MI-62 HEAT EXCHANGER STAND

(Part 1)

ABSTRACT

The goal of this project is the design of a stand capable of safely supporting a large shell-and-tube heat exchanger in the correct position. Furthermore, the purpose of this report is to show an overview of the steps taken in the design of the different components of the stand and to show the final product.

INTRODUCTION

Fermilab hosts some of the largest and most intricate machines in the world. These machines need to be cooled and maintained at an adequate operational temperature. Some of this cooling is done with water, which has to transfer heat from the machines to the ponds in several stages. The pond water does not cool the machines directly. A heat exchanger is used to cool clean, de-ionized water with pond water. Then, the clean, de-ionized water-cools the machines. Because of the use of pond water in the heat exchanger, a shell-and-tube heat exchanger has to be utilized. This type of heat exchanger is large and heavy, and has to be placed in a special position. A stand capable of sustaining the loadings from the heat exchanger at the correct position has to be designed.

dimensional requirements

The shell-and-tube heat exchanger to be installed in the MI-62 building has to be placed in the correct position. That is, it should be placed longitudinally against the wall, allowing enough space for a man to walk between the heat exchanger and the wall. The middle-line of the heat exchanger has to be placed at a height of 75 ½ inches from the floor. The tube side of the heat exchanger will have pond water circulating through it. Therefore, the end of the heat exchanger with the tube nozzles has to face the sliding door, for cleaning purposes. These are some of the dimensional requirements for the stand, they will be better appreciated with the drawings shown later in the report.

SECTIONS OF THE REPORT

The report has been divided into four sections to make it easier to read. These are the following

1. Column Member Selection

2. Welding

3. Base Plate Design

4. Final Product

Back to index

COLUMN MEMBER SELECTION

force analysis

The most obvious force that has to be taken into consideration is the weight of the flooded heat exchanger. Nevertheless, other forces may act upon the exchanger as well, like the vertical and horizontal forces exerted by the tubes and water flow on each nozzle. Consideration must also be made for a side force that may be caused by an earthquake. Although Fermilab lies in a zone 0 seismic area, a 15% of the heat exchanger weight was used.

The values used for the forces are as follows:

Fv = 5000 # “vertical” forces

Fh = 5000 # “horizontal” forces

W = 10180 # weight

S = 1527 # side force

The vertical and horizontal forces received a number according to the nozzle they act upon. These forces can only change sense, that is horizontal forces either go left or right, and vertical forces either go up or down. It is important to say that rotational moments at the nozzles are assumed to be negligible. The weight and side force act on the center of gravity. The side force either goes left or right. Left and right are italicized because they are relative to the frame of reference used. Two frames of reference were used to solve this problem.

Figure 1. The freebody diagram of the side view of the heat exchanger.

Figure 2. The free body diagram of the front view of the heat exchanger.

The detailed hand sketches showing the dimensions used for the free bodies and the stand can be found at appendix I. If the appendix takes a long time to load, try saving the file into your hard drive and then open it with Adobe Acrobat. From these two free body diagrams, the maximum values for the different forces can be calculated by applying the conditions of static equilibrium, which are the following.

å Fx = 0

å Fy = 0

å M[about any point] = 0

Both free body diagrams have three unknowns. The three conditions showed above give three different equations for each diagram. We have three equations and three unknowns; consequently the variables can be solved. In spite of the apparent simplicity of this system, the fact that the horizontal and vertical forces can change direction will yield multiple solutions for the system. Hence, Scientific Workplace built in Maple software was used to get the maximum and minimum values of each force. The document that shows the solutions for the two frames of reference is available at appendix II.

The following table shows the maximums and minimums for each force divided by two. Please note that the minimum values had a “minus” sign. Therefore the affected member would be under tension. The table also shows the affected members. To clarify this point, please refer to the figure 3 and try to relate it to figures 1 and 2.

FORCE	MAXIMUM COMPRESSION / 2	MAXIMUM TENSION / 2	MEMBERS AFFECTED
Fa	23 386 #	19 822 #	I & II
Fb	9 149 #	4 059 #	III & IV
Fc	15 222 #	15 222 #	V & VI
Fd	21 527 #	21 527 #	VII & VIII
Fab1	13 471 #	8 381 #	I & III
Fab2	21 263 #	15 173 #	II & IV

I. Vertical column @ NW corner

II. Vertical column @ NW corner

III. Vertical column @ NW corner

IV. Vertical column @ NW corner

V. West longitudinal brace

VI. East longitudinal brace

VII. North lateral brace

VIII. South lateral brace

IX. South horizontal member

X. North horizontal member

Figure 3. Isometric drawing showing the different elements of the stand.
Material Selection

Due to the fact that the built-in bases of the heat exchanger are 4 inches wide, 4 X 4 steel structural column was preliminarily chosen. The thickness however was not yet determined. According to the 9^th edition of the Steel Construction Manual by the American Institute of Steel Construction a thickness of 3/16” with a yield stress of 46 ksi of the column mentioned above was adequate for the specified loadings and members’ lengths. But still other considerations had to be made, including corrosion and the bending moment on members IX and X.

Figure 4. Forces acting on members IX and X.

f_allowable < 0.66 f_y

M_max = (F ) (L)/4

f_max = (M_max) (y_max)/I

Where,

M_max à maximum bending moment

F à Force applied at the center of the beam

Là Length of the beam

f_maxà Maximum flexural stress

y_maxà Maximum distance from the centerline of the beam’s cross section

Ià Moment of inertia of the beam’s section

f_yà Yield stress

Using the conservative diagram and the formulae shown above and the data from the Steel Construction Manual (AISC, 3-43) it was easily shown that the 3/16 thickness was not suitable for the top members (IX and X). Therefore, an appropriate thickness of 5/16 was chosen. Ultimately, the drafting department changed it to 3/8 since that thickness is easier to obtain from distributors.

