Special Section: Networking/Ethernet
A crude vision
Vision-based Ethernet system gives real-time sight in oil, gas applications
By Thomas M. Canty
Vision analysis can provide a dimension into process and product control that has not been previously available. The technology affords the process engineer a view into the live process and enables direct calibration of the control parameters based on what is actually occurring and not indirectly based on previous data. With current Ethernet technology, the information can be broadcast to key decision-makers as process events occur, enabling a concentration of expertise to more quickly realize new discoveries and resolve process aberrations as they occur.
Vision systems are now developed to the point where they can provide analysis and control in the oil and gas industry where other instrument types cannot. Improvement in design of hardware and software, and advancements in Ethernet communication technology have allowed vision-based systems to see into process environments, such as crude oil, to detect and analyze particle constituents. Vision can also effectively handle applications involving the monitoring and control or produced water concentration, water-in-oil analysis, multi-cut sand, oil and water analysis, and even flame and tube analysis in the refining process with the benefit of the operator’s visual optimization.
Vision analysis, as proven in many applications such as polystyrene bead manufacturing, sugar crystal growth, cell growth, and abrasive manufacturing to name a few, provides an extra dimension to process analytics that enables the process engineer to make evaluations based on his own expertise and to use this information to better set software parameters that enable the system to more accurately analyze and control the process. Newly developed hardware systems can see into these processes and provide a static or streaming video image of the process fluids in real time. Incorporating Ethernet connectivity into the hardware system then allows information to transmit and see immediate use anywhere an Internet connection is available.
The conditions of the application require a vision system to survive in the drilling environment, maintain explosion-proof electrical ratings, and be easy for operators to use to be effective. It is important to first look at hardware mounted to the process.
The viewing mechanism of the vision system is most important. Rugged high-pressure windows through which users can conduct illumination and imagery must fit with the vision system. These windows must also be adjustable to allow proper control of the measurement zone for accurate results. The best option is the fused, glass-to-metal seal that meets the criteria and allows the formation of an adjustable viewing cell inside the process that can withstand thermal and physical shock, as well as several thousand pounds of pressure. A typical system will be comprised of an inline body with camera, light, and light pipe attached. The high polish of the glass surface helps to maintain the windows free of process buildup.
Illumination is critical to the success of vision-based technology, and a consistent, high intensity source is the building block that allows software to define particles consistently and accurately. Another important feature of a complete system is the ability to control the measurement zone where the particles are illuminated and which the camera sees. The light pipe component has adjustment capability in order for the gap between the camera and light pipe (measurement zones) to be adjusted to optimize the sensitivity of the system. Fluids can range from relatively clear (low turbidity) to quite heavy in particulate matter (high turbidity), and the control of the measurement zone enables the control of illumination by means other than adjustment of intensity from the bulb.
Vision systems are generally based on the same concept as the human eye—brain interaction. We see objects in our field of vision because they stand out from the background. This delineation is created by differences in light intensity around the object caused by shadows or colors. The vision system detects particles in much the same way. Intensity differences from the background field occur at particle boundaries. The vision system detects these and is able to form the particle based on defined threshold information, which sets the level of intensity change that defines a particle boundary. The operator inputs the threshold information based on what he sees on screen. This is the powerful advantage of vision over other types of instrumentation. Reliance on indirect calibration is not required. The actual particles define what is a correct size detection and what is not. The hardware and software principles are largely the same regardless of applications.
The major by-product of oil and gas production in the U.S. is produced water, which must undergo processing for injection back into the formation or disperse to surface water bodies. The latter requires disposed water to meet standards of the U.S. Congress’s Clean Water Act. The vision system can monitor this water for multiple particle constituents, such as oil droplets or other entrained solids and bacteria. It can also provide a turbidity reading to indicate total clarity. Additional applications include the monitoring of effluent for fuel transfer stations, fuel storage drainage systems, vehicle wash-down stations, whip bilge water, oil/water cooled machinery runoff, and others.
Water in oil
Separating water from oil is mostly accomplished by chemical treatment, which encourages the coalescing of small water droplets into larger droplets or formation of droplets so you can more efficiently remove them. The vision system provides the capability to monitor this process as it occurs and provides either automatic or manual control of the treatment process.
In the coalescing process, water droplets coalesce together and become larger in size. This is the desired effect, and the vision system provides a method for achieving this visual feedback that can assist in optimizing the extraction process. The software can, in both images, detect the water droplets and formations in order to provide real-time and continuous data on the changes in the crude as the process moves forward.
Crude oil analysis
Analysis of the crude is necessary in the refining process. Crudes contain varying levels of water, sand, asphaltenes, resins, paraffin, and other constituents that may impact refining, storing, and transportation methods. In these refining steps vision can be of great assistance in defining when a process has reached an appropriate level of development.
This same camera technology, in a slightly different configuration, can see use to view the inside of extreme atmosphere enclosures such as reformers and boilers. Significant maintenance issues regularly arise in the reforming furnaces whereby process irregularities, such as burner malfunction, catalyst changes, feedstock variations, and the like, can cause imbalances in the thermal profile of the tube, which can lead to rupture. The result is an expensive repair and lost production of the furnace.
Proper thermal design can ensure a camera system that can withstand the rigors of a high temperature environment. The software then allows for analysis of various aspects of the furnace operation, including the temperature and flame size and temperature. The vision system can detect thermal profiles and shapes (flame) as the human eye would. Metal temperatures above 750˚F are visible to camera technology, and thus you can monitor and analyze them with a vision system.
ABOUT THE AUTHOR
Thomas M. Canty is president of J.M. Canty Inc., a consultant in vision analysis across process industries, now specializing in manufacturing sight glass, industrial surveillance, and fiber optic lighting, in Lockport, N.Y. E-mail him at email@example.com.
Image is everything
Improved hardware and software, along with advances in Ethernet capabilities, now allow vision-based systems to see into process environments, such as crude oil, to detect and analyze particle constituents.
The top left image shows particles in the stream, which you can use to define to the software what the particles actually are. The next step is to digitize the image and test thresholds to properly define what a particle is, based on the actual view of the particle.
Comparing the top right and the bottom left, it is easy to see in the bottom left image where the threshold is too low, and the digitized particles are obviously too small, while in the top right case, the digitized particles detected are over-sized and a good deal of noise is detected as well.
When you compare the bottom right image, you can see the proper threshold has been selected, as the digitized particles define the actual particles very closely.