-
LOCATION:HOME >> NEWS >> Industry news
-
Industry news
The Wind Speed That Breaks a Stack Is Not the Fastest One
Design for the gale, check the breeze
A freestanding exhaust stack is checked against the design gust. Take a 600 mm outside diameter, a 6 mm wall and 18 m of height, all inside the round tube range of 8 to 680 mm at 0.5 to 75 mm wall. At 40 m per second the dynamic pressure is half of 1.225 times 40 squared, about 980 Pa. Over 10.8 m2 of projected area that is 10.6 kN, a root moment of 95 kN.m and a bending stress near 58 MPa. Against the 205 MPa minimum yield on the 310S data sheet there is a factor of three and a half in hand. The stack then cracks at a weld toe in its second winter, in a wind of 5 m per second that nobody calculated.
The wind has a rhythm as well as a force
A crossflow past a circular cylinder sheds vortices alternately from each side. The shedding rate is a Strouhal number times velocity over diameter, with the number near 0.2 for a round section here. At 5 m per second and 0.6 m that is 0.2 times 5 over 0.6, or 1.67 Hz. Every shed vortex changes the side force, so the stack is pushed left and right at that frequency. This is a crosswind motion, at right angles to the gust the structural check was written for, driven by a steady wind rather than a turbulent one.
Work out the stack's own frequency
For a cantilever the first mode frequency is 1.875 squared over 2 pi times the root of EI over m times L to the fourth. From the 310S sheet E is 200 GPa in tension and density 9.01 g/cm3. A 600 by 588 mm annulus has I of 4.94 by 10 to the minus 4 m4 and a mass of about 101 kg per metre. With L at 18 m, EI is 9.87 by 10 to the 7 N.m2, m times L to the fourth is 1.06 by 10 to the 7, and the root of their ratio is 3.05. Multiply by 0.56 and the first mode sits at 1.71 Hz.
Two numbers meet at five metres per second
Set the shedding rate equal to 1.71 Hz and solve for velocity: 1.71 times 0.6 over 0.2 is 5.1 m per second. That is 18 km per hour, an ordinary afternoon. In a band around it the shedding locks to the structure's motion instead of following the wind, so amplitude builds rather than passing through. The gust check at 40 m per second never visits this speed; the stack spends hours there every month. What fails is not strength but repetition: 1.71 cycles per second is 6,100 an hour, near 49,000 in an eight-hour blow, and a few seasons put the weld toe past a million cycles.
Making the wall thicker does not move the critical speed
The instinct is to add thickness. For a thin circular shell the second moment of area goes as diameter cubed times wall, and mass per metre goes as diameter times wall, so their ratio, the only place wall thickness appears in the frequency, cancels. Double 6 mm to 12 mm and I doubles, m doubles, and f stays at 1.71 Hz. The critical wind speed does not move at all. Thickness buys bending strength against the 40 m per second gust that was never governing; it buys nothing against the 5 m per second one that is.
Grade behaves the same way. 201 stainless carries an elastic modulus of 197 GPa against 200 for 310S, a 0.75 per cent drop in frequency and no useful change in critical speed. E, geometry and height set the frequency, not the alloy.
The two levers that do work
Height is the strong one: frequency goes as one over L squared. Bringing the free length from 18 m down to 12 m with a guy wire set or an intermediate frame multiplies f by 2.25, to 3.84 Hz, and lifts the lock-in speed to about 11.5 m per second, above most everyday winds. Diameter is the other: f rises roughly with D, and critical velocity goes as D squared, so an 800 mm stack of the same wall shifts the band to about 9 m per second. Both trade a support or a larger tube for a frequency, cheaper than a thicker wall.
The 304 and 316 rope is one way to buy the shorter length; a square section is not, because its shedding number is different and corner vortices are less predictable. The guy anchors need a welded lug that is not itself the new fatigue site.
Break the shedding instead of the stack
Three helical strakes at about a tenth of a diameter high and a five-diameter pitch stop the vortices forming in step along the length. They are folded on a press brake and welded on site. A perforated screen round the top does a similar job by spoiling the flow. Either way the fix is aerodynamic, and neither shows up in a wall-thickness calculation.
Where the crack starts
Cyclic crosswind bending puts the highest stress at the bottom weld and at any change of section. A welded pipe leaves a continuous toe running the length; a seamless pipe removes it. The checks tabulated on the 347H pipe page, flattening, flaring, visual and chemical, say nothing about a million cycles, and neither does a hydrostatic test. If the gas is hot, the 310S pipe rated for continuous service near 1150 degrees C is the right metallurgy, but the weld detail and the restraint decide the fatigue life.
Materials for the shell and the liner
For a cold or moderate stack the 304L pipe programme runs outside diameters from 3 to 2,200 mm, and the large-diameter range goes from 6 to 2,500 mm at walls up to 150 mm. Where the gas is hot, the 316L coil and the 310S pipe take over. For a low-cost prefabricated flue, the AZ150 aluminium-zinc coil is rated above 300 degrees C and named on chimney tube duty; its 55 per cent aluminium-zinc coat is less dense than zinc. The AZ120 sheet and the Galvalume comparison make the same choice. Carbon alternatives such as API 5L line pipe have the same modulus and a heavier wall, which as shown does not move the critical speed.
The four numbers on the enquiry
Give the outside diameter, the wall, the free height and every intermediate restraint, because those four set the frequency. Give the base weld detail and say whether strakes, guys or a damper are in scope. Ask for the coated plate for the strakes on the same line. A certificate will list yield strength and a hydrostatic test; neither tells the buyer what the stack does at five metres per second, which is the number that decides whether it is still standing in five years.
Next:No More!