December 2016

Process Engineering and Optimization

Optimize steam usage in refinery flares

In this study, methods to optimize the consumption of dispersion steam in flares were examined at a large refinery that burns a considerable quantity of gas. The desired optimization was achieved by monitoring hydrocarbon emissions using an infrared camera (thermography).

In this study, methods to optimize the consumption of dispersion steam in flares were examined at a large refinery that burns a considerable quantity of gas. The desired optimization was achieved by monitoring hydrocarbon emissions using an infrared camera (thermography). Utilizing this methodology, it was determined that an excessive quantity of steam was used in the flares under study; plumes of hydrocarbons that had not been converted in the flares could be seen exiting the stacks. The steam flowrate was reduced until the optimal level was reached, preventing both the incomplete combustion of hydrocarbons and the formation of soot. The results indicated that an infrared camera could be helpful for optimizing steam consumption, which can yield significant savings for refineries.

In 2012, the US Environmental Protection Agency (EPA) recognized that open and steam-assisted flares do not have the 98% conversion efficiency claimed by their manufacturers, and that regulations fail to ensure this performance level. Teaming with industry leaders and manufacturers, the EPA has undertaken studies that should result in new rules for flares. These rules include new operating and monitoring guidelines involving a number of intrusive and non-intrusive methods, such as the use of infrared cameras, to determine process parameters to improve flare efficiency measurements.

Proving percentage efficiency in this area is difficult, and a good deal of uncertainty about the values exists. The difficulty of obtaining the minimum 98% conversion is related largely to the way flares are operated globally. Originally, the main functions of flares included:

  • Protecting equipment from catastrophic faults (pressure relief)
  • Shielding employees/neighboring communities from exposure to toxic gases
  • Promoting general refinery safety: the flare system should be large enough to cope with worst-case conditions, when gases must be released simultaneously from multiple units.

Flare systems, which were originally designed solely for emergencies, are now habitually used in refineries for continuous discharges, and have even become an alternative mechanism for controlling the pressure in process units. In this scenario, it is easily understandable why the daily quantity of flared gas at some refineries is so high. This misuse is arguably the main problem facing flare systems.

Meanwhile, one of the issues that influences the conversion performance of flares is their operation with a gas stream that is far below their installed capacity. However, as observed above, flares should be sized to cover the worst-case scenario—the simultaneous release of gases from multiple units. In view of this fact, some facilities around the world operate one flare exclusively for processes and another exclusively for emergencies.

Refineries should be operated to keep flare streams to a minimum, as that equates to reduced product (and, therefore, economic) losses, and decreased environmental impact. Nonetheless, if gases must be discharged, then they should be sent to the flare.

Another challenge worth mentioning is that workers operating flare systems are trained to add enough steam to prevent visible emissions. This has resulted in a culture operating under the understanding that the more steam the flare system uses, the better. In fact, the optimal point of operation of a flare system in terms of its combustion efficiency is not when there is an excess of steam, but just before the flare begins to form soot. Therefore, it should be regarded as good practice to operate flares with a visible flame (but without soot), as this will bring about a reduction in emissions of unburned hydrocarbons to the atmosphere, improve the air quality around the refinery and enhance public health. An excess of steam actually diminishes the efficiency of hydrocarbon removal.



This non-destructive, non-intrusive technique produces thermal images (thermograms) of a given body or surface. The technique is based on the thermal radiation emitted by any material with a temperature higher than absolute zero, and it attributes wave properties (wavelength and frequency) to this radiation.

Infrared rays are used to measure temperatures, or observe differential temperature distribution patterns, to provide information about the operating conditions of a component, equipment or process. Thermal imaging cameras are used for this purpose, producing an artificial view of the infrared spectrum, which is invisible to the human eye.

Thermal imaging is widely used in preventive maintenance. It helps eliminate many production problems and prevents electrical and mechanical faults, as well as material fatigue.

Thermal imagers, infrared cameras

Thermal imaging cameras transform infrared radiation into visible images. They detect the energy emitted by objects, modify the frequency of the energy detected, and produce corresponding images in the visible range of the electromagnetic spectrum. To produce a thermal image, a temperature difference must exist. If the temperature of a surface is constant, then no image is formed.

A proprietary portable infrared camera has applications in different industries and uses the principle of infrared thermography to detect leaks of specific gases. These leaks are imperceptible to the naked eye, but seen through the infrared camera, they appear as plumes of gas deriving from a clearly identifiable point of origin. Therefore, this camera is invaluable for leak detection purposes.

The portable infrared camera uses optical systems and sensors that are adjusted to detect narrow ranges of the infrared spectrum. With thermal sensitivity of up to 35 mK, the camera is capable of detecting subtle temperature changes in the image, meaning that the camera can “see” the energy given off by particular gases. The images are processed and improved to clearly show the presence of gases against stationary backgrounds. The gases that can be detected by the camera appear on the screen as white or black smoke, depending on the contrast (which can be automatic or adjusted manually). The images produced by the portable infrared camera can be saved as short films or still pictures.


The methodology adopted for the tests consisted of varying the flowrate of the steam in the flares to obtain optimal flow, when there was enough steam to prevent the formation of soot but no emissions of unburned hydrocarbons. The optimal steam flowrate identified for each flare was specific to the operating conditions being used when the tests were conducted, such as stream flowrate, composition of the flare stream and height of the flare seal.

