates are used to connect the steel columns to the foundations or shear walls located on the first floor. The base plates vary in size, but conform to ASTM A-36 type steel.

Figure 6: Partial Column Schedule, Courtesy of Baker and Associates

Advisor: Dr. Memari

Page | 13

Spring Thesis


Lateral System The main lateral system used in NBP-300 consists of four different eccentrically braced frames. These occur at column lines “D”, “E”, “4” , and “11”, each spanning about one full bay (See Figure 9 on the following page). This provides two lateral bracing frames in each direction. The frame at column line “D” begins on the first floor, w hile the other three frames begin on the second floor, supported by concrete shear walls on the first level. Two wind bracing frames are shown below, in Figures 7 and 8, to show the difference in height. The shear walls at the base of the three shorter frames are 1’-2” thick and are reinforced with #7 bars at 10” on center, each way on each face of the wall.

HSS tubes are used for the bracing, and vary in size from frame to frame, but conform to ASTM A-500 Grade C steel and have a yield strength of 46 ksi. All connections in the braced frames are slip-critical. Shear studs at wind bracing beams occur at 12” on center, and are ¾” in diameter and 5 ½” in length. At

the base plates, an 8x4x3/4 angle at 1’-6” in length is used on each side of the gusset plate to resist 100% of the design shear. Additionally, 7/8” diameter by 8” long shear studs are used at the shear wall connections for the three elevated braced frames.

Figures 7 and 8: Wind Braced Frames, Courtesy of Baker and Associates

Advisor: Dr. Memari

Page | 14

Spring Thesis


Figure 9: Location of Braced Frames at Core of building, Courtesy of Baker and Associates

Load Path As wind lateral forces are applied to NBP-300, they are transferred from the façade to the

composite floor system through bearing connections. From there, the load is transferred from the floor system to the four braced frames, which travel through the height of the building to foundations and shear walls on the first and second floor. Where braced frames connect to shear walls, the load is then resisted by the shear wall and ultimately transferred to the foundations below. Ground motion due to seismic loads is resisted by the foundations, first floor shear walls, and braced frames that run the height of the building. As each floor is seismically loaded, the force is transferred to the braced frames, which transfers the load to the foundations and shear walls at the base of the building.

Advisor: Dr. Memari

Page | 15

Spring Thesis


Impact of Lateral System on Calculations The lateral system used in the design of NBP-300 consists of four eccentrically braced frames. In

ASCE 7-05, this system provides a Response Modification Coefficient of 7, but since the structural documents provided for NBP-300 specifically state that seismic resistance was not a controlling factor in design (see Figure 10 below), an R value of 3 was used instead, to allow comparison between the calculations presented in this document and the design criteria of NBP300. This assumption effected the seismic calculations, namely the value of the seismic response coefficient. By using R=3, a higher base shear was calculated, making this a conservative assumption for design. In regards to detailing, this system may be more complicated and expensive. Eccentrically braced frames are used more in regions with high seismic activity, but by using an R=3, the stability will be affected during design and will have to be compensated through other detailing methods.

Figure 10: Response Modification Coefficient Assumption, Courtesy of Baker and Associates

Advisor: Dr. Memari

Page | 16

Spring Thesis


Design Codes National Business Park- Building 300 was designed in 2007 and constructed in 2009. At this point in time, the following codes were used to complete the design of the project:        

International Building Code (IBC) 2003 (with local amendments for Anne Arundel County, Maryland) The Life Safety Code, 2006 (NFPA 101) ASCE 7-02 AISC, 13th Edition ACI 318-02 AISI Specification for the Design of Light Gage Cold-Formed Structural Steel Members and the Steel Deck Institute’s Design Requirements Specifications for Masonry Structures, ACI 530.1/ASCE 6/TMS 602-92 Structural Welding Code, 2006, D1.1

In the redesign, the codes listed below were used to complete the analysis of National Business ParkBuilding 300.       

International Building Code (IBC) 2009 ASCE 7-05 AISC, 13th Edition ACI 318-05 Vulcraft 2008 Decking Manual UFC-340-02 AISC Seismic Design Manual

Advisor: Dr. Memari

Page | 17

Spring Thesis


Material Properties Original Design Reinforcement:  Reinforcing Bars

ASTM A615, Grade 60



Welded Wire Fabric (WWF) Reinforcing Bar Mats

ASTM A-185 ASTM A184



Lap Splices

ACI 318



Structural Steel:  Grade 50:

ASTM A992, Grade 50, F y= 50 ksi



HSS Pipes:

ASTM A500, Grade C, F y = 46ksi



Steel Tubes:

ASTM A500, Grade B, F y = 46ksi



All other steel:



Bolts:

ASTM A 36, F y = 36 ksi ASTM A325, with threads included in shear planes, ¾” or 7/8” diameter



Shear Studs:

¾” diameter x 5” long, uniformly spaced at 24”

maximum Metal Deck: 

Floors:

3”x20 gage, bonding type



Roof:

3”x22 gage

Concrete:

Minimum Concrete Compressive Strengths (f' c) Member

28 Day Strength (psi)

Elevated Slabs

3500

Slab-on-Grade

3500

Walls, Piers, and Grade Beams

4000

Interior Concrete Topping Concrete Exposed to Freezing

3500 4500

Grout

3000

All Other Concrete

3000

Table 1: Minimum Concrete Compressive Strengths

Advisor: Dr. Memari

Page | 18

Spring Thesis


Blast Redesign The following material properties were used in the redesign:

Reinforcement:  Reinforcing Bars

ASTM A615, Grade 60



Welded Wire Fabric (WWF)

ASTM A-185



Lap Splices

ACI 318

Structural Steel:  Grade 50: 

All other steel:



Bolts:

ASTM A992, Grade 50, F y= 50 ksi ASTM A572, Grade 50, F y = 50 ksi ASTM A325, with threads included in shear planes, 1 .5” diameter

Metal Deck: 

Floors:

3VLI16, composite

Concrete:

Minimum Concrete Compressive Strengths (f' c) Member

28 Day Strength (psi)

Elevated Slabs

4000

Slab-on-Grade

4000

Walls, Piers, and Grade Beams

4000

Interior Concrete Topping

4000

Concrete Exposed to Freezing

4000

Table 2: Minimum Concrete Compressive Strengths

Advisor: Dr. Memari

Page | 19

Spring Thesis


Gravity Loads Using IBC 200, ASCE 7-05, and estimation, the dead, live, and snow loads were calculated for NBP-300. Tables 2, 3, and 4 that follow show the loads used for the analysis presented in this paper compared to the loads used in the original design of NBP-300.

Dead Load

Superimposed Dead Loads Area

Design Load

Floors

15 psf

Roofs

15 psf

MEP

20 psf Table 3: Superimposed Dead Loads

Live Load

Live Loads Area

Design Load

ASCE 7-05 Load

Floors (including partition load)

100 psf

80 + 20 psf

Mechanical Room

125 psf

-

Elevator Machine Room

150 psf

-

Penthouse Floor

150 psf

-

Stairs

100 psf

100 psf

Slab-on-Grade

150 psf

-

Screen Enclosure and Roof Area 60 psf Table 4: Live Loads

60 psf

Note: (-) signifies that there was no recommended value from ASCE 7-05 for the specified loading condition and the design load was assumed to be correct.

