Process Engineering Department is the main driver of any EPC Engineering Group. Because they provide information related to the actual process of what is going to happen. Other downstream departments use the information provided by the Process department in their design. The Process Department of any EPC organization normally generates the below-listed deliverables:
Process Engineering Deliverables
Process Engineering deliverables can be grouped into the following two classes:
Process engineering Documents, and
Process Engineering Drawings
Process Engineering Documents
The following deliverables come under the process engineering documents
Start-Prof is a part of PASS software suite for piping stress analysis, hydraulics analysis, boiler & pressure vessel, heat exchanger, column, tank design & stress analysis is available worldwide since 2018.
Article consists of 2 parts:
Part 1. Unrestrained, Totally Restrained and Partially Restrained Pipes. Bourdon Effect
Part 2. Restrained and Unrestrained Zones in the Buried Pipelines. Interpretation of Strength Criteria in ASME B31.4 and B31.8
ASME B31.4 and B31.8 codes divide pipes into restrained and unrestrained. Which part of pipe is restrained and which is not? Many engineers have a misconception about this. We will explain the difference and suggest new universal strength criteria, which cover both restrained and unrestrained pipes.
Before we begin, let’s say that actually, there are three types of pipe behavior instead of two described in ASME B31.4 and B31.8 codes:
Unrestrained pipe expansion from the pressure load consists of two parts. The first part is the expansion due to the pressure load on the end cap. The second part is pipe shortening due to Hook’s law.
Pipe expansion from the pressure load on the end cap is:
L – Pipe Length
E – Modulus of Elasticity
Pipe cross-section area is
D – Pipe Outer Diameter
t – Pipe Wall Thickness
N – Axial Force in the Pipe
Axial force N is equal to the force acting on cap
P – Internal Pressure
Pipe expansion will be
Sh – Hoop Stress in the Pipe
According to Hooke’s law the axial deformation of the pipe under axial stress is:
v – Poisson’s Ratio
Pipe Shortening Under Internal Pressure
Pipe shortening due to internal pressure:
Total pipe expansion from pressure load is
If we add thermal expansion the equation will be:
– Temperature Difference between Installation and Operation temperature
– Coefficient of thermal expansion
Longitudinal stress caused by internal pressure is
If the left end is connected to pressure vessel nozzle or rotary equipment, then axial force in the equipment nozzle will be N as calculated above. But when equipment manufacturers calculate allowable loads, they assume that nozzle has end cap and vessel is under pressure. This means that axial stress caused by pressure is already included into allowable loads and should not be considered twice.
This means that we must exclude the pressure thrust load from axial force to calculate the support load that can be compared to allowable load on nozzle. To do this we must assume that pipe has two caps on the both ends. In this case the support load R will be equal to internal force N minus thrust force on the end cap, i.e. zero
A strength criterion for unrestrained pipe is:
Sallow ‑ Allowable stress.
If we add here bending stress and axial stress from loads other than pressure, we get
If we want to add torsion stress, we should calculate equivalent stress:
Allowable stress value depends on the code. Usually it is Sh or 0.75Sy for sustained loads, kSh or 0.9Sy for occasional loads, 0.9Sy…Sy for test state. Occasional load factor k=1.15…1.8 depends on selected code. Sy is yield stress, Sh – code allowable stress at operating temperature.
Thermal expansion has no effect on unrestrained piping systems, i.e. this equation usually used for sustained and occasional stress check in piping systems from pressure, weight and other force-based loads.
The code equations were created for manual calculation. But now most of pipe stress analysis software can consider Bourdon effect. This means that code equations should be modified to match the current level of technology.
If axial force N is calculated using software that considers Bourdon effect, then we should subtract PD/4t value from axial force otherwise it will be included twice:
The criteria for software analysis where M and N calculated with Bourdon effect should be just:
This has been already done in ASME B31.3 for Process Piping, GOST 32388 for Process Piping, GOST 55596 for District Heating Networks, SNiP 2.05.06-85 for Gas and Oil Pipelines, but still not fixed in all other ASME B31.X and EN 13480 codes.