The camera was first used to record images of flares operating with the steam valve apertures set by the flare system operators. The decisions about which apertures to use were made in the field after observing the unburned hydrocarbons and the presence of soot. In other words, each flare was tested under normal working conditions (normal aperture of the steam valve), and then under a series of other conditions with sequentially smaller apertures, until soot began to be observed in the flare. The aim of this exercise was to identify, under test conditions, the optimal aperture for the steam valve to prevent the formation of soot and the presence of unburned hydrocarbons.

To correctly assess the behavior of the flares under each valve aperture size, some apertures were repeated. This was necessary to get clearer images, and to differentiate more clearly between valve apertures, since the process is dynamic and many factors can interfere in the performance, such as wind and solar irradiance, which affect the formation and visualization of plumes of unburned hydrocarbons.

FIG. 1 shows Flare 1 operating with the valve that controls the steam flow 75% open. FIG. 2 shows an infrared image of the same flare produced by the infrared camera. The plume circled in red is of hydrocarbons that were not destroyed by the flare, which are invisible to the naked eye (FIG. 1), but can be detected at an infrared wavelength.

Fig. 1. Flare 1 operating with the steam valve 75% open.
Fig. 2. Infrared image of Flare 1 operating with the steam valve 75% open.


Test summary

The consolidated results of the tests are presented in TABLE 1. The tests served to identify the optimal lift of each steam valve, enabling a comparison with the valve normally used in operations. These optimal values were specific to the process conditions at the time the tests were conducted. Based on the observation that excessive quantities of steam were being used, the next stage of the study involved measuring the potential financial and operational savings. It should be noted that using an optimal steam flowrate is important not only for reducing wastage, but also for the environment, as it prevents the emissions of large quantities of volatile organic compounds (VOCs) and greenhouse gases to the atmosphere.

To quantify the amount of steam saved, the plant information system was consulted to retrieve operational data from the tests. However, inconsistencies were found between the valve aperture data and steam flowrates, making it impossible to use this system data. It was decided to track data over periods for which more coherent steam valve aperture data and their respective steam flowrates were available, to derive an equation for each valve and thereby represent as closely as possible their real behavior.

Selection of experimental curve for steam valve

Four sections of data were identified for Flare 1, which enabled experimental equal-percentage curves to be plotted for the steam valve, as shown in FIGS. 3–6.

Fig. 3. Experimental curve A for Flare 1.
Fig. 4. Experimental curve B for Flare 1.
Fig. 5. Experimental curve C for Flare 1.
Fig. 6. Experimental curve D for Flare 1.

By inserting different valve lifts into each equation, the respective steam flowrates were obtained. Generally speaking, the values were coherent (especially curves B, C and D), which tends to validate the results and confirms the reliability of the selected curve.

The first step in selecting the curve that best represented the behavior of the steam valve in Flare 1 was to identify the curves that had the most similar results. Curve A was eliminated because its values were the most dissimilar from the others. The results of curves C and D were very similar, but as the correlation coefficient (R²) of curve C was better, it was the equation selected for use.

Selection of curve for steam valve of Flare 2

For Flare 2, only two sections of valid data (FIGS. 7 and 8) were identified as showing the kind of behavior expected for equal-percentage valves. By inserting the different apertures into each equation, the respective flowrates were obtained. Although only two sections of data were identified, the curves are reliable because both are similar and demonstrate coherent behavior between the valve opening and flowrate for an equal-percentage valve. Curve B was selected as the equation for the steam valve of Flare 2 because its correlation coefficient (R²) was better.

Fig. 7. Experimental curve A for Flare 2.
Fig. 8. Experimental curve B for Flare 2.

Potential savings by reducing steam consumption

The results obtained from this analysis of two flares indicated that more than $2,300/d could be saved, which would amount to approximately $850,000/yr. TABLE 2 shows the final results of the tests.

By using the infrared camera, it was possible to ascertain that the studied flares contained a large quantity of unburned hydrocarbons, indicating the excessive use of steam. The tests also indicated what the optimal flowrate of steam to these flares would be under operating conditions when the study was completed, resulting in the absence of black smoke and unburned hydrocarbons in the flare. The research determined that it is common practice to use excessive quantities of steam in flares, and this has been corroborated by studies undertaken at refineries in other countries. For a flare to operate optimally, the wide range of conditions under which it is operated must be constantly monitored. HP


GasFindIR™ is a portable infrared camera produced by FLIR Systems.


  1. Venkoparao, V. G., et. al., “Flare monitoring for petroleum refineries,” 4th IEEE Conference on Industrial Electronics and Applications, Xian, China, 2009.
  2. US Environmental Protection Agency (EPA), “Petroleum refinery sector risk and technology review and new source performance standards,” Proposed rule, Federal Register, Vol. 79, No. 125, 2014.
  3. Torres, V. M., S. Herndon and D. T. Allen, “Industrial flare performance at low flow conditions: Steam and air-assisted flares,” Industrial & Engineering Chemistry Research, 2012.
  4. US Environmental Protection Agency (EPA) website:

The Authors

Related Articles

From the Archive