Snow Load

Snow Loads Load Type

Roof Snow Load Drift Load

Advisor: Dr. Memari

Design Load

ASCE 7-05 Load

20

17.5

Not available Table 5: Snow Loads

81.94

Page | 20

Spring Thesis


Structural Proposal National Business Park-300 was designed initially as a typical office building. For the purpose of this thesis, it will be assumed that National Business Park- Building 300 will be designed as a Federal office building with the potential to be a target of terrorism. This means that, to protect the life and safety of the occupants of NBP-300, it will be required to be designed for blast loading. The first step in designing for blast is to determine the type and level of threat expected. It was decided that the site would be redesigned and the security of the site increased to eliminate a vehicular explosion from being the controlling blast load. It was then determined that the controlling load would be a small-device, and that it most likely would be an interior explosion. Using these specific design constraints, a typical bay of National Business Park- Building 300 was redesigned to withstand the blast load. This includes the design of a column, a beam, a girder, and the floor slab. A typical moment connection will also be discussed. Due to time constraints, progressive collapse and a lateral system analysis were not completed, but would be investigated in depth if given more time. The intent of this redesign is to show a difference in the amount of structural damage between a building designed for blast and a building not designed for blast, and to determine the effectiveness of minimizing structural damage using blast design and analysis. This thesis will also look at the effectiveness of modeling only critical sections of a building designed for blast. This will be achieved through the use of LS-DYNA Blast Modeling software, a finite element analysis program. Comparisons between the two structures and their reactions to an explosive event will be discussed in detail, and through design and analysis, it will be shown that the linear elastic method is highly conservative.

Advisor: Dr. Memari

Page | 21

Spring Thesis


Blast Design Until the 1960’s, blast design was reserved for facilities where accidental or chemical explosions could occur. Blast design was not considered for ordinary structures. In the 1960’s, though, design guides for blast started to emerge, mostly for buildings with the potential for chemical explosions. In today’s world, terrorism is an unfortunate reality, and designing for the safety of the occupants in high risk structures has become more important. After the Oklahoma City bombing in 1995, blast design gained significant popularity as a design consideration for life safety. Since 9-11, blast design has become a well-sought after design not only for federal and military building, but other high risk buildings such as hospitals, banks, and international business buildings, to name a few, and is an issue of life safety. Blast design has to take into consideration both impact loads from the initial wave front of the blast, and additional time-dependent pressures, which can occur due to thermal effects behind the wave front. Reflected pressures must be taken into account as well as ventilation to relieve built up pressures in confined explosions. In some cases, shrapnel from the explosive container may act as a projectile, and could cause serious damage, such as breaching. Much of blast design is based on the members behaving plastically. Therefore, there are strict limits on their ductility ratios and support rotations. The ductility ratio is “an approximate measure of plastic strain based on the assumption that the curvature in the maximum moment regions increase proportionally with deflection after yielding and plane sections remain plane,” as stated in the “Handbook for Blast Resistant Design of Buildings.” The limit on the ductility ratio makes sure the

member’s plastic deflections are within the allowable range. The limit on the support rotation maintains that tension membrane action that may develop in a member will be limited so as not to cause connection failures. Due to the unpredictable nature of blast design, often times it is overly conservative. Sometime, though, this over conservative nature leads to unreasonable designs or costs. For that reason, structural elements are allowed to be damaged, as long as collapse is prevented. The principles of blast design were researched throughout this redesign and will be presented through calculations, diagrams, tables, and figures throughout this thesis.

Advisor: Dr. Memari

Page | 22

Spring Thesis


Procedure The first step in designing for blast is to decide what the maximum threat on the building will be during the life of the building. As stated in earlier section, in this case, it was determined that an interior blast from a briefcase sized explosive would be the design scenario. To eliminate the possibility of any other type of blast controlling, both building setback and site control had to be determined. This information will be presented in the Site Redesign Breadth. Next, the blast load was determined by using engineering judgment to estimate the type of explosives in the blast. In this scenario, C-4 was used to estimate the blast load. By following the UFC-340-02 (seen to the right in Figure 11) prescribed technique, blast pressures were determined for the specific situation set forth in the problem statement. After determining blast loads, the member designs were completed. A typical column, an interior beam, the exterior girder, and the floor slab were designed for the determined blast pressures. Plasticity was considered in each design. Additionally, a moment connection design was investigated. After the hand calculations were completed for the member designs, LS-DYNA was used to build a partial model of NBP-300. A 3 bay by 3 bay by 3 story model of NBP-300 was built to demonstrate several findings.

Figure 11: UFC-340-02 Cover Page

Firstly, it was a goal of this thesis to show that since the stress in structural members decreases significantly as distance from the bay of detonation increases it is not necessary to show an entire model. The blast forces may not even be significant in bays that are two or three bays from the initiation of the blast in some cases. Additionally, due to the sophistication of the LS-DYNA program and the tedious nature of the input, it was deemed not in the scope of this thesis to build the entire model. All hand calculations, Excel spreadsheets, tables and figures used in the design process will be presented in the appendices of this document, and will be referenced in the detailed explanations in each section.

Advisor: Dr. Memari

Page | 23

Spring Thesis


Blast Load Determination Blast loads are pressure waves caused by the rapid release of energy during a chemical reaction. The wave propagation is spherical in nature and dissipates with distance from the blast initiation, which can be seen in Figure 12 to the right. Although the spherical nature of a blast wave will cause differential pressures along the length of certain elements, such as columns, it can be assumed to be linear for simplification of design and analysis. Generally, the design process using a single

Figure 12: 3D Blast Wave Projections

degree

of freedom method is conservative, so this assumption will still provide a valid design. The first step of the actual design process in blast design requires that the blast event load be determined. As C-4 was assumed to be the explosive compound being used in this analysis/redesign, a determination of a reasonable weight of the explosive needed to be calculated. Since C-4 has properties similar to clay, it was assumed that about 50 pounds of C-4 could fit inside a large briefcase or small rolling suitcase. Since NBP-300 is an office building, it was assumed that the explosives would have to be camouflaged to fit into the everyday environment of the workspace, hence a briefcase or a small business suitcase. The weight of C-4 then needed to be converted into equivalent TNT weight. Not a lot of research has been done with explosives other than TNT to date, so all the tables and charts in the UFC used in blast design were designed around data collected with TNT testing. Much is known about the explosive power of this compound, but not of many others. Therefore, the weight of actual explosives assumed in this scenario had to be converted into equivalent pounds of TNT. The conversion was performed by dividing the weight of C-4 by the weight conversion factor found in Table 6 .1 of “The Handbook for Blast Resistant Design of Buildings,” and an excerpt from the table can be seen in Figure 13 on the following page. The equivalent mass for pressure was used since design pressures would be the critical values for the structural design.

Advisor: Dr. Memari

Page | 24

Spring Thesis


Figure 13: C-4 Equivalency Conversion Factors, from “The Handbook for Blast Resistant Design of Buildings”

The equivalent weight was found to be 36.5 pounds of TNT. Next, the scaled distance needed to be found, which is representative of the front of the pressure wave. The equation

 

was used to find the scaled distance, which equaled to 4.5 ft/lb1/3 assuming a

standoff distance of 15 feet since that would be the average distance of the explosive to the closest structural elements in the lobby. See Figure 14 below for a simplified floor plan of the lobby area. From this information the peak incident overpressure value of 50 psi was found using Figure 6.6 of “T he Handbook for Blast Resistant Design of Buildings”, which can be found in Appendix B.

