Hydrocarbon Processing Copying and distributing are prohibited without permission of the publisher
Email a friend
  • Please enter a maximum of 5 recipients. Use ; to separate more than one email address.

Improve process control for natural gas heat exchangers

01.01.2012  |  Lai, H-M.,  Jacobs Canada Inc., Calgary, Alberta, Canada

Simulation model shows how to optimize plant controllability and safety

Keywords: [natural gas] [preheating] [heat exchanger] [process control] [control valve] [hydrate formation] [heat transfer] [simulation models] [safety]

Dynamic simulation is becoming an important tool for engineering design and plant operation.1,2 In this case history, a first-principle dynamic model of a natural gas/steam heat exchanger system is built using a commercially available dynamic simulator. Four scenarios for operability and safety are investigated to demonstrate how a process and associated control system will respond to various disturbances as a function of time.

Case history. Preheating of natural gas (NG) is frequently used to prevent hydrate formation due to the Joule-Thompson effect of the NG let-down stations. The typical NG heater system consists of a shell-tube heat exchanger, a condensate receiver and a steam trap, as shown in Fig. 1. The steam control arrangement is also shown in Fig. 1. This system includes:
• One temperature control valve on the steam inlet line
• One level control valve on the condensate outlet line
• One pressure control valve on the vapor outlet line.

  Fig. 1. Simplified process scheme for the NG steam heater.

As a part of the plant design, a steady-state simulation of the system is done to check the heat-and-material balances and equipment sizing. Table 1 lists the process conditions and major equipment sizing data.


Scenario 1: Process upsets.

In this scenario, the impacts of both inlet NG temperature changes from 0°C to 10°C and NG demand changes from 100,000 kg/h to 140,000 kg/h have been analyzed. The simulator logic unit operation—the transfer function block—is used to simulate sine wave changes of NG demand and inlet temperature.

Fig. 2 shows the process response to an inlet temperature change of the NG from 0°C to 10°C. As shown in Fig. 2, when the NG inlet temperature rises from 0°C to 10°C as a result of falling heat load, the steam pressure in the heat exchanger will drop about 200 kPa. Fig. 3 shows the process response to a change in NG flow from 100,000 kg/h to 140,000 kg/h. Note: The steam pressure also drops over 250 kPa, while the NG demand declines from 140,000 kg/h to 100,000 kg/h. These dynamic simulation results confirm that steam pressure in the condensate receiver cannot be maintained at stable ranges during process upsets. If a steam trap is used, then the steam control scheme will lead to reduced condensate flow from the steam heater system, and it will form the so-called “stall behavior.”

  Fig. 2. Process responses to NG inlet temperature changes:
  0°C to 10°C.

  Fig. 3.  Process responses to NG demand changes.

Scenario 2: Stall behavior.

This scenario discusses condensate removal from the heat exchanger. As mentioned before, the temperature control valve on the steam line maintains the NG outlet temperature by opening or closing to adjust the steam flowrate, thereby varying the steam space pressure. When the steam pressure in the heat exchanger is equal to, or less than, the total backpressure imposed on the steam trap, then the reduction or cessation of condensate flow from the heat exchanger occurs. The condensate will back up in the drain line and will flood back into the exchanger. This condition can damage the control valve and may cause corrosion of the exchanger. This symptom is called the “stall behavior.”

Based on current heat exchanger sizing data, the dynamic heat model for this steam heater system was built using dynamics and spreadsheet tools. These conditions were assumed for the model:
• NG gas inlet temperature rising to 10°C
• Low-pressure (LP) steam pressure of 442 kPa and steam-trap backpressure of 338 kPa
• NG consumption of 132,000 kg/h.

Fig. 4 illustrates the simulated stall behavior. Due to the 10% over design margin, the heat exchanger has more heating area than required. So, the operating steam pressure will be much lower than needed.

  Fig. 4. Simulated results of stall behavior for the steam heater.

When the condensate is waterlogged in the heat exchanger, the surface area available to condense steam is reduced. The heat flow drops, and NG outgoing temperature begins to fall. While the temperature sensor detects this change, the controller will open the steam control valve. This raises the pressure in the steam space to above the trap-back pressure and causes condensate to pass through the trap. The condensate level falls, and the NG temperature climbs. When the sensor detects this, the controller closes the control valve. The steam pressure falls, and then flooding begins again. The result is a continual cycling of opening and closing the steam control valve.

The side effects of stall include damaging the control valve and water hammer along with corroding and leaking heat exchangers. These operating conditions will increase maintenance incidents and reduce the service life of the steam heater and associated equipment.

Scenario 3: Alternative control scheme.

There are different ways to prevent stall.3 Normally, we could use an alternative means to remove condensate from the exchangers by installing a pumping trap, instead of using steam traps if the pressure in steam space may be less than the backpressure. We could also size the heat exchangers and steam traps properly to ensure that the pressure in steam space is stable and always higher than the backpressure under all operating conditions. Or we should reduce the backpressure of condensate discharge lines. In reality, this can’t always be done.

