At most in situ oil sands operations, the bitumen is extracted by injecting steam into the reservoir, causing it to soften to a point where it can be pumped to the surface. Typically, the steam is produced using boilers called once-through steam generators (OTSG). These boilers use a series of tubes that run along the outside of a heat source. As the water passes from one end of the tube to the other, heat is progressively added, transforming water into steam.
Currently, most OTSGs operate at about 75- to 80-per cent steam quality, meaning this percentage of the water is converted into steam. The rest becomes “blowdown,” a mixture of water, salt and other solids. It becomes a challenge to increase steam quality because the more water that is converted into steam, the more scaling that builds up in the boiler tubes, increasing the risk of fouling.
Based on the quality of water being used in the boiler, manufacturers usually set a maximum steam quality limit of 80 per cent to avoid tube scaling and failure. According to Subodh Peramanu, Process and Technology Specialist at Canadian Natural Resources Ltd. (Canadian Natural), operators would prefer to maintain higher steam quality to maximize oil production and reduce blowdown. However, operators have to keep the steam quality below the maximum to accommodate for changes in steam quality caused by fluctuations in temperature, the rate of water coming into the boiler and the amount of contaminants in the boiler feed water itself.
Canadian Natural is testing a technology for use at its Primrose North in situ facility with the potential to reduce the variability of steam quality, enabling OTSGs to run at higher steam quality. The technology called “soft sensors” was developed for OTSGs by Dr. Biao Huang at the University of Alberta (U of A), in partnership with Suncor Energy.
Soft sensors are virtual sensors that use existing measurements to calculate the unknown values of a quantity of interest using advanced algorithms. They are also called inferential, proxy or surrogate sensors
Soft sensors are commonly used in industrial applications where the conditions make hardware sensors unsuitable or too expensive to implement. For example, they are used in the refinery industry to minimize hydrogen sulfide and sulfur dioxide emissions.
The analogy of a car can be used to explain how soft sensors work.
If the speedometer in a car stopped working, other data like the engine’s revolutions per minute (RPM) and the gear ratio could be used to determine how fast the vehicle is moving. Soft sensors work the same way; they use known information from hardware instruments in other areas to calculate the conditions in areas where instrumentation isn’t quite successful — such as an OTSG — in real time. They also have the ability to self-calibrate whenever the actual value is available, allowing for faster adaptation to prevailing conditions. For example, in the OTSG application, laboratory steam-quality values from periodic boiler blowdown samples can be used for soft sensor self-calibration.
The measured steam qualities from soft sensors can then be used to apply corrective action in OTSG controllers so that steam quality can be controlled in a tighter range with reduced variation in the amount of fuel used.
Suncor worked with the U of A to develop and test a steam quality soft sensor for its Firebag in situ facility. The technology was implemented for a number of OTSGs in 2014. The result of this work has significantly reduced the variability in steam-quality measurements, in turn enabling the OTSGs to be run closer to the limit, increasing overall steam quality.
Building on the success of Suncor’s use of soft sensor technology, Canadian Natural worked with the U of A to conduct a feasibility study for applying the technology to the OTSGs at the company’s Primrose North in situ operations.
Canadian Natural provided the U of A with operating data from its OTSGs in order to complete a preliminary evaluation. The goal was to determine if the U of A’s algorithms would apply to its OTSG configuration, which is slightly different from Suncor’s.
The U of A’s feasibility study report indicated a number of potential benefits related to steam quality control using the soft sensor approach. From there, Subodh and his team worked with the U of A to develop the final algorithms for the soft sensor model that would be implemented for selected OTSGs at Primrose North for further testing.
The report from those preliminary tests indicated an improvement over the current steam quality calculations. The next step is to implement soft sensors into the control systems for selected OTSGs at the facility for commercial evaluation.
The initial feasibility study conducted by the U of A and tests on Canadian Natural’s OTSG data indicated soft sensors could allow Canadian Natural to improve steam quality by up to two per cent. That doesn’t seem like much but, according to Subodh, a two per cent increase in steam quality at a SAGD facility would result in an eight per cent reduction in boiler blowdown, up to a one per cent reduction in greenhouse gas emissions and a two per cent increase in oil production. Production and efficiency improvements would pay for the technology within a couple of months.
The original technology was developed by the U of A in partnership with Suncor. Canadian Natural learned about the technology at COSIA’s Water Conference in March 2014.
The ability of this technology to achieve higher steam quality by operating the OTSGs more smoothly was what caught Canadian Natural’s attention at the conference. The fact it could be implemented quickly and at a low cost was a potential added benefit.
Canadian Natural was also able to use the results of Suncor’s field trials to accelerate the timeline of its own project.
“This is a great example of how the COSIA model allows members to leverage the work done by companies to improve the environmental performance of their own operations,” says John Brogly, Director, COSIA Water EPA. “Canadian Natural will then share the results of their own work with the Water EPA, allowing other companies to use that information to improve the technology further.”