Figure 14: Simplified Floor Plan of Lobby Area with members to be designed highlighted Advisor: Dr. Memari

Page | 25

Spring Thesis


From these values the peak reflected pressure was found using the UFC figures. Interpolation between Figures 2-100 and 2-97 provided a peak reflected pressure of 117 psi. Peak reflected pressure is important in confined explosions because as the wave front encounters barriers, some of the pressures are reflected off these materials, thereby increasing the pressure created by the blast. This is a critical component for determining design loads. Additionally, the



value was found similarly by interpolation between Figures 2-149 and 2-146

to give a value of 81 psi-ms/lb1/3. These values were then used to find the effective duration of the blast from the UFC. An effective duration of 4.6ms was calculated. This is only half of the battle, though. Since it is assumed, for the purpose of this thesis, that the explosion will be a confined event, the gas pressure must also be calculated to account for the pressure build up behind the initial front that occurs during an explosion due to the heat released during the explosion. Gas pressure loading density was calculated by dividing the equivalent weight of TNT by the free volume of the space. Then, Figure 2-152 of the UFC was used to determine the peak gas pressure. Additionally, the scaled gas impulse was found using Figure 7- 15 of “The Handbook for Blas t Resistant Design of Buildings.” No fragmentation was accounted for in the determination of blast loads because it was assumed that debris would be negligible. As stated previously, the explosives would likely be contained in a briefcase or suitcase, which is assumed to disintegrate from tremendous heat and pressure at the initiation of blast. A summary of the pressure and impulse design values are summarized in Table 6 on the next page and a diagram of the bilinear pulse blast loading expected can be seen in Figure 15 on the next page. The full calculations are provided in Appendix B.

Advisor: Dr. Memari

Page | 26

Spring Thesis


Table 6: Summary of Blast Load Data

Figure 15: Peak Reflected Pressure and Peak Gas Pressure vs. Time

Advisor: Dr. Memari

Page | 27

Spring Thesis


Column Design In blast design, there are many additional factors that affect the design process as compared to a simpler design for service loads only. Certain ductility requirements must be met. Additionally, there are both dynamic increase factors and overstrength factors that have to be considered during design. The dynamic increase factors account for the increased yield and tensile strengths of steel due to rapid loading. The overstrength factor accounts for the fact that steel actually has a higher yield stress than specified by the AISC. The Department of Defense Explosives Safety Board allows the use of this higher yield stress in design and analysis. There is an additional load combination to be checked, also: 1.0B+1.0D+0.25L. Since the controlling load combination in the original design was 1.2D+1.6L for the gravity system, this was again used as the controlling gravity load. Both load combinations were checked to determine the overall controlling design load for the column. In the original design, all connections were shear connections. As stated previously, the lateral system was composed of four eccentrically braced frames, so the columns in the lobby took no lateral load as they were not part of the original lateral system. Although the redesign requires all connections to be moment connections, the controlling load combination for blast does not require other lateral loads to be accounted for due to the probability of having maximum blast load simultaneously with maximum wind or earthquake loads being very minimal. Therefore, the only lateral load applied to the column was the blast load. Another assumption made was that the column would behave plastically during the blast loading. This allows the use of higher allowable strength values described previously. To begin the design, several more assumptions had to be made. Firstly, the column was assumed to be pinned-pinned. This assumption follows the procedure set forth in the UFC to allow for more simplified calculations in the linear elastic analysis procedure. Additionally, due to the orientation of the columns in the bay, it was assumed that only the flange would be loaded during the blast. Realistically, the load would be distributed to both the web and the flange as it would originate at about a 45  angle to the flange face. Finally, it was assumed that the 117 psi peak reflected pressure would be the controlling pressure in this design. The blast pressures were determined assuming that the explosion would detonate in the center of the bay at floor level. As stated previously, the relationship between pressure and distance is not linear, and the closer the blast is to a structural element, the higher the blast and impulse on that member. Although the space will be fully vented and the reflected pressure is unlikely to ever reach 117 psi, the possibility of the blast being detonated much closer to the column exists. This would inflict the most stress and have the highest probability of damaging the structure, so the 117 psi was used to account for the possibility of this situation. This may be conservative, but column failure could result in Advisor: Dr. Memari

Page | 28

Spring Thesis


progressive collapse of the structure or “pancaking” of floors from above , so it is important to design the column to withstand a direct blast load. Once these assumptions were established, the distributed and axial loads on the column could be calculated. Then the maximum moment and shear values were calculated. An unbraced length of 15’ was used. The dynamic yield strength was calculated by multiplying the 50 ksi yield strength by both the dynamic increase factor for yield and the overstrength factor to get 71 ksi yield strength. A preliminary section was selected based on the required section modulus, calculated by dividing the maximum moment by the yield strength, and the requirement that b f /2tf ≤7. A W14x159 was found to be the smallest section that met the required 8.64 in 3 section modulus and the compact section criteria. Figure 16: 3D model of typical bay with column called out The W14x159 was then analyzed to determine if it met all the requirements for blast. This included checking the d/t w ratio, web shear, the plastic section modulus, plastic moment, plastic axial load, slenderness ratio, allowable axial stress, allowable moment, and the Euler buckling load. Modified interaction equations for blast loading and plastic behavior were used to account for the combined axial and bending. These calculations can be found in Appendix C. It was determined that the member met all the design requirements, and so a W14x159 was used as the typical column in the final design. Figure 16 above shows a typical column for the redesign. Additionally, after designing a typical beam and girder, the column was again checked to make sure it had the capacity for the transferred elastic moment from the moment connections in each frame. It was determined to be adequate and a W14x159 was used as the column members in the LS-DYNA model, which will be discussed further in later sections.

Advisor: Dr. Memari

Page | 29

Spring Thesis


Beam Design As for the column design, some assumptions had to be made before jumping right into the design process. The beam was assumed to be simply supported with a maximum support rotation of 2 . Additionally, it was assumed that the blast load would only act over the width of the flange of the beam. This system was not designed as a composite system, so tributary uplift between beams would not be an issue. Had the system been designed as a composite system, the uplift on the floor slab from the blast would be transferred to the beams, and the beams would have to be designed to withstand the larger blast load. Finally, it was assumed that the pressure-time loading on the beam would be the most critical. The reason for this is that the longer duration will have an

Figure 17: 3D model of typical bay with beams called out

overall larger effect on the structure. Additionally, since the probability of the space reaching 117 psi is minimal it can be considered not to be the most critical condition. Two different bay configurations were designed. Initially it was believed that using three infill beams as opposed to two would allow a significant decrease in the weight of the system. Since only the flange of the beams were assumed to be taking the blast load (not a tributary width of floor slab transferring the load to the beam as with composite systems), it was believed that a beam with a smaller flange width would carry less blast load and could therefore be designed as a lighter member. A second design was also performed using the original bay configuration with two infill beams. In the end, it was realized that the larger members obtained in the second design with the original beam layout was a much lighter system even though the beams had to withstand a larger blast load. Figure 17, above, shows the beams that were designed and the final bay layout using two infill beams equally spaced. Due to the strict rotational limitations, many iterations of the design process had to be completed. After performing several iterations by hand, an Excel spreadsheet was set up to calculate most of the values, with the exception of a few values that had to be read from tables in the UFC. An example hand calculation of the first iteration can be found in Appendix and the Excel spreadsheets for the two different bay layouts can be found in Appendix D. To begin the design, both service load combinations and blast load combinations were checked to find the controlling combination. Blast was found to be the controlling load, again using the combination Advisor: Dr. Memari

Page | 30

Spring Thesis


1.0D+0.25L+1.0B. This equation accounts for the direction of the loading, in this case uplift from blast but gravity loads from the superimposed dead loads, estimated self-weights, and live loads. From the controlling distributed load, the moment was calculated. Plastic moment capacity was used to determine the required section modulus for the beam. Additionally, the local buckling requirements were determined. From these requirements, an initial member was chosen. Plastic moment was determined to make sure it was less than the plastic moment capacity of the chosen member. The natural period of vibration was calculated to find the ductility ratio of the beam. Additionally, the strain rate of the beam was calculated to determine if the DIF originally estimated was accurate. The ductility ratio and rotation of the member were calculated and compared to the criteria established at the beginning of the design. Finally, ultimate shear capacity was compared to maximum support shear using the dynamic shear yield stress. Several iterations of this process were completed in order to determine that a W 27x161 was the smallest beam that worked. The ductility ratio of this member was found to be 1.0, which means that the DIF assumed at the beginning of the design was adequate. This means that the maximum deflection is expected to be the equivalent elastic deflection. Also, lateral bracing was checked, and was calculated to be required at every 3 feet along the length of the beam. This requirement will be met by the slab on deck floor system that will produce an unbraced length of zero across the length of the beam. Additionally, rebound of the member was checked. Rebound is the reversal of load and deflection that occurs after the blast wave has passed. The member must have a rebound resistance greater that that specified in Figure 5-13 of the UFC. This allows the member to remain elastic during the rebound phase. It was determined that the member had sufficient rebound resistance. Live load and total load deflections were also checked under service loads only. The beam was found to be adequate in live load and total load deflections.

