Unai Saralegui Vallejo of Tecnalia tells us about his work in the project
Tell us about your work in the HypeCOG project so far
TECNALIA is involved in various activities in HYPERCOG and we have a team made up of experts in artificial intelligence (AI), statistics, materials, BlockChain and computer vision is working on the project. My team is mainly focused on the cement use case, more specifically working on the data acquisition and analysis of different sections of the cement plant by using statistical analyses.
Currently, we are working closely with experts at the ÇIMSA plant to better understand the cement production process and identify which are the most relevant sensors installed and the key parameters related to them which affect to final product quality. At the moment, we are mostly focused on the development of soft sensors for the measurement of particle size distribution (PSD).
Can you provide some further details about the progress of this work?
We are following the “Hardware In the Loop” (HIL) research line, which states that in a real physical system, a set of different virtual additions can be added with the aim of studying the behaviour of the system under certain circumstances. In our case, we are creating a virtual sensor, also known as a soft sensor, which is a virtual entity that acts like a sensor giving as output a value we want to measure. Instead of getting the information directly from the physical world, however, it estimates the value from other related sensor readings.
The use of soft sensors is interesting in cases when a variable cannot be measured directly because it is impossible or too expensive to install a physical sensor or there is no real sensor that can do the job. Considering that, we have made a huge effort to understand the process in this use case and to analyse the data. We will use this data in the development of the soft sensor to be able to estimate the Particle Size Distribution (PSD) of the mill output.
The prototype to be installed is aiming to achieve a more energy-efficient process for cement manufacturing. How does your work on the soft sensors contribute to this aim?
When referring to an energy-efficient process, we must also pay attention to the fact that the quality and performance of the process should be kept at optimal levels, too. To do so, we need to be able to evaluate the quality or performance of the process to make modifications to the behaviour of the process.
In this case, the measurement of the PSD is a complicated process that can only be done in a laboratory by using expensive machines. Although there are commercial solutions that can be installed in the process line, the cost of implementing them is too high. The development of a PSD estimation soft-sensor, however, will allow us to estimate the PSD values and will act as a quality measurement device to evaluate the process in near-real time, too. Our development will help in determining the quality obtained in the production in near real-time, thus enabling the development of advanced control strategies that look for resource-efficient processes while maintaining the quality of the generated materials.
How does this work link with other work being done in the project and how are you working with other partners?
TECNALIA is involved in the majority of the activities in HYPERCOG, so we have an overall knowledge of all the tasks taking place in the project. That gives us a good overview, meaning we can detect parts that may be reused or adapted, especially in terms of the data acquisition and analysis at the three use cases.
For example, we are closely collaborating with UPEC in the development of the knowledge factory by helping them gather data and knowledge from the cement use case. We are also closely collaborating with EKODENGE, CYBERSERVICES and ÇIMSA to be able to securely extract data from their systems and understand specific details of the cement plant.
What other potential uses will there be for the work you do in HyperCOG?
TECNALIA is also working in the steel use case at SIDENOR in Spain, which mainly focuses on the development of a hyperspectral camera-based sensor to estimate the composition of the ladle slag. This is interesting, as one of the aims here is to be able to reuse a material that is a by-product in one industry but is critical in other, related industries. Furthermore, this tool will also help the plant reduce the effort and duration of the current method in use. So, by using the HyperCOG solutions here, the slag generated at the Sidenor steel plant may be able to be used in the cement production process at ÇIMSA. A blockchain-based tool is also being developed to ensure the quality and security of the properties of the materials to be shared and maintain confidence in the characteristics measured in the steel plant.
What do you hope to achieve through your participation in HyperCOG?
Our aim is to develop a PSD estimation model that can be easily adjusted to be used in other cement factories different to the use cases being used in the project. That could make it easier and more affordable to measure relevant values and develop more advanced control strategies in these types of industries, leading to the development of more resource-friendly strategies.