1
$\begingroup$

I am an instrumentation engineer and I need your help with designing a support structure for a gas-measuring (precision) instrument. The instrument works in an emitter-detector configuration, and measures gas concentration inside a duct. Alignment is key and the detector must always see the emitter. Thermal expansion is the enemy. I am using bellows to create a cushion between the duct and the instrument - the idea is that as the duct heats up, the movement of the duct will minimally interfere with the alignment of the instrument. The emitter and detector each weigh 20lbs. They sit 90" above the ground.

Please refer to the drawing that has 'A', 'B' and 'C' marked. I need your help with the following:

A: the length, width and thickness of the floor plate, and the size of the bolts. Without much knowledge of structures, a 12 x 12 would probably work but that will be the absolute max that can fit at that location. In other words, space is at a premium.

B: size of the square tube. I was advised to use square tube by a mechanical engineer.

C: the purpose of this gusset is to prevent warping or bending of the flat plate that mates with the duct; it's referred to as "process connection" in Drawing 1. I had two designs in mind: one gusset in the middle vs two on the edges, on each side of the plate. I decided to go with the latter as there is more steel to prevent warping. What are your thoughts? I am mounting 50 instruments on 50 ducts so if I can optimize steel usage that will save me money.

PS: Some of you might think this is an overkill for mounting an instrument. This is my third design iteration, and the previous two "minimally invasive" designs failed in giving continuous reliable data. At first, I had mounted the instrument directly to the duct, without any support structures, etc.

Drawing 1 Drawing 2

$\endgroup$
6
  • $\begingroup$ Can you quantify the alignment needed? And how did the previous designs fail? $\endgroup$
    – Pete W
    Oct 8, 2021 at 3:37
  • 3
    $\begingroup$ I would consider attaching them to each other with a C shaped frame, rather than attaching each sensor to the floor. The gussets are only important to the extent that the duct will be applying a load to the sensors. If you want things to stay aligned over time I would try to make sure the duct will not apply pressure to either sensor even if it shifts slightly. $\endgroup$
    – Drew
    Oct 8, 2021 at 4:22
  • 1
    $\begingroup$ The idea with the bellows and the flanges is good. What Drew said about the C-shape. You either need a way to adjust the position the sensors, manufacture and mount the whole setup with really high precision. $\endgroup$
    – mart
    Oct 8, 2021 at 8:53
  • 1
    $\begingroup$ I am guessing that the duct is already elevated by some form of support structure. Would it be possible to get with their people and design attachments to the supports that will perform both functions? At least that way if there is ground movement of the supports, your detectors would follow. If these dectectors are deemed required, I would expect this addition to become common. You also have the benefit of their Structural People making the system secure. $\endgroup$
    – Jim Clark
    Oct 8, 2021 at 15:32
  • $\begingroup$ @Drew Thanks for your feedback. Can you give me an example of a C-shape frame? a rough sketch maybe? $\endgroup$
    – vasiqshair
    Oct 11, 2021 at 3:05

2 Answers 2

1
$\begingroup$

This is my suggested details for the post and the base plate with sufficient safety margins (for the static load as shown). I would weld all around the tube to the base plate ($F_y = 36 or 50 kai$). If you have significant torsion, please check the torsional stress against yield by the formula $\sigma_t = \dfrac{F_t*e}{J}$, in which $J = 4.28 in^4$ for the 3x3 tube. I have no suggestion on the gusset though.

enter image description here

Note: this is an update from the previously suggested 2x2x3/16, which will be too flexible for its height (90"). You may consider a larger baseplate for the ease of anchor bolt installation.

This is valid for the INDOOR application only. A Lateral tie is recommended.

$\endgroup$
5
  • $\begingroup$ Thanks for your feedback. It is actually an outdoor application. Also, can you tell me what A, Sx and Fy are? $\endgroup$
    – vasiqshair
    Oct 11, 2021 at 3:32
  • 1
    $\begingroup$ A is the net cross-sectional area of the tube. Sx is the section modulus, which is required to check the bending stress, fb = M/Sx. Fy is the yield strength of the tube or plate (46 kai & 36 or 50 ksi respectively). fb is usually kept around 60% of yield strength, fb <= 0.6*Fy. $\endgroup$
    – r13
    Oct 11, 2021 at 3:56
  • 1
    $\begingroup$ Forgot to mention, for the outdoor application, I would be concerned with wind pressure, which could be a killer for your project. Without specific information, I strongly suggest tying the top of the post to a fixed (non-moving) structure/object. Also, are you at a seismic active zone? It is a critical factor for both indoor and outdoor applications. I really wish that you have a structural guy on your team to consult with. $\endgroup$
    – r13
    Oct 11, 2021 at 4:55
  • $\begingroup$ I am in a seismic active zone. You are right, it is worth consulting a structural engineer, especially since I am mounting 50 of these. Thanks for all your help. $\endgroup$
    – vasiqshair
    Oct 12, 2021 at 16:18
  • 1
    $\begingroup$ You are welcome. Good luck. $\endgroup$
    – r13
    Oct 12, 2021 at 16:31
1
$\begingroup$

If we were to deal with this as a post with just the weight of the instrument, and say we would estimate the bolts at 10 inch of center, we would have a moment of
$$ M=20* (14+4)/2 lbs inch$$

and a pull up force on the bolts of

$$180/10/2=9lbs \ each \ bolt$$

and we could multiply loading by 3 as a dynamic load factor and be done with it.

But the vibration of your hot pipe could resonate with that of the post and throw all the readings off.

so we need more information about your pipe section, its supports, and the natural frequency of those to design the post with a natural frequency away from that.

$\endgroup$
1
  • $\begingroup$ Thanks for your feedback. I have not noticed significant vibration in the duct. With that been said, it'd be a good idea to measure it. Also, can you provide information about the numbers you have used. It is not clear to me what 14, 4, 180 are. $\endgroup$
    – vasiqshair
    Oct 11, 2021 at 3:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.