Advisor: Dr. Memari

Page | 31

Spring Thesis


Girder Design The girder design followed the same design principles and procedure as the beam design, differing only in their distribution of loading. The interior girder was designed using point loads from the reaction values from the second scenario of the beam design. This meant that there were two point loads equally spaced along the length of the beam with combined uplift and gravity, and a distributed uplift blast load acting on the bottom flange of the girder. Again, timepressure loading was determined to be the more critical design pressure as it occurs over a much longer period of time, relatively, than the peak incident pressure. This design may be conservative for the exterior beam, but will be used for simplicity of design.

Figure 18: 3D model of typical bay with girders called out

Again, plastic moment capacity was used to determine the required section modulus for the beam. Additionally, the local buckling requirements were determined. From these requirements, an initial member was chosen. A W30 x116 was chosen as a trial member. The following checks were performed on the member to find if it met the blast requirements: 

Plastic moment capacity must be greater than member plastic moment.



The strain rate of the girder was calculated to determine if the DIF originally estimated was accurate.



The ductility ratio and rotation of the member were calculated and compared to the criteria established at the beginning of the design.



Finally, ultimate shear capacity was compared to maximum support shear using the dynamic shear yield stress.



Rebound resistance of the member must be greater than the required resistance.

Since the W30x116 met the specified requirements, it was chosen as the most economical member for the design. An additional check for deflection under gravity loads only was performed. Both live load deflection and dead load deflection were found to be less than the allowable deflections. The final member chosen for the design was the W30x116. Figure 18, above, locates the girders that were designed. Hand calculations for the girder design can be found in Appendix E. Advisor: Dr. Memari

Page | 32

Spring Thesis


Slab Design Initially, it was intended to design a slab on deck for the redesign. After a few initial calculations, though, it became evident that the capacity of such a slab would not be enough to withstand the timepressure loading caused by the explosion. At this point, it was decided that a two-way slab on shored form deck could be designed to satisfy the requirements instead. It was chosen to design the slab as a Type II Slab as defined by the UFC. This means that rebound reinforcement will not be required because damage will be allowed and accepted in the slab if rebound does occur. Therefore, rebound calculations were not performed on the slab, as they were unnecessary. The slab, needing to withstand a significant uplift force without intermediate supports, was designed with two-way reinforcement at the top. Being the more critical force on the slab, this reinforcement was also used on the bottom of the slab to provide the same amount of resistance if the blast were from above. Calculations for the two-way slab on shored deck form can be found in Appendix F. The design yielded a 16.5” concrete slab on shored form deck held in place by shear studs. This was not designed as a composite system, though, so the shear studs are intended only for holding the slab in place. Figure 19 on the next page shows a detailed section of the slab. Although this slab design provides a solution, there are some problems. This slab is 10” deeper than the original system. This means that the ceiling height on each floor may have to be decreased in order to fit all the mechanical, electrical, and telecom equipment, especially since the girders and beams remained at nearly the same depth as designed originally. Reducing the ceiling height to less than 9’ may make the space feel very crowded and cramped, which would be unpleasant for the occupants. The other solution would be to increase the building height by 10” per floor, which would result in an overall building height increase of 70”. This would add significant cost to the building, and would likely be upsetting to the owner. The system also does not provide the required lateral bracing for the beams. Although lateral bracing can be provided by other means than a floor slab, it will add additional cost to the design. Utilizing the floor slab as lateral bracing for the beams is the most common method of providing stability to beams due to the economy of the design. Therefore, this would be an inefficient design in that respect, as well. In addition to the issue of system depth, there are the issues of constructability and reasonable solutions. A slab with a depth of 16.5” seems highly conservative for a floor system for day to day loads. There is a lot of extra weight and cost to this system as compared to a traditional slab on deck. Many times, it is not economical to design slabs for blast. There are rarely solutions that are both cost effective and provide complete safety to the occupants. For this reason, it is suggested that the redesign utilize a composite slab on deck system. Damage will most certainly occur, including but not limited to cracking and spalling of concrete and possible breach. As stated previously, damage is allowed as long as collapse does not occur.

Advisor: Dr. Memari

Page | 33

Spring Thesis


After further research into slab design, it was determined that since the space would be vented at 5 psi, this pressure could be used to design a floor system. This allowed a second look into slab on deck systems. A composite deck with slab was designed, and it was found that a 3VLI16 with a 4.5” topping would provide adequate strength for the gravity system. These calculations can also be found in Appendix F. This gives a total system depth of 7.5”, which is o nly a 1” increase in system depth from t he original. A 1” decrease in the ceiling cavity depth should not be an issue with coordinating the MEP.

Additionally, this system will be much less costly, both in materials cost and labor. Overall, this system provides an alternative solution if damage is not considered to be an issue. Additionally, further investigation into Aluminum Foam Composite Sandwich Panels , lightweight sandwich panels composed of aluminum foam cores, could be utilized. This innovative material takes advantage of the high stiffness and high compressive strains of aluminum foams. Large deformations of the sandwich panels give them the ability to absorb high quantities of energy. Placed underneath the slab on deck system, this material could possibly provide the necessary energy dissipation to mitigate structural failures thinner floor systems. Cost data was not looked into for this system, so no recommendations can be made on the economy of the system as a whole. With more time, research into this new discovery would have been completed, and a design capitalizing on this technology may have been possible.

Advisor: Dr. Memari

Page | 34

Spring Thesis


Figure 19: Two-Way Slab on Shored Form Deck Detail

Figure 20: 3VLI16 Slab on Deck Diagram taken from Vulcraft 2008 Cut Sheet

Advisor: Dr. Memari

Page | 35

Spring Thesis


Moment Connection Design

Figure 21: 3D model of typical bay with connection location called out

A typical moment connection at the exterior of the building between a beam and a column was designed as part of the MAE requirement for thesis. Figure 21 above shows the location of the moment connection at the exterior of the building. This is important to note because the columns at the interior may need to be upsized to meet the strong column- weak beam requirements. With more time, an interior connection between two beams and a column would have been investigated. In this design, the moment connection must allow the beam to develop its full plasticity. That said, the overstrength factors that are present in the beam calculations are not accounted for in the connection. This is because it is necessary for the connection to be designed to develop the full strength of the connecting members. It was decided to use seismically prequalified connections in this design. This is because seismic detailing has many of the same requirements as blast design. Additionally, the design can be significantly simplified using the prequalified connections specified in the AISC Seismic Design Manual. The column must have a higher plastic moment capacity than the beam in both situations, providing a strong column- weak beam connection. In this type of framing, the intention is to have failure occur in the beam before the column, since column failure could lead to pancaking of floors and ultimately building collapse. This also helps distribute the inelastic deformations to the rest of the structure. Another similarity is that member rotation is limited to .04 radians. Since the beam was designed for this limit, the prequalified connections could be utilized for the design. Finally, seismic design using prequalified connections ensures that ductile limit states will control, which is the objective in blast design. Advisor: Dr. Memari

