Behind the Science: Using Smart Tomographic Sensors in Process Control

Behind the Science: Using Smart Tomographic Sensors in Process Control

Author: Barbara Boeck, Uwe HampelORCID iD

Dr. Barbara Boeck, Editor-in-Chief of Chemie Ingenieur Technik, talked to Professor Uwe Hampel, Technische Universität Dresden, Germany, about his article on controlling a fluid separator with tomographic sensors. The work was recently published in Chemie Ingenieur Technik.

What is tomography? What was the inspiration behind this study?

Many people know computed tomography from medical imaging, where it has originated. In the 1970s, X-ray tomography revolutionized medical diagnostics when doctors were supplied with detailed images from the interior of the human body. Obviously, tomographic imaging techniques are also of great interest for industrial processes, because you can get insight into opaque process vessels. However, there are some fundamental differences between medical tomography and what we call process tomography.

In process diagnostics, we try to avoid radiation and, hence, we use different tomographic modalities, such as electrical tomography. Industrial processes need to be controlled and almost every sensor in a process plant is in some way connected to a control system. This is exactly where the inspiration sparked. Can we use process tomography techniques, which produce massive amounts of data and require complex processing algorithms, for industrial process control?

The answer is clearly yes, but it requires a new type of thinking in sensor, control, and automation engineering. We have to turn bulky tomographic scanners into smart sensors that can be integrated into industrial plants. Tomographic data processing, which consists of solving very large linear or non-linear systems of equations, must be made very fast in order to deliver data in milliseconds. Moreover, a new control theory for massive amounts of data is required.

Separation techniques seem to be a popular topic at the moment, with a focus on digitalization and “Industry 4.0”. Why was your attention focused on this particular area?

In fact, fluid separation is a prototypical operation that is nicely suited to demonstrating how fast and smart tomographic sensors can be used for process control. The in-line fluid separators that we look at rely on fast flows and, hence, the control action needs to be fast, as well. This is why we have chosen this application to demonstrate the capability of this type of control system. Besides that, fluid separation is widely used in chemical and petrochemical processing. Thus, we think that providing solutions here will also have a great technological impact.

make imaging sensors smart using sophisticated technological concepts, such as new materials, intelligent algorithms, and fast parallel computing. This fits very nicely with the trend of digitalization in the process industry and could be groundbreaking for the field of process sensors and automation.

Could you please summarize your findings?

We have introduced the principle of controlling a fluid separator with tomographic sensors. The task is to separate fluids of different densities while they are flowing in a pipe. We have studied this based on the example of a gas-water flow. However, it may be also water-oil or something else.

The separator essentially consists of a so-called swirling device that sets the flowing mixture into swirling motion. Just as in a washing machine, the denser phase—i.e., the water—will be accelerated towards the pipe wall, while the lighter phase—i.e., the gas—remains in the center. This way, the gas can be extracted with an extraction tube in the center of the pipe.

This concept has been used for decades. However, it is not controlled. This means that whenever the flow changes—e.g., when the gas fraction increases or decreases—the separation is less optimal and water can get into the extraction tube or gas is carried over into the liquid outlet. Therefore, we propose installing tomographic sensors upstream and closely downstream of the swirling device and using the tomographic data to drive a control valve in the gas line. Depending on the flow velocities, we need a quick action within milliseconds. This is challenging for both the sensors and the actuator. In order to support the system design, we also use computational fluid dynamics to simulate the flow conditions in the device.

How soon could your results lead to widely used applications in the industry?

Very soon, I hope. In our European training project, we are aligned with 15 industrial partners who advise us, but also have a decent look at technology transfer options.

What is the longer-term vision for your research?

In-line fluid separation is just an example of using smart tomographic sensors in process control. Within our European training network, we also pursue other applications. These are, for example, the control of microwave drying processes, the control of batch crystallization, and the control of continuous steel casting. This way, we have the chance to demonstrate the capabilities of this new concept in a broad range of industrial applications. The long term vision is clearly to deploy such technologies and, in this way, provide a new type of process control.

The article they talked about




Leave a Reply

Kindly review our community guidelines before leaving a comment.

Your email address will not be published. Required fields are marked *