For the present NG steam heater, the most cost-effective solution is to use an alternative control scheme—a bypass control. This control approach bypasses a partial NG stream around the exchanger and blends it with a fraction that has passed through, as shown in Fig. 5. The temperature control valve is relocated from the original steam line to the NG bypass line.

  Fig. 5. Alternative control scheme for the NG steam heater.

System dynamic responses to the process upsets over NG demand and inlet temperature are illustrated in Figs. 6 and 7. The results show that the maximum change of steam pressure is much less, lower than 30 kPa. The stall behavior will not happen, as the pressure in the steam space is always greater than steam trap backpressure. Compared with the regular steam control, the simple bypass control greatly improves the operating performance of the steam heater.

  Fig. 6. Process responses to NG inlet temperature changes:
  0°C to 15°C.

  Fig. 7. Process responses to NG demand changes.

Scenario 4: Tube rupture contingency.

Pressure-relief systems are a critical part of any process design. Proper design of these systems is required by regulation and industrial codes. Due to the large operating pressure difference between the exchanger tube and shell sides (flange rating 900 lb at tube side vs. 150 lb at shell steam side), the case of complete tube rupture is a valid case in the steam heat exchanger.

Although the simulator cannot predict the instantaneous pressure wave at the rupture site, it does provide important insights on the dynamic system behavior under the tube-rupture conditions. Normal operating data and pressure safety valve (PSV) sizing results by the conventional method are listed in Tables 1 and 2. These parameters were set to generate the initial values of the dynamic model for the tube-rupture case:
• UA value was set for the steam heat exchanger
• Condensate receiver was set to real sizes to simulate steam/liquid accumulation and liquid level variations
• Normal valve with a customized spreadsheet was used for constant NG rupture flow into the steam condensate system.


In general practice, to protect overpressure of steam system from the high pressure of NG, a check valve should be installed on the upstream steam line, and a PSV shall be provided on the top of the vapor line in the condensate receiver.

The dynamic simulation with two different PSV sizes was verified, Figs. 8 and 9 summarize the results. Some highlights of the dynamic simulated results are discussed here:

• Pressure in the condensate receiver begins to build up immediately following the tube-rupture event. After about 3 seconds, the receiver pressure reaches the set pressure; then the PSV starts to relieve.

• PSV would work fine if the normal PSV of 4M6 sized by a conventional method is installed on the top of shell side in the steam heater. However if this PSV is relocated to the top of the condensate receiver, a 40% overpressure in the receiver would occur, as shown in Fig. 8. The major reason is that, under the upset conditions of the tube rupture, the NG has a strong stripping effect (due to vapor/liquid equilibrium) that carries the steam out of the condensate phase. This causes the PSV peak relief load (30,530 kg/h) from the condensate receiver to be about 23% higher than the tube-rupture flow (24,920 kg/h) estimated by API 521 method.

  Fig. 8. Tube rupture profiles for 4M6 PSV with 6 in. inlet/
  10 in. outlet piping.

• As evaluated in Fig. 9, if installing a PSV on the top of the condensate receiver, a larger sized 4P6 PSV and associated larger inlet/outlet piping should be installed.

  Fig. 9. Tube rupture profiles for 4P6 PSV with 8 in. inlet/
  12 in. outlet piping.

This example shows that, when upset conditions occur, equilibrium conditions in vessels are changing, and the safety system design must be adjusted to account for those changes.

Options. This case study illustrates how critical it is to consider vapor/liquid equilibrium changes and interaction of process with controls in the system design, and how dynamic simulation can improve plant performance, controllability and safety in design and operation. HP


Special thanks to Alan Childs, manager of the process department, for the valuable discussions, review and comments.


1 Dissinger, G. R., “Studying simulation,” Hydrocarbon Engineering, May 2008.
2 James, G. and J. Reeves, “Dynamic Simulation Across Project and Facility Lifecycles,” 6th World Congress of Chemical Engineering, Melbourne, Australia, Sept. 23–27, 2001.
3 www.spiraxsarco.com/Resources, “Practical Methods of Preventing Stall.”

The author 

Hai-Ming Lai is a principal process engineer in Jacobs Canada Inc., Calgary, Alberta, with over 26 years of experience in process research and development, design, and engineering of oil and gas, refining/upgrading, and petrochemical projects. His specialties include simulation studies, conceptual and front-end engineering design. He holds a PhD in chemical engineering from Beijing University of Chemical Technology (BUCT), P.R. of China., and is a registered professional engineer in Alberta, Canada. Prior to joining Jacobs, Dr. Lai worked for Aspen Technology, Calgary, Canada, and Research Institute of Chemical Technology in BUCT, Beijing, P. R. of China. 

Have your say
  • All comments are subject to editorial review.
    All fields are compulsory.

Related articles


Sign-up for the Free Daily HP Enewsletter!

Boxscore Database

A searchable database of project activity in the global hydrocarbon processing industry


Is 2016 the peak for US gasoline demand?




View previous results

Popular Searches

Please read our Term and Conditions and Privacy Policy before using the site. All material subject to strictly enforced copyright laws.
© 2016 Hydrocarbon Processing. © 2016 Gulf Publishing Company.