Totally Restrained Pipe
For a restrained pipe with two anchors on both ends, thermal expansion should be zero
The axial force required to compress the pipe back to its original length can be calculated from this equation:
Therefore support load should be:
After substitution the thermal expansion equation we got final support load for restrained pipe:
The value of axial force can be obtained from the equilibrium conditions near the anchor. Axial force is equal to reaction in anchor minus the pressure thrust force that is received by anchor and doesn’t acting on the pipe:
Final equation for axial force in restrained pipe is
Axial stress in the restrained pipe will be
A strength criterion for totally restrained pipe is:
If we add here bending stress M/Z and axial stress N/A from loads other than pressure, we get
If we want also consider torsion and hoop stress, we should use the equivalent stress equations like described for unrestraint pipes.
If axial force is calculated using software that considers Bourdon effect, then we should subtract pressure axial stress:
The criteria for software where M and N calculated with Bourdon effect and thermal expansion should be:
A criterion is the same as for unrestrained pipes, but allowable stress is usually 0.8Sy…1.0Sy to prevent the Yielding through all pipe length.
The maximum temperature difference for fully restrained pipe, ignoring longitudinal buckling effect, can be found by equation:
If pressure is zero, this value is about 80…110 C for steel pipes.
Partially Restrained Pipe
If we add flexible spring instead of rigid anchor on the right end of the pipe, we will get the third pipe condition – partially restrained.
We will pass the derivation of equations process and just show the final equations in table below.
The strength criteria for partially restrained pipes should be
From sustained primary loads:
From occasional primary loads
From both primary and secondary loads acting simultaneously
Primary Loads – are force driven not self-limiting loads like weight, pressure, relief valve thrust, wind, etc.
Secondary Loads – are displacement driven self-limiting loads like thermal expansion, anchor movements, support or soil settlement, etc.
Unrestrained and fully restrained pipe conditions can be easily calculated manually, but third condition require using of pipe stress analysis software, because spring stiffness k depends on connected pipes.
Bourdon Effect Model in PASS/Start-Prof
Now I will explain how PASS/Start-Prof software considers pressure Bourdon effect in arbitrary piping model. Start-Prof model the pressure loads consist of two parts.
Firstly, Start-Prof adds pressure thrust force on each end of the pipe.
Secondly, Start-Prof adds axial deformation for each pipe. It equal to pipe thermal expansion minus pressure shortening.
The combination of these two loads allows to model correctly any type of piping: unrestrained, restrained, and partially restrained.
Bourdon effect makes a significant contribution to the support loads, displacements, and stresses for
High pressure piping
Plastic piping (PE, PP, PB, PVC)
FRP/GRP/GRE piping
Start-Prof always preforms analysis with Bourdon effect, it is non-disabling function.
Every EPC company must have project-specific pressure vessel nozzle loading tables which are used for comparing allowable nozzle loads for vessels, columns or towers, heat exchangers, Drums, or any similar type of equipment. Normally forces and moments at the nozzle and shell interconnection are provided in a tabular format. These force and moment values are decided based on the following major factors:
Using these tables is quite simple. However, we must keep in mind a few points while using those tables. This article will list those important points for using these tables easily.
1. Before checking the tables, find out the load and moment directional drawing from which we have to correlate the Caesar II axis.
2. Each nozzle, including those designated “spare” but with the exception of man-holes and instrument nozzles shall be designed to withstand the forces and moments specified herein. The indicated loads are to be considered to act at the shell/head-to-nozzle intersection.
3. For nozzles matching with any global direction (other than head nozzles) compare the values mentioned on the tables with global force values in CAESAR II output.
Typical Pressure Vessel
4. For inclined nozzles in the horizontal plane (with respect to any global direction) there are 2 options as listed below.
Compare loads mentioned in the tables above with local element forces in CAESAR II output. In that case, the local X force will be a radial force and compare other directions to get proper forces.
Otherwise rotate the CAESAR II input model to match the nozzle axis with any global Caesar II axis and compare the loads and moments.
5. For Head nozzles (nozzle axis and equipment axis same direction) compare Mx and Mz as per √[{(Mx)2+(Mz)2}] ≤ √[{(ML)2+(MC)2}]
6. In the case of any vessels in the packaged area, these values shall not be applicable and nozzle loading shall be coordinated with the vendor.
7. In the case of any licensor / proprietary item, these values shall not be applicable and nozzle loading shall be confirmed by them.
8. Allowable for self-reinforced nozzle shall be more than as mentioned in the above table. In that case, allowable shall be exercised from the vendor.
