FRP crack propagation rates / activity

[YungPlantEng]

Hello,

We have a FRP tank in atmospheric service for an acid. Surface cracks were recognized at a nozzle mated to a stainless flange a couple days upon a prior torque-in-service due to a leaking gasket. Based upon the lack of an expansion joint and appropriate pipe supports (as well as previous visual inspections indicating small cracks) it is believed that excess compression from the torque activity on the upper section of the nozzle increased crack diameter and depth.

From approximately a week of marking we haven't noticed any indication of crack propagation. I was wondering what literature is out there for crack propagation on fiber-glass reinforced systems (non-carbon fiber). My admittedly primitive understanding of FRP as a material and how crack propagation occurs in stainless vessels leads me to believe the crack is inactive until further stress placed upon it via expansion/compression loading from temperature fluctuations.

We're currently coordinating a complete nozzle replacement but I wanted to do a little more research to see if delaying the repair to a later date would be viable with remediation of the pipe support / expansion joint install.

Thank you!

 

[Dhurjati Sen]

Have you taken the tank out of service by now?
Hope you have spare tanks.

DHURJATI SEN
Kolkata, India

 

[Louisiana Inspector]

Look at the first obvious reasons for failure.

Was the Flange over-torqued in the Past initiating the cracks or incorrect torque pattern?

Assuming the FRP flange is flat face is the stainless flange flat face also? If raised face this could be your issue.

What type of gaskets are being used? Full face?

If all things listed above are scratched off I would certainly perform stress analysis, Vibration analysis, etc., and have an engineer involved in the design of a support or Expansion joint

 

[EdClymer]

YungPlantEng,

I have looked carefully at your photograph.

Under high magnification, a GRP(FRP) (probably polyester/glass)laminate can be seen below the very thick layer of flow-coat.

In a physical battle between GRP and steel - steel will always win - (much) stronger, harder, tougher.

Polyester resin (cured) is brittle - which is why it is combined with glass fibres for all practical uses.

The flow-coat - a thixotropic, pigmented layer is brushed or rolled on to give a final (thin) finish (rather like paint!)

In your case the flow-coat is grossly thick, with entrapped air below.

In flange bolting the gasket does the work, a suitable grade of ??? is chosen with a modest Modulus to allow half thickness
compression at modest bolt torque.

A GRP (FRP) must be mated to full face flat surface flange via a suitable gasket. No raised face. Also the gasket seals over a
narrow band only - located just within the PCD.

A GRP flange (thermoplastic lined or not) always has to be used with a backing ring/spreader plate - to spread the compressive forces
induced by the bolting. You appear to be using a standard washer and full nut only. If I assume (why not) M16 bolt, torqued to seal
at say 250 Nm (newton - meter)- This can induce a force of 57,459 N (newton) and a (standard) washer stress of 119 MPa (N/mm[sup]2[/sup]), but a
form F or G washer stress of 33 MPa (N/mm[sup]2[/sup]). With a backing ring much less.

Under high compressive loading; washers tend to embed into the laminate (over time) thereby reducing the clamping force and leaks result.
Suggest backing rings (and longer bolts), or at very least use over size washers which are also thicker than standard (in UK we use Form F and G).
Not sure if it camera angle - but space between branch and bolt seems 'small'.

Probe the cracks with a suitable tool to check depth and extent of crack. Monitor.

Sorry about use of SI units - but I have forgotten how to use the Imperial (English) units (an age thing!): regret to say can now only think
in metric units.

Trust this is helpful.


Regards

Ed


Ed Clymer
Resinfab & Associates
England

 

[YungPlantEng]

Appreciate the response! We’ve since appended our standards to specify flat face flanges for mating to FRP systems. I’ll have to check with our bolting specialist on your statements related to the washers - unfortunately I’m not as much an expert in appropriate bolting components.

The cracks extended through the paint into the fiberglass so we took the system down to grind them out and apply new rather than wait for the cracking to lead to a failure and nozzle replacement.

Going to look at the paint statement you made and see if NACE has any applicable guidelines to support the next tank construction.

Once again appreciate the steady eye and comments!

 

[EdClymer]

YoungPlantEngh

The 'paint' is not paint.It is a thixotropic coat comprising of laminating resin plus a viscosity
enhancing gell (aka pre gell).

Purpose is to seal the laminate from external environmental attack and improve visual appearance.

Backing rings of metal (suitably protected) are always used, minimum 6 thick, on GRP/FRP flanges.
(aka spreader rings/plates: as they spread the bolt load and reduce bolt induced compressive stress).

Standard washers to GRP/FRP, given reasonable bolt torque will always embed over time. Avoid.

Regards

Ed



Ed Clymer
Resinfab & Associates
England