Energy Monitoring & Optimization for Regenerative Thermal Oxidizers

Ajinomoto Omnichem (AOC) is a subsidiary of Ajinomoto Group, a Japanese multinational food and biotechnology corporation active in a wide variety of areas including: seasoning & food, frozen foods, healthcare, chemicals, and semi-conductors. AOC specialises in pharmaceutical CDMO (Contract Development & Manufacturing Organization) services, agrochemical solutions, botanical extractions, and amino sciences. Ajinomoto operates a Regenerative Thermal Oxidizer (RTO) on its plant in Wetteren. In essence, an RTO is a combustion device that controls the emission of volatile organic compounds (VOCs) and other hazardous air pollutants (HAPs) by converting the toxic particles into CO2 and H2O.

Rock Around the Clock

This conversion within the RTO takes place by burning natural gas to heat the contaminated air up to 800°C. Through a vent, contaminated air is drawn into the inlet of the RTO. The contaminated air flows through the RTO and is heated. At this moment, when the contaminated air is hot enough, the conversion takes place and the clean air leaves the outlet of the RTO. Obviously, making sure that the temperature is kept high enough to convert the VOCs and HAPs, significant energy is required in the form of natural gas. Ajinomoto’s RTO is operating 24/7, even when it’s not really needed. The RTO is responsible for around 20% of AOC’s total gas consumption, representing a significant cost. Ajinomoto is looking for a way to decrease the RTO-related energy bill and teamed up with Faktion to investigate this matter and find a data-driven solution to optimize energy use for the RTO.

Balancing Volume and Concentration: a Legal Limitations Trade-off

Ajinomoto is able to adjust the pressure differential, which is the difference in pressure between the right before and after the inlet fan of the RTO. This difference is created by a fan that is spinning. The fan and thus the pressure differential lead to air being sucked into the RTO. The bigger the pressure differential is, the bigger the volume of contaminated air that flows into the RTO. If the pressure differential is high, the RTO has to burn more natural gas in order to keep the temperature high and to process the increased volume of contaminated air. Hence, a higher pressure differential results in more natural gas use and thus an increased energy cost.

One might argue that Ajinomoto Omnichem could just lower the pressure differential, leading to less contaminated air flowing into the RTO and hence less natural gas use. However, AOC needs to comply to legal emission limitations. Exceeding this legal threshold would lead to a fine. The concentration of toxic particles in the contaminated air is not a constant, since the air is coming from different chemical processes that are by definition not constant. Generally speaking, lowering the inlet flow of contaminated air increases the concentration of toxic particles, making it difficult for the RTO to convert all particles. As a consequence, more unconverted particles are emitted, leading to potential emission breaches. Hence, there is a trade-off between pressure differential (or inlet volume) and legal emission breaches, ensuring that no one-size-fits-all solution is available to decrease the RTO’s energy consumption. The proposed solution entails a scenario analysis where recommendations can be extracted to decrease the energy consumption of the RTO.

Data-driven Insights

In a first stage, Faktion conducted an exploratory data analysis to gain insights into the RTO data at hand. Together with AOC engineers, our ML engineers conducted a thorough analysis in order to understand the different factors that influence the energy consumption of the RTO before making recommendations for improvement. In essence, there are five interlinked building blocks towards an optimization solution. In the exploratory data analysis, these five building blocks were quantified as they were needed as input for the scenario analysis. The following insights where found:

  1. As stated above, there is a positive correlation between pressure differential and natural gas use, as explained above.

  2. There is a positive relation between the pressure differential and the inlet flow, as explained earlier.

  3. There are periods where a lot of emission incidents take place and periods where AOC remains well below the legal emission boundary.

  4. There is a negative relation between the efficiency of the RTO in converting toxic particles and the pressure differential. The higher the pressure differential, the more volume is pumped into the RTO, the lower its efficiency will be.

  5. Generally speaking, the lower the number of toxic particles coming in, the lower the number of toxic particles leaving the RTO will be. A regression model that predicts the outlet of toxic particles as a function of the inlet of toxic particles showed that there is a certain area of opportunity where the outlet of toxic particles is well below the legal limit.

All these insights were combined in a scenario analysis. The logic behind the simulation is that gas consumption can be reduced by reducing the pressure differential. However, because of the trade-off between lowering the pressure differential and the risk of incidents, the question becomes: how much should we reduce pressure and when can we do that without causing emission incidents? In essence, the answer to this question boils down to predicting the outlet of toxic particles and comparing it to the legal limit. In periods of time where the outlet of particles is well below the legal boundary, we can calculate how much additional inlet hydrocarbons are allowed before the legal limit is reached, taking into account a certain safety threshold. This additional inlet can be converted into a flow reduction, as a lower volume increases the concentration of particles. To achieve this flow reduction, the pressure differential is decreased, leading to less energy consumption.

The Result: Insights and Optimization of the RTO’s Energy Consumption

This solution, together with the relations stemming from the five building blocks described above, were embedded in an interactive visualization tool. The tool automatically calculates the optimized pressure differential and the energy and cost savings, compared to the non-optimized situation, based on some input parameters, specified by the user (e.g. the natural gas price or the safety threshold). As an initial scenario with base settings, we estimated a monthly energy cost saving of 4,2%. The solution proved to be a very useful and insightful tool for Ajinomoto’s engineers to extract relevant insights about the working of the RTO, enabling them to optimize Energy Consumption.

Get in touch!

Inquiry for your POC

Scroll to Top