If you wish to learn more about Pressure Vessels, their design, fabrication, installation, etc in depth, then the following online courses will surely help you:
This presentation is prepared by Mr. Deepak Sethia who is working in ImageGrafix Software FZCO, the Hexagon CAS Global Network Partner in the Middle East and Egypt. He has extensive experience in using Caesar II, PV Elite, AFT Impulse software, and troubleshooting. The points that will be covered in this article are:
If the vibration frequency is at the natural frequency it is called “acoustic resonance”
PFA PROCESS IN AFT IMPULSE:
PFA Process:
Build the model
– This model represents the suction side of a system with two PD pumps (J8 & J9) in parallel fed by a centrifugal pump (J1)
– We will analyze the upper PD pump (J8)
Select Pulsation Setup from the Analysis menu
▪ This is where the “ring” will be defined
– Specify junction (only one in the system) and pulse start time
– The automatic magnitude of the strike will be twice the steady-state flow
Details of the PD pump are entered on the PD Pump Setup tab
– Used to generate the flow vs. time pump curve for the child pump scenarios
With everything defined we can view the flow pulse that will be used to “ring” the system
– Click Show Pulsation Graphs… in the lower-left corner
The pulse spike and the FFT of the pulse without filtering are shown in the first two tabs
Once the low-pass filter is applied (140 Hz in this case), the pulse spike is now a decaying sine wave
Evaluate Frequencies:
After running the model, go to the Graph Results window
Select the new Frequency tab
Select the pipe location(s) to analyze and click Generate
The spikes show the frequencies that excite the system at that location
Right-clicking on the annotation brings up a menu, allowing the user to Evaluate Excitation Frequency
For PD pumps,
the speed (RPM) will be determined at that frequency for the harmonic multiples
– A red dot appears indicating the frequency is being evaluated
These scenarios have special properties since they are dependent on the configuration of the parent scenario
They are read-only so most of the model cannot be changed
They will be deleted when the parent is changed
They are named with the pump speed and frequency
Analysis:
When the model is run, there will be several blocks of time steps used to determine when the model reaches an equilibrium– This eliminates any artificial fluctuations caused by the PD pump flow, and leaves just the steady-state harmonics of the system
Graphical Report:
The frequency response graph can again be generated– Notice the spike occurs at the frequency of interest.
Output Report:
At the end of the run, the maximum peak-to-peak pressure levels are checked against the API-674 standard
– If the levels are OK there will be a confirmation
– A warning is given if there are locations that exceed the maximum set by the standard. It will list the location where the largest DP occurred.
The elevated flare system consists of a flare header, a knock-out drum, and a flare stack. The waste gas and condensate are collected from the whole plant through the flare header and then the condensate is separated in the knock-out drum finally, the gas is burnt in a stack at a high elevation. As the combustion of gases (toxic) is done at the flare tip at a high elevation, the complete system is called Elevated Flare System.
Purpose of flare system
The primary function of a flare system is to use combustion to convert flammable, toxic, or corrosive vapors to less objectionable compounds like CO2.
Why not cold vent instead of flaring?
Methane is roughly 30 times more potent as a heat-trapping gas than CO2. Hence cold venting of HC gases is not allowed as per pollution control board directives.
Design standards for Flare System
API 521: Pressure-relieving and Depressuring systems
DEP 80.45.10.10: Flare and vent systems (amendments to API 521)
API537: Flare details for refinery and petrochemical service
DEP 80.45.11.12: Flare details (amendments to API 537)
Types of Flares (Fig. 1)
Elevated flares: Commonly used in the oil and gas industry and the most economical.
Enclosed flares: Used in plants where a visible flame is not acceptable. Also used for offshore facilities.
Advantages: Low noise and radiation levels;
Disadvantage: Poor dispersion of gases during flameout condition (flare needs to be tripped on gas detection)
Ground flares: Used for liquid or two-phase relief flaring.
Advantages: Low radiation, low noise;
Disadvantage: accumulation of vapor cloud, high initial cost.
Fig. 1: Types of Flares
Types of Elevated Flares (Fig. 2):
Self-supported stacks
Simplest and most economical design; Stack height up to 100 ft overall height; As the flare height and/or wind loading increases, the diameter and wall thickness required become very large and expensive.
Guy wire-supported stacks
Most economical design in the 100- to 350-ft height range. Normally, sets of 3 wires are anchored 120 degrees apart at various elevations.