Page | 36

Spring Thesis


One difference to note between seismic connection design and blast connection design is that there are different compact section requirements. The AISC Seismic Design Manual specifies the limiting ratios for seismically compact sections. These differ somewhat from the blast requirements for compact sections. The blast requirements were met in the design of the column, and therefore engineering judgment was used to follow those requirements rather than the seismic requirements. Additionally, the seismically compact limiting ratios only apply to members of seismic load resisting systems, which this is not. The design of the connection followed the requirements for Special Moment Frames, which are defined as frames “expected to withstand significant inelastic deformations…” in the AISC Seismic Design Manual since the members in blast design are expected to withstand the large inelastic deformations that may occur as a result of blast loading. The guidelines for prequalified connections design was used, which can be found in the specifications of the AISC Seismic Design Manual. The beam limitations in section 5.3.1 of the specification for the AISC Seismic Design Manual were met in accordance with the blast design requirements. Additionally, all column limitations in section 5.3.2 were met. Additional requirements in Table 6.1 were met throughout the design process. Finally, the beam-column relationship requirements of sections 9.6 and 5.4 of the specification were met. The overstrength factor used in seismic design for this calculation should not be used in blast calculations. It was determined that the plastic capacity of the column was greater than the plastic capacity of the beam, and therefore met the requirements for a strong column- weak beam connection. To begin the actual design, a prequalified connection was chosen. Based on the large expected plastic moments, an Eight-Bolt Stiffened connection was chosen, or 8ES. Next, a “dog-bone,” otherwise known as a reduced beam section, was added to the beam to ensure that a plastic hinge occurs in the beam and not at the connection. The reduced beam section was sized according to AISC design requirements. The reduced beam section modulus, reduced beam shear, and the expected plastic capacity upon yield were then calculated. Next, the end plate and bolts were designed. It was found that shear yield in the plate controlled the plate design. The plate was determined to be a 2.25” thick plate. The bolts were determined to be 1.5” A325N bolts. On the column side, it was determined that stiffeners would be required. Web crippling was determined to be the controlling limit state, and stiffener plates were designed accordingly. It was determined that a 1.5” thick stiffener plate was required at the top and bottom beam flange locations. The same stiffener was used in both locations to account for load reversal during rebound of the beam. Full penetration groove welds were determined to be required at all welded locations. It is noted that this is quite costly, but it is also understood that this is the unfortunate downside to blast design. Figure 22 on the following page shows a side and top view detail of the exterior moment connection.

Advisor: Dr. Memari

Page | 37

Spring Thesis


Figure 22: 8ES Moment Connection Detail

Advisor: Dr. Memari

Page | 38

Spring Thesis


Final Design Figure 23 below shows a portion a floor plan of the redesigned structure with members labeled. This was used for the LS-DYNA Redesign model, which will be discussed in later sections.

Figure 23: Partial Floor Plan of Final Design with Members Labeled

Advisor: Dr. Memari

Page | 39

Spring Thesis


LS-DYNA Modeling LS-DYNA Blast is a finite element analysis software used for modeling and analyzing blast events. Although not created specifically for use in the building industry, it can be used to model a building structure to determine member reactions during a blast event. In this thesis, LS-DYNA computer modeling and analysis will be used to meet the AE597A MAE requirements. A partial model of NBP-300 was created in LS-DYNA to compare the results using a blast modeling software to the hand calculations performed in the first half of the structural depth. A partial model, as opposed to the model of the entire building was used for several reasons. Firstly, blast loads decrease significantly with increasing distance from the origin. By building a partial model, it can be shown that members as close as one to two bays away from the blast really don’t see much stress. Therefore, it can be accurately assumed that a partial model will provide adequate stress-strain data. A full model would be required for a lateral analysis though. Second, each member had to be created individually in a text file as opposed to the user interface. This made the process of building the model very tedious and time consuming. The scope of this thesis did not allow for that in-depth of a modeling study. Finally, since LS-DYNA is a finite element analysis software full of dynamic analysis equations and coding, the run time for large, complicated models increases drastically as compared to smaller less complicated models. In order to run several iterations, this time issue had to be taken into account. The model was created using a base keycard file. This file has preset “cards” which are recognized

commands for the LS-DYNA programming. This is where the blast load is defined using the *LOAD_BLAST card. The equivalent weight of TNT, the (x,y,z) location of the blast initiation, and the delay time between the start of the model and the blast initiation were identified to define the blast. This keyword takes into account incident pressures and reflected pressures. The first member, a W14x159 was created in the user interface using a block mesh command (BMESH). Figure 24 on the next page shows an example of the mesh for the beam. Initially a rectangular section was created using five indices in the x-direction, 5 indices in the y-direction, and 6 indices in the zdirection. The distances between indices in each direction were entered to develop the outline of the cross section of a W-flange. Finally, material was removed from the square on either side of the outlined web to create the W-flange shape. This created the “I” shape used for wide flanges. Then the dimensions for the web and flanges were defined and a height was given to the column.

Advisor: Dr. Memari

Page | 40

Spring Thesis


Figure 24: View of Block Mesh used in LS-DYNA with Indices Labeled in “I” and “j” directions

Each member was created using this process. In the cases where beams framed into girders, copes were created by adding more indices in the x-direction and z-direction, and then subtracting out the space where the cope was required. A standard 3” cope was used and every location required for simplicity. Once one member was created in the user interface, the file was saved, and then the text file for that column was opened. From here, each member was created using the format from the first member. The model was checked periodically during the text input to make sure that the geometry created was accurate. A sample of this text is available in Appendix H. The floor slab was created using the block mesh command again. Individual quadrants were created and restrained at interfaces with one another. Cutouts were created at the column corners. This can also be seen in the sample text file. Fixed end moment connections at the ends of the beams and girders were modeled using the *CONTACT_TIED_SURFACE_TO_SURFACE_CONSTRAINED_OFFSET option. Once the model was built and the *LOAD_BLAST card was defined, the model was run. The blast was defined to detonate at 2.5ms into the run time of the model. This process was repeated for the redesigned structure using the blast members. Figures 25 through 28 on the following pages show several isometric snapshots from the LS-DYNA output model, including the frame, the frame with slab, and a connection detail. Advisor: Dr. Memari

Page | 41

Spring Thesis


Figure 25: Isometric view of structural elements of partial LSDYNA Model (slab not shown)

Figure 26: Isometric view of structural elements of partial LSDYNA Model (slab shown)

Advisor: Dr. Memari

Page | 42

Spring Thesis


Figure 27: Isometric view of underside of structural elements of partial LS-DYNA Model (slab shown)

Figure 28: Close-up of Connection in LS-DYNA model

Advisor: Dr. Memari

Page | 43

Spring Thesis


Comparison of Results The results of the LS-DYNA model revealed a lot about the blast design process. It showed that the stresses are lower than expected in the redesign, and that the hand calculations may be too conservative. Plastic strain was never reached in the redesign, so the members should not yield. Spikes in stresses and strains occur as pressures are reflected back into the structure, causing the stresses in the members to build. This is more noticeable in the redesigned structure since it would be stiffer, and therefore would have a less flexible response to the blast load. Von Mises stresses were compared to the allowable stresses for the structural steel in dynamic loading. Von Mises stresses are calculated using the following equation:

