Static Stress Analysis of an Underground GRE Firewater System

ON THIS PAGE

Seventeen underground GRE firewater subsystems at a degassing station were analysed in CAESAR II for three load cases, hydrotest, operational and sustained, to ISO 14692. It may be concluded that, upon implementation of the recommendations, the static pipe stresses are acceptable and the flange loads are moderate.

Underground GRE Firewater Piping at a Degassing Station

A main degassing station required an upgrade of its firewater system. The underground piping is glass-reinforced epoxy (GRE) and connects to fire hydrants and monitors, steel piping and existing GRE firewater lines. The system consists of seventeen subsystems.

The design pressure is 20 barg, the maximum design temperature is 52.0 °C and the installation temperature is 21.0 °C. The static stress analysis assessed the pipe stresses and flange loads of each subsystem and defined the adjustments to the standard arrangement needed to bring stresses within allowable limits.

GRE Material Data, Stress Envelope and Flange Allowance

The pipe is GRE type EST 20 with a 55 degree winding angle. The axial elastic modulus is 12,000 MPa and the hoop elastic modulus is 20,500 MPa. The shear modulus is 11,500 MPa. The Poisson ratio is 0.38 for hoop load (hoop/axial) and 0.65 for axial load (axial/hoop). The thermal expansion coefficient is 20 x 10⁻⁶ mm/mm/°C and the specific gravity is 1.85. These values and the allowable stresses are based on material data provided by the pipe manufacturer.

Nominal diameter

Structural wall thickness (mm)

ND 100

2.4

ND 150

2.4

ND 200

3.3

ND 250

4.1

The stress assessment used the allowable load case dependent stress envelope. The assessment plots show the operational, sustained and occasional envelopes and the five highest stress conditions for each load case, as axial stress against hoop stress in MPa. The hydrotest is labelled pressure test on the plots.

Because the EST 20 GRE flanges are used at design conditions, no allowance is available for external loads. According to a memo from the flange manufacturer, the qualified pressure is two times the rated pressure. Combined flange loads, meaning internal pressure and external loading, are checked against the rated pressure corrected by the applicable load dependent part factor. Where GRE piping connects to steel piping, the GRE flange has to be protected from loads applied by the steel piping, which may be done by applying restraints in the steel piping.

CAESAR II Models and Load Cases for Seventeen GRE Subsystems

Each subsystem was modelled in CAESAR II. Fifteen subsystems are named after the lowest pipe number they contain; the other two, S29 and S30, are the systems on isometrics sheets 29 and 30. Four of the fifteen contain more than one line: P3 (lines P3 and P7), P4 (lines P4, P5 and P6), P10 (lines P10, P11, P12, P13 and P17) and P20 (lines P20 and P21).

CAESAR II model plot of a GRE pipe run coloured by wall thickness in mm. The pipe shown is the 2.4 mm section.
CAESAR II model plot of a GRE pipe run coloured by wall thickness in mm. The pipe shown is the 2.4 mm section.

Three load cases were analysed. Weight is that of the filled pipe, with water at 1,000 kg/m³ as the content.

Load case

Pressure

Temperature

Loads

Hydrotest

30 barg

21 °C

Weight plus test pressure

Operational

20 barg

52 °C

Weight, maximum design temperature and design pressure

Sustained

20 barg

21 °C

Weight and design pressure

GRE Hydrant Flange Restraint and Boundary Conditions

At the hydrant connections the piping is anchored at grade level, where a concrete floor was applied and the piping is assumed to be encased and fixed to it. To limit the moment loads transferred to the GRE when the hydrants are manipulated, the top flange of the valve connecting the steel hydrant piping to the GRE piping is restrained in all directions except the vertical. A rigid connection from the steel hydrant piping to the concrete floor is required to prevent operating loads on the hydrants from overloading the GRE flanges. It is implemented by a clamp with brackets that connects the flange to the anchor foundation.

Piping that is not visible on the isometrics but is connected to the existing GRE firewater lines was included in the models for overlap purposes, as a boundary condition. These overlaps are based on the general arrangement drawing. Steel piping and some existing GRE piping outside the scope of the assessment were modelled from a general overview drawing as boundary to the GRE piping. In some subsystems, axial stops were introduced to generate a boundary condition, and their positions were chosen to prevent high loads and moments on the GRE flanges.

Where the connecting piping was not known, two situations were analysed. In the first, an anchor was introduced at the connection and the restraint loads were established. In the second, the connection was unrestrained and the movements of the free end were established. The stress assessment and the measures for improvement are based on the worst case, and the measures give acceptable stresses in both situations. In the actual installation it must be assured that the displacements at the connection with the existing piping do not exceed the values presented on the stress isometrics.

Stresses and Flange Loads in GRE Subsystems to ISO 14692

The pipe stresses were assessed in accordance with ISO 14692, using the material data provided by the manufacturer. The adjustments to the standard arrangement are indicated on the stress isometrics. Where the wall thickness was increased, the effective thickness TE is shown, and the model plots show the applied wall thicknesses by colour for each subsystem. Anchors that have to be applied, other than at the hydrant connections, are also shown. The locations of the axial stops are given in the appendices of the report.

Highest operational and sustained stresses and pressure test stresses for one GRE subsystem, plotted as axial against hoop stress in MPa with the operational, sustained and occasional stress envelopes.
Highest operational and sustained stresses and pressure test stresses for one GRE subsystem, plotted as axial against hoop stress in MPa with the operational, sustained and occasional stress envelopes.

It may be concluded from the computer results that, upon implementation of the recommendations, the static pipe stresses are acceptable and the flange loads are moderate. The report attributes the moderate flange loads to the measures taken.

Cover Depth of Buried GRE Piping

Cover depths of at least 747 mm were applied in all models. The cover depth is normally indicated on the isometrics. Where it was not, it was estimated from nearby hydrant locations with specified burial depths or from the general arrangement drawings. In most cases the cover depth turned out to be approximately 800 mm, measured from the top of the pipe, and in some cases it was less than 800 mm.

A cover depth of at least 800 mm is advised to assure buried stability and prevent pipe upheaval. Sufficient cover depth is also required to give the pipe the support it needs, meaning sufficient friction and side support, which limits the pipe stresses. The cover depth must be at least 800 mm in all subsystems, and if the actual cover depth differs from the values indicated on the isometrics, the stresses have to be rechecked.

Frequently Asked Questions

How is a GRE flange protected from loads applied by connected steel piping?

Restraints may be applied in the steel piping. At the hydrants, a rigid connection from the steel hydrant piping to the concrete floor is required, implemented by a clamp with brackets that connects the flange to the anchor foundation. The top flange of the valve between the steel hydrant piping and the GRE piping is restrained in all directions except the vertical.

At least 800 mm in all subsystems. This depth assures buried stability, prevents pipe upheaval and provides the friction and side support that limit the pipe stresses. If the actual cover depth differs from the values on the isometrics, the stresses have to be rechecked.

Two situations are analysed. One model has an anchor at the connection, which gives the restraint loads. The other leaves the connection unrestrained, which gives the movements of the free end. The stress assessment and the measures for improvement are based on the worst case, and the displacements at the connection must not exceed the values on the stress isometrics.