Derrick supported stacks
The most feasible design for stack heights above 350 ft. Derrick supports can be fabricated from pipe (most common), angle iron, solid rods, or a combination of these materials. They sometimes are chosen over guy-wire-supported stacks when a limited footprint is desired.
Fig. 2: Types of Elevated Flares
Non-Assisted / Assisted Flares:
Non-assisted flares are the flares that do not use any assist media and are typically used for hydrocarbon or vapor streams that do not cause smoking (i.e. For clean-burning gases like methane, hydrogen, carbon monoxide, ammonia, hydrogen sulfide) or when smoke is not a consideration.
The incomplete combustion of heavy HC gases produces Carbon monoxide, which is the main component to create smoke. For flaring heavy gases, a smokeless operation can be achieved by assisting media such as steam, air, or gas which improves the mixing of flare gas with air.
Steam-assisted flares (Fig. 3) for smokeless operation. Steam increases the momentum of flare gas which enhances fuel-air mixing leading to complete combustion. Also, the water-gas shift reaction converts CO to CO2
CO + H2O ⇌ CO2 + H2
Air-assisted flares (Fig. 3) are used where smokeless burning is required. It is used when steam is not available or where low-pressure air delivery offers a lower cost. (the only fraction of the requirement of air is mixed with flare gas to promote momentum which effectively entrains additional combustion air from the surrounding).
Fig. 3: Steam-Assisted and Air-Assisted Flares
Flare load estimation (Fig. 4)
Fig. 4: Example of Flare Load Estimation
Fire zone: Wetted areas within a 300 m2 (3200 ft2) plot area shall be considered when a system’s relief loads are calculated.
Flare gas flow rate: Tip diameter is decided based on the design flow rate.
Mach number in tip: 0.5 to 0.8 (depends on allowable pressure drop)
Lower gas velocity (Fig. 5): When the gas flow is so low that the local gas velocity is less than flame velocity, air entrains into the flare tip leading to burning back / flashback. At very low gas velocities, the flame can travel back through the mixture (flashback) into piping and KOD.
Fig. 5: Effect of Lower and Higher Gas Velocity
Higher gas velocity (Fig. 5): When the gas flow is higher than the design capacity, then the local gas velocity becomes higher than the flame velocity leading to detached flame or flameout (higher velocity leads to turbulence, which in turn reduces HC component concentration below LFL)
Flame velocity: The burning velocity or flame speed is the velocity at which a flame front moves through the unburnt gas/air mixture. This flame speed varies with the air/gas mixture ratio and the chemical makeup of the gas.
Purge gas requirement: To avoid air ingress down the flare stack purge gas is injected in the flare header. The injection rate should be controlled by a fixed orifice, rotameter, or other devices that ensure the supply remains constant and is not subject to instrument malfunction or maladjustment.
Purge reduction seals (Fig. 6): To reduce the purge rate purge reduction seals are used.
Liquid seal
The liquid seal drum shall be designed as a pressure vessel with a design pressure of at least 7 bar (100 psig) to maintain containment against internal deflagration.
Where there is a risk of an obstruction in the flare due to process flows creating an ice plug with the liquid seal, alternate sealing fluids such as a glycol/water mixture (60% EG & H2O freezing point –45 deg C) or other means to prevent freezing SHALL [PS] be implemented.
For LNG facilities, liquid seal drums shall not be used, since in the event of a cold release this may form an obstruction in the flare relief system.
Allowable Liquid droplet size (to avoid burning rain): Burning rain occurs when the rate of burning (depends on the type of flare) of liquid droplets is lower than the rate of settling of droplets (depends on droplet size).
Fig. 6: Purge Reduction Seals
Drift distances of burning liquid droplets from an inadequately designed flare system can be considerably greater than 200 ft (60 m).
If the liquid is not drained from flare gas, at a gas velocity of 3-4 m/s – liquid droplets of 1000 microns can be entrained which can cause burning rain in the flare.
Liquid droplet size allowed without burning rain
Unassisted flares: <600 micrometer
Steam or air-assisted: <600 micrometers (less than 1% mass)
High pressure (if operated at least 200 kPag): <1000 micrometer (less than 1% mass)
5 barg (50 psig) when a liquid seal drum is located between the KO drum and a flare stack.
7 barg (100 psig), if there is no liquid seal drum in the system.