            This defines a stress plane based on stresses in the 1 st, 2nd, and 3rd principle stresses. Within this plane, stresses are below allowable. Outside the plane, yield occurs. The Von Mises stresses obtained from the LS-DYNA output should be below the dynamic yield stresses of the structural steel, and were compared to determine the effectiveness of the redesign. In the original structure, the stresses are well below dynamic yield. Additionally, in the model designed for blast, we can see that no failures occur, as the stresses are below dynamic yield stress, or 71,000 psi. This can be seen in Graphs 1 and 2 on the following page. Figure 29 (pages 46-47) shows several still pictures of the blast wave propagation through the structure. As stated previously, the blast was detona ted at 2.5ms into the model. Significant stresses aren’t immediately felt by the members, though. It takes several more milliseconds for the blast wave to induce stresses in the members. Additionally, it can be seen that the pressure distribution over the members is not equal. The base of the column sees the pressure wave many milliseconds before the top of the column. In the hand calculations, it was assumed that the pressure would be evenly distributed along the beam. In reality, this would not be the case, and so the member may be overly conservative in its design. This is true of other members in the bay as well, such as the beams, which would see the blast load at the center of the span before either of the ends. In addition to checking the stresses in the members, the pressure from the blast was checked. It was found that the pressures from the blast were very close to what was calculated by hand. This data was presented in the “Blast Load Determination” section previously. The maximum pressure was found to be 147 psi, which occurred at 7 ms. The blast was detonated at 2.5 ms into the analysis. This drops off quickly and remains between 22 psi and 15 psi between 14 ms and 20 ms. This is very close to the 117 peak pressure and 21 gas overpressure that was found through hand calculations. On the following pages a progression of still images from the pressure analysis video obtained from LS-DYNA output can be seen. Advisor: Dr. Memari

Page | 44

Spring Thesis


Graph 1: Stress in Original Structure

Stress over Time

25000

20000

19116

Columns Beams b/t Columns Infill Beams

) 15000 i s p ( s s e r t 10000 S s e s i M n5000 o V

Girders Slab

0 0

5

10

-5000

15

20

25

Time (ms)

Graph 2: Stress in Blast Designed Structure

Stress over Time 70000 60000 Columns

50000 ) i s p ( 40000 s s e r t S 30000 s e s i M n 20000 o V

Beams b/t Columns Infill Beams Girders Slab

10000 0 0 -10000

Advisor: Dr. Memari

5

10

15

20

25

Time (ms)

Page | 45

Spring Thesis


Figure 29: Blast Pressure Progression

Advisor: Dr. Memari

Page | 46

Spring Thesis

Advisor: Dr. Memari


Page | 47

Spring Thesis


Additional Comments Comments and and Conclusion Lateral System Implications The redesign of NBP-300 will likely have several implications to the lateral system. Originally, as stated previously, the lateral system system was composed of four eccentrically braced frames. frames. Of the changes made during blast design, one of the most important to note is that moment frames would be utilized. Moment frames can provide more ductility than the present eccentrically braced frames in the building, which would help distribute the large amount of energy and pressure produced by the blast. With the increased weight of the structure, the seismic lateral loads would increase, and produce a larger design load in both directions. Seismic load controlled the design design in the east-west direction originally, originally , and would remain the controlling design load. In the north-south north-so uth direction, though, wind loads controlled the lateral design, so this would need to be checked to determine if that were still true. Seismic loading may control in the redesign due to the incredible amount of weight added to the structure. A further look into how ground shock affects the building’s foundations would be the next step to

determine additional strength requirements for the foundations. foundations . This is another area where additional research and design could yield some interesting interesting results, but as it was not in the scope of work for this thesis, redesign of the foundations was not performed.

Effects on Foundations As stated previously, the redesign adds tremendous weight to the structure. This means that the current foundations may not be large enough or strong enough to support the additional weight. Additionally, the lateral loads would increase, so overturning moments in the foundations would have to be rechecked. Finally, due to ground shock from blast, the slab on grade may need additional strength. A further look into how ground shock affects the building’s foundations would be then next step to

determine additional strength requirements for the foundations. foundations . This is another area where additional research and design could yield some interesting interesting results, but as it was not in the scope of work for this thesis, redesign of the foundations was not performed.

Advisor: Dr. Memari

Page | 48

Spring Thesis


MAE Requirements As an MAE student, certain masters’ courses were utilized in this thesis redesign. Overall, three masters’

courses were used throughout this thesis. A moment connection was designed using the information taught in AE 534. Computer modeling was used to model and analyze the original and redesigned structure, using material and information from AE 597A. 597A. Finally, AE 542 information on building façade systems and glazing was utilized to complete a façade redesign.

Advisor: Dr. Memari

Page | 49

Spring Thesis


Breadth Topic I: Site Redesign To redesign the site of NBP-300 several aspects of the site had to be taken into consideration. There was originally parking located on-site, as well as truck access for garbage collection and delivery dropoffs. Several pedestrian access points existed on multiple sides of the lot. Waste water management also had to be considered during during any modification to the site. Native plants and materials to the region were researched. Any plants, trees, or shrubs that were to be planted had to be able to grow in the climate and soil conditions present for the NBP-300 site. Figure 30 below shows a comparison between the original site and the redesigned redesigned site. Firstly, all on-site parking was eliminated. eliminated. This negates the possibility of a car being used as an explosive at a close range to the building. It will be assumed that the parking structure to the north-east north-eas t of the NBP-300 lot has the capacity to take on parking load eliminated from the site. Additionally, Additionally, this lot will also require a security system to prevent it from becoming a hazard. The site was also expanded to meet the setback requirement of 400 ft. This means that at any point on the building, the closest vehicular access way must be at least 400 ft away. Since there was only one road running in front of NBP-300, the building could easily be moved further back on the site. The driveway at the top of the site will remain, but will be for deliveries only and will have a security checkpoint. checkpoint . The problem with moving the building back on the site is that this will add to the cost of the project. An additional 80,856 square feet, or 1.86 acres would would be required to meet the setback rule. This is about a 50% increase from the original size. The cost benefit tradeoff makes this purchase worth the money. If no setback is maintained, then the structure structure would be required to be designed for other, possibly higher, blast loads. An important change to the waste-water management had to be made. The distance to the main hookup was increased with the increase in setback. Fortunately, a clear path between the main run and the building was able to be maintained, and so this wasn’t se en as a critical issue in the redesign.

The other issue that had to be dealt with is the grading of the site. Since NBP-300 has a partial basement that is exposed at one side, it was preferable to keep this architecture undisturbed. After looking at the site grading from the original CAD file, it was determined that the site slopes gradually away from the back side of the building toward the east. Ultimately, it was not seen as an issue for the architectural integrity or constructability of the building. Site access for delivery trucks and maintenance maintenance vehicles was maintained, but moved. moved. A gated access road to the north of the site leads to the loading dock at the northwest corner of the building. This provides a security checkpoint for any vehicle entering the site, and meets the blast requirements for site security. The road loops around to provide both entrance and exit points from site. There is a spacious turn around area and parking parking area for any vehicle that may require require such. The road maintains at minimum a 24’width to provide adequate space for trucks.