Back to index

Back to top

WELDING

At this point, a person without welding experience would just say, “Well, just weld everything together.” However welding is not as easy as it looks. Several design considerations have to be made to get everything welded together. Next, calculations had to be made to ensure that the weld is strong enough to support the loadings on the joints.

Different kinds of welds had to be used to weld all the members together. In spite of the differences, all the welds made had an average strength of a ¼” fillet weld. According to the AISC Steel Construction Manual Volume II (2-29, 2-59), the formula for calculating the strength of a fillet weld is the following:

P_allow = 0.707 * W * L *(F_y * 0.3)

Where

P_allow àAllowable loading

W àLeg length

L à Length of the weld

F_y à Weld material yield strength

Using this formula the minimum force of a welded connection between any two members can be obtained. The weakest welded connection is the one between two parts of the 4 X 4 column joined in a straight angle. The strength of this connection, leaving 7/16” free on each corner, is 50 kips. This is more than enough to pass the loadings from one member to the other.

Despite the simplicity of this explanation, more involved calculations were made to ensure there was enough force on each welded connection. Including the determination of the radius of curvature of the corners [appendix III] of the 4 X 4 column, which was necessary to determine the maximum size and resulting strength of the flare-bevel groove weld. However, all the different kinds of welds used yielded a strength greater than the strength of a standard ¼” fillet weld. Therefore, the explanation in the paragraph above is valid.

In conclusion, the types of weld used were the following. Please refer to the diagram below to observe the places they were used. Note that continuous lines denote visible welds [in the picture] and dashed lines denote welds that are not visible in this view. The base plates’ welds are the only ones not shown here. Those are welded by ¼” fillet welds, leaving ¼” un-welded at each corner.

Figure 5. Illustration of the different kinds of welds used.

Back to index

Back to top

BASE PLATE DESIGN

TWO DIFFERENT DESIGNS

The designs below were chosen for the base plates. The base plate with six holes is designed for members I, II, VII, and VIII. The rest of the members use the base plate with four holes, for they are expected to sustain lower tensile load. The four-hole plate was designed originally for all the members. However, it was realized that four anchors were not strong enough to sustain the expected tensile loading for some members.

The six-hole design can be either used for straight members I and II or slanted members VII and VIII. However, members VII and VIII have to be connected with the base plate in the way shown in figure 7.

Figure 7. Six-hole base plate joined to a slanted member.

DETERMINING THE BASE PLATE THICKNESS

The American Institute of Steel Construction recommends that base plates should be analyzed as cantilevered beams. The most susceptible part of the plate is where there is the largest distance between the bolt and the column.

Using the following formulas the solution may be obtained:

f_allowable < 0.66 f_y [the base plate material shall have a yield stress of 46 ksi]

M_max = (F ) (L)

f_max = (M_max) (T/2)/I

The last formula however could not be valid given that the moment of inertia varies with the position “L”. Appendix IV analyzes the problem; the graphs shown below were extracted from the same appendix.

Figure 8. Flexural stress in ksi versus position “L” (in inches) for a base plate thickness 5/8”.

Figure 9. Flexural stress in ksi versus position “L” (in inches) for a base plate thickness ¾”.

In this case, flexural stress does not behave linearly. However, the maximum value still occurs where the maximum moment occurs, that is at the edge of the column. Flexural stresses of 22 and 14 are acceptable, for they are below 30 Ksi (.66Fy = 30.36). Therefore the plate thickness can be 5/8” or ¾”

ANCHORS

In the design of the base plates, it was also necessary to take into account the anchor resistance. Several times the hole positions had to be changed to meet the minimum required distances between the anchors. The base plate dimensions had to be changed accordingly.

Five eights Hilty adhesive anchor rods were chosen because of their high strength. Several factors had to be taken into account in the selection of the anchor such as the concrete strength and depth. Anchors do not always have the same allowable tensile and shear loads. The allowable loads diminish with two factors:

1. Distance between anchors.

a. Spacing for maximum load 5”

b. Minimum spacing 2 ½” (reduce both shear and tensile allowable loadings by 30%)

2. Distance from edge.

a. Spacing for maximum load 7 ½”

b. Minimum allowable edge distance 2 ½” (reduce shear 50%, tensile 40%)

The maximum working loads for each of these anchors are tensileà 4520 # and shear à 3000 #. By measuring the distances between anchors for different plates, the reduction factors above can be applied. Linear interpolation is required to obtain the exact reduction factor. The resistance of the anchors was more than enough to sustain the tensile forces on each member.

Back to index

Back to top

FINAL PRODUCT

At the time of the writing of this report, the blueprints for the construction of the stand were ready. Since the blueprints were made according to the specifications of this report, this is how the stand should look like. These are pictures from a model created in IDEAS computer aided design software.

Figure 10. The stand with the heat exchanger on top.

Figure 11. A wire-frame detail of figure 10, were the base of the heat exchanger joins to the stand.

Figure 12. Wire-frame detail showing the other base of the heat exchanger sitting on the stand.

Figure 13. Wire-frame detail of one of the bases of the stand.

Figure 14. Detail of one of the bases of the stand.

Back to index

Back to top