For a multi-process unit facility (e.g., refinery) based flare KO drum where it may not be immediately clear which unit is sending liquid to the flare, liquid space on top of LA (HH) SHALL [PS] be designed to contain the maximum emergency liquid relief rate from the largest single contingency for a period of at least 15 minutes for the unit KO drum and at least 20 minutes for the flare KO drum, without taking credit for pump out capacity.
Flare Height
The height of the flare (Fig. 7) is established based on allowable thermal radiation levels. Flare height depends on the available plot and the distance of nearby equipment from the flare stack.
More plot area: Low flare stack height
Less plot area: Higher flare stack height
Fig. 7: Flare Height vs Available Plot Area
Thermal radiation:
The effect of thermal radiation on a person at grade or at an elevated platform shall be checked by radiation calculation.
Thermal Radiation affects human skin (skin burn).
Exposure Times Necessary to Reach the Pain Threshold
31 kW/m2 – Up to 20 s
15 kW/m2 – Up to 1 hour
58 kW/m2 – Continuous
If personnel exposure to radiant heat exceeds the guidelines provided, then shielding should be considered.
Depending on the location the thermal radiation limit is provided in Fig. 8
The solar radiation need not be added to calculated thermal radiation values (0.79 to 1.04 kW/m2) from the Flare.
A wind velocity of 10 m/s (22 mph) at the elevation of the flare tip, blowing towards the receiver, is a typical assumption for flame tilt assessment.
When two flares are located in close vicinity, combined radiation effects shall be calculated.
Fig. 8: Thermal Radiation Limit
Dispersion Analysis
To ensure safe operation during periods when the flame might have extinguished, the concentration of hazardous components should be determined using dispersion analyses, assuming the flare is functioning as a vent only.
Level of Concern
Hydrogen Sulphide (Concentration, Time)
Sulphur Dioxide (Concentration, Time)
8 Hour TWA (Threshold Limit Value)
5 ppm, 8 hours
2 ppm, 8 hours
15 Minute STEL (Short-Term Exposure Limit)
10 ppm, 15 minutes
5 ppm, 15 minutes
Short-term exposure limits (STELs) are set to help prevent effects, such as eye irritation, which may occur following exposure for a few minutes
Smokeless requirement
Local rules and regulations shall be followed. Typically flare combustion quality shall meet Ringelmann Index 1 criteria (Fig. 9).
Smokeless flowrate shall be the normal flow that is expected in day-to-day operations. Do not specify the design capacity for smokeless operation.
A scale used to define levels of white, gray, and black i.e. intensity of smoke
Ringelmann No. 0 is clear smoke
Ringelmann No. 5 is 100 percent black.
Ringelmann No. 1 is equivalent to 20 percent black
Fig. 9: Ringleman chart
Other requirements
Noise: For normal flow rate (including starting-up and shutting-down): 85 dB(A) at the sterile radius. For emergency conditions: 115 dB(A) at sterile radius
Combustion efficiency: greater than 98%
The number of pilots (Fig. 10): The number of pilots required is a function of the flare burner diameter. For very small flares, a single pilot will reliably light the flare gas. However, it should be noted that if only a single pilot is used, a single pilot failure would represent a complete failure of the ignition system. Recommended installing at least 2 pilots for tip size of up to 8″ to increase reliability. As the flare burner diameter increases, the number of pilots required to reliably light the flare, regardless of wind direction, increases.
Fig. 10: Number of Pilots
Flare gas recovery system Safety considerations
Path to flare: PRVs, depressuring systems, etc., shall always have flow paths to the flare available at all times.
Reverse flow: Because flare gas recovery systems usually involve compressors that take their suction directly from the flare header, the potential for the reverse flow of air from the flare into the compressors at low flare gas loads shall be considered.
Monitor oxygen content in the flare header and provide recovery system trips. Provide a low-pressure trip on the recovery system suction to avoid air ingress. Liquid seal drum (not practical in AP flare systems)
All Piping stress engineers who use Intergraph’s Caesar II software must be aware that there is an inbuilt API 610 module for API centrifugal pump nozzle load checking. By the use of this module, you can directly check if the pump nozzle loads are within acceptable limits provided by the API 610 code. Searching the code for allowable load or asking the vendor for the nozzle limiting force and moments are not required.
The method of using the API 610 module is fairly simple. But before you start using the module you have to perform static analysis following conventional methods. In this video tutorial, the API 610 module is explained clearly.
– Temperature Difference between Installation and Operation temperature
– Coefficient of thermal expansion