Advisor: Dr. Memari

Page | 50

Spring Thesis


Next, a perimeter wall was designed to surround the site of NBP-300. This consisted of an 8’ high cast in place concrete wall. The wall was sized for wind loads and overturning moments. Foundations were sized up to a standard size rather than using the actual calculated sizes, as they were too small for the wall size. A CAD drawing of the wall can be seen in Figure 31 on the following page. In front of the wall, decorative boulders will be used to protect the wall from vehicular impact. More sidewalks and patio areas were added to the site to form a friendlier environment for the occupants. There are three circular paths with bench seating connected by curved paths. This provides a nice outdoor walking space for the occupants. Additionally, there is an outdoor patio with tables, chairs, and umbrellas for those occupants who may wish to eat outside. Birch trees were chosen to line the perimeter wall to obscure the view of NBP-300 from outside people. Birch trees were chosen for their clean look, expected growth height, and ability to grow well in the area. Additional plants, such as annual and perennial flowers would be expected on site in planters and flower beds. These were not specifically detailed as this would be up to the owner and/or occupants. This would have drastic impact on the industrial park as a whole. As it is now, most of the lots are open and unobstructed from view. All buildings are visible from the roads and sidewalks through the industrial park. By adding a site wall, barriers, and tall trees, NBP-300 will become almost invisible to drivers and pedestrians in the park. On one hand, this creates a peaceful and quite atmosphere on site for the occupants. There will be less noise from traffic due to the perimeter wall, and less exposure to the public. With more outside areas, the occupants may feel more comfortable with eating lunch outside or taking a walk around the site after their lunch break. On the other hand, the building had a very unique architecture, even after the façade change that is discussed in the next sections, and it is a shame to hide it from the public eye. Unfortunately, for safety reasons, it is a precaution that must be taken. There is also the possibility of the occupants finding the site and building too closed off from the rest of the park. Both perspectives are important to consider in this redesign. Appendix I contains additional information to the redesign of the site.

Advisor: Dr. Memari

Page | 51

Spring Thesis


Figure 30: Comparison between original site (left) and redesigned site (bottom).

Advisor: Dr. Memari

Page | 52

Spring Thesis


8’-0”

Figure 31: Perimeter Wall Section

Advisor: Dr. Memari

Page | 53

Spring Thesis


Breadth Topic II: Façade Redesign and Heat Transfer In order to vent the interior in the case of an explosion, special care needed to be taken in the design of the façade. It was determined that the glazing would be required to fail, but this presented a challenge. Glazing is not allowed to shatter or become a projectile during blast because it becomes a safety issue. Research into blowout panels and cable catch systems led to some interesting new research in anchored glazing system. This proved to be very valuable information as it allowed the glazing to blowout, but remain in the framing so as not to become a danger to outside persons. The façade redesign utilizes this unique anchored film system. Wet glazing is used at the head of the window, while dry glazing is used around the jambs and sill. The reason for the two types of glazing because the glass is supposed to dislodge from its supports so as to provide a pressure release for the interior. The top of the glass must remain in its supports so as not to become a dangerous projectile.

Figure 32: Comparison between original façade (left) and redesigned façade (bottom)

Advisor: Dr. Memari

Page | 54

Spring Thesis


Façade Design The façade redesign presented many challenges. The need to ventilate the interior space during the explosion combined with the strict provisions against glass debris made finding a so lution somewhat tricky. Initially, the intent was to design a system that would allow the glazing and framing to be forced from the wall system at 5 psi, but would be tethered to a short retractable cable system so as not to become debris. This proved to require more research than the tight schedule permitted, and so the idea was put on the “back burner” so to speak. It is believed to be a viable system, but would require testing and

significant research, which was not in the scope of this thesis. Instead, anchored glazing film systems were researched. This proved to be a much more constructible system that would likely be significantly cheaper. The glazing needed to be designed to withstand estimated wind pressures, but needed complete “wet blanket” failure at 5 psi in order to properly ventilate the interior space during the blast. A “wet blanket” failure is a failure mode of glass where the pane remains in tack but loses all stiffness and load

carrying capacity. It behaves as a blanket and becomes very pliable. The glazing in the façade was sized to withstand the design wind load calculated in Technical Report 1, or 23 psf. An IGU (Insulating Glass Unit) unit composed of two lites of FT (fully tempered) glazing was decided on as the glass type for the redesign. The reasons for this choice were two-fold. One, this type of glazing would provide the desired failure mode. Two, this glazing would provide good resistance to heat loss and gain through the façade. A GTF of 4 was used since FT glass was used. Load sharing factors were used, but it was decided to use the same thickness for both lites, so LS1 and LS2 were determined to be 2.0 each. A 3/8” air space was used between the two lites of glass. The overall design yielded a glazing system of two lites of ¼” FT glass. The overall thickness of the system was determined to be 7/8” thick. This glazing was then verified to have less strength than the blast force of 5 psi, or 720 psf. With only 25.08 psf capacity, the window system will definitely fail before the 5 psi blast load is reached. Next, the wet glazing needed to be sized in order to provide adequate strength in tension so as not to come lose when the glazing fails. This is the anchorage part of the system. Through research, it was found that GE provides high strength glazing, specifically for use in blast design. From their manufacturers catalog it was determined that UltraGlaze SSG4000 could be used as the wet glazing. It has a tensile strength of 340 psi, well over the 5 psi blowout pressure and the peak pressure of 117 psi determined in the blast analysis. A ½” bead of the UltraGlaze SSG4000 will ensure that the top of the glazing will remain in its joint even when the other three sides of the glazing are dislodged. Additionally, GlassLock wedge gaskets will be used at the top to restrain the glazing in its joint as part of the anchorage system.

Advisor: Dr. Memari

Page | 55

Spring Thesis


Finally, and most importantly, an 8mil safety film was applied to the interior of the IGU. This is twice the required thickness for post-breakage retention of glazing. The factor of safety of 2 was used to account for tearing due to “flapping” motion of the broken panels. This would, of course, need to be verified with testing, but such was not in the scope of work for the redesign of the façade system. It should be noted that the connections were not designed for the new façade, but would be required to withstand over 5 psi of pressure since that is the highest vented pressure that would occur. Figure 33: Glazing Anchor System Detail

The spandrel area at each floor level was changed to the precast panel system used on the un-curved portions of the façade. This allows for a better anchorage system at the top of the window. The corners were left fully glazed to create architectural interest. The anchorage system would still be used at these locations, but would have to be installed at the horizontal mullions in each glazing panel. The solid glass column at the corners eases the building into the surroundings. The precast panels create a very harsh

Advisor: Dr. Memari

Page | 56

Spring Thesis


feeling, so the glass creates a nice transition between the harshness of the concrete and the softness of the surrounding nature. A comparison between the facades can be seen in Figure 32, and a detailed section through the head of the glazing system can be seen in Figure 33. Hand calculations and additional information for the façade redesign and glazing system design can be found in Appendix J. Additionally, the work performed for the redesign meets MAE requirements.

Heat Transfer Heat loss and heat gain through façade systems are of great significance to building systems. The more heat that is transferred through the wall system, the more energy is required to condition the inside environment to mitigate the transfer. Poor façade and glazing systems can cause a lot of energy costs, not to mention unhappy occupants. After the façade system was decided upon and the glazing sized, heat transfer through the spandrel and glazing areas were calculated. This was done for both mid-summer (July) and mid-winter (January), when outside temperatures are at their extremes. Weather data for Annapolis Junction, Maryland was research. It was found that the extreme high expected in the summer would be 88 F, and the extreme low in the winter would be 24 F. These values were used for the outside air temperatures to calculate the heat loss and gain due to air temperature differences. A general wall section for the precast spandrel area was decided upon. After choosing the materials and components of the wall section, HAM was used to find the best location for the vapor barrier. It was determined that the vapor barrier would be best located between the air cavity and the plywood sheathing. No condensation occurred in the wall system using this configuration for either winter or summer. This data can be viewed in Appendix K. One assumption that was made during this process was that the heat gain and loss values would be calculated for one half of a single floor level. This was for ease of calculations and to determine if the output data seemed reasonable. The values presented in the following tables would be required to be multiplied by 2 to get the heating and cooling loads per floor level. The equipment would also need to be sized according to the adjusted loads. R-values for each component of the wall were found using tables provided in “ Mechanical and Electrical Equipment for Buildings.” HAM was used to check the values. The total R-Values were very close. The data from HAM for the R-Value Analysis can be found in Appendix K. Table 7 on the next page summarizes the R-Value data.

Advisor: Dr. Memari

Page | 57

Spring Thesis

Rebecca Dick | Structural Option R-Values

Component

R

Inside Air Gyp Board

0.68 0.32

Batt Insulation plywood

15 0.62

vapor retarder

-

Air Cavity

0.68

Rigid Insulation

11

Precast Concrete

0.87

Outside Air

1.35

IGU Glazing

3

(two FT lites with film) Table 7: R-Values of Spandrel Wall Section An Excel spreadsheet was used to calculate the U-value of the wall assembly and the glazing in the winter. The U-value of the wall assembly and the glazing was used to calculate heat loss in the winter. Tables 8 and 9 below show the calculations for heat loss through the wall components and overall heat loss in the winter. A total heat loss of over 77,000 Btu/hr is expected in the winter, or 154,000 Btu/hr per floor level..

Component Room Air

Uconcrete=

0.0328

Uglazing=

0.3333

Winter Thermal Gradient R R

R/RT

T

-

-

70.0

Inside Air Film Gyp Board

0.68 0.32

Batt Insulation

15

0.68 1

1.0 1.5

69.0 68.5

16

24.1

45.9

0.62

16.62

25.0

45.0

-

16.62

25.0

45.0

0.68

17.3

26.1

43.9

11

28.3

42.7

27.3

Precast Concrete

0.87

29.17

44.0

26.0

Outside Air Film

1.35

30.52

46.0

24.0

Plywood Vapor Retarder Air Cavity Rigid Insulation

Outside Air 

Advisor: Dr. Memari

-

-

-

24.0

30.52 Page | 58

Spring Thesis

Rebecca Dick | Structural Option Table 8: Winter Thermal Gradient Winter Heat Loss

A=

4575

Uconcrete=

0.0328

Uglazing=

0.3333

ft 2

Twinter=

46

F

Tsummer=

18

F

through concrete

through glazing

6895.5

70150

sensible heat loss(winter)-=

latent heat loss can be neglected due to high efficiency vapor retarder Total:

77045.5 Table 9: Winter Heat Loss

Btu/hr

Summer presents different heat transfer issues than winter. In the summer, heat gain through a wall assembly must take into account radiant heat from the sun, outside air temperature, electrical heat loss from lighting, equipment, and people. These factors were taken into account during the heat gain calculations. Additionally, latent heat gain was estimated as a percentage of the sensible heat gain and was included in the total expected heat gain. Tables 10 and 11 below show the data used for the summer heat gain calculations.

Summer Thermal Gradient Component Room Air

R

R/RT

T

-

-

-

75.0

Inside Air Film

0.68

0.68

0.3

75.3

Gyp Board

0.32

1

0.4

75.4

15

16

6.8

81.8

0.62

16.62

7.1

82.1

-

16.62

7.1

82.1

Air Cavity

0.68

17.3

7.4

82.4

Rigid Insulation Precast Concrete

11 0.87

28.3 29.17

12.1 12.4

87.1 87.4

Outside Air Film Outside Air

1.35 -

30.52

13.0

88.0 88.0

Batt Insulation Plywood Vapor Retarder

Table 10:

R



Advisor: Dr. Memari

-

30.52

Summer Page | 59

Spring Thesis

Rebecca Dick | Structural Option Thermal Gradient Summer- Heat Gain

A=

4575

Uconcrete=

ft2

0.1700

Uglazing=

0.81

DETD=

10

DCLF=

56

IF=

1.1

N=

300

people

54600

Btu/hr

9600

Btu/hr

computers

6000

Btu/hr

printers

10000

Btu/hr

copiers

2000

Btu/hr

misc.

Plights= Pequipment=

Sensible Heat Gain: Envelope:

2567

Btu/hr

Glass: Outdoor Air:

171654

Btu/hr

5033

Btu/hr

People:

69000

Btu/hr

Lights Equipment:

54600 27600

Btu/hr Btu/hr

Total:

330453

Btu/hr

Total:

66091

Btu/hr

Total Summer Heat Gain: 396544 Table 11: Summer Heat Gain

Btu/hr

Latent Heat Gain:

Based on the calculations, the HVAC system would need to be sized to accommodate about 400,000 Btu/hr of cooling in the summer and about 77,000 Btu/hr of heating in the winter. This seems reasonable since there is floor to ceiling glazing on each floor level, so high heat gain would be expected in the summer and high heat loss would be expected in the winter. Advisor: Dr. Memari

Page | 60

Spring Thesis


Cost Estimates Minor cost estimates were performed for this redesign. A comparison of cost between an original typical bay and the redesigned typical bay for blast can be seen in the Table 12 below. It is apparent that the cost of design for blast is tremendously larger than a typical design, with a price increase of over 40% from the original. The increase in member sizes combined with the labor costs creates a much more expensive design. The original cost of construction of National Business Park- Building 300 was $38 million dollars. With a price increase of 40%, the new cost of NBP-300 would be $53.2 million dollars. This is a significant price increase, but if life safety is the issue then the extra cost would be well worth it. This estimate only includes the structural steel in a typical bay. It does not account for the façade redesign, the additional cost from the site redesign, or the cost of connections or foundations. A much more detailed cost estimate would show a truer comparison of costs, but was not in the scope of work for the breadth topics.

Advisor: Dr. Memari

Page | 61

Spring Thesis


COST ANALYSIS OF STRUCTURAL MEMBERS ORIGINAL DESIGN

MEMBER

LENGTH ( ft)

NUMBER OF MEMBERS

W14X99 W14X159* W27X84 W27X94 W27X102*

15 15 35.1 40.9 40.9

2 2 1 2 3

TOTAL COST PER LINEAR FOOT

TOTAL W/ O&P (PER LINEAR FOOT

$ $ $ $ $

$ $ $ $ $

142.02 225.00 120.64 133.64 144.94

158.10 252.00 134.00 149.00 161.00

TOTAL

TOTAL

TOTAL W/ O&P

$ $ $ $ $

4,260.60 6,750.00 4,234.46 10,931.75 17,784.14

$ $ $ $ $

4,743.00 7,560.00 4,703.40 12,188.20 19,754.70

$

43,960.95

$

48,949.30

TOTAL

TOTAL W/ O&P

$ $ $

13,500.00 36,940.88 11,566.85

$ $ $

15,120.00 41,063.60 12,776.40

$

62,007.73

$

68,960.00

BLAST REDESIGN

MEMBER

LENGTH ( ft)

NUMBER OF MEMBERS

W14X159* W27X161 W30X116

15 40.9 35.1

4 4 2

TOTAL COST PER LINEAR FOOT

TOTAL W/ O&P (PER LINEAR FOOT

$ $ $

$ $ $

225.00 225.80 164.77

TOTAL

252.00 251.00 182.00

% INCREASE FOR REDESIGNED STRUCTURE:

141%

141%

NOTE: Members wi th (*) were interpolate d/extrapolat ed from data in RSMeans 2 012

Table 12: Cost Data for Structural Steel in a Typical Bay

Advisor: Dr. Memari

Page | 62

Blast Design and Analysis

Recommend Documents