CHENNAI: An IIT-M research team has developed and used an artificial intelligence tool to study the production of gaseous fuel from biomass, which is expected to save time for experimental researchers involved in developing reactors and new techniques to convert biomass into products.

The researchers said they developed a recurrent neural network (RNN) model, a machine learning model, to study the reactions that occur during the conversion of biomass into biofuel and chemicals. Researchers can use this model to come up with solutions, processes and plants to convert biomass to biofuel. "Experimental researchers can directly use this machine learning based model that is expected to save the time it takes for them to develop new processes. Using this technique, the researchers can come up with more optimal and economical ways to convert waste biomass into fuels and chemicals," said professor Himanshu Goyal from department of chemical engineering, who worked with professor Niket S Kaisare also from department of chemical engineering, and Krishna Gopal Sharma, a fourth year B Tech student from department of computer science and engineering.

The researchers said at present there are highly accurate and detailed models available, but most models are computationally expensive i.e take a long time to become operational. Artificial intelligence tools such as machine learning can hasten the modelling process. Also, experimental researchers working to develop new reactors and new techniques to convert biomass into useful products work on a trial and error method which is time consuming and expensive.

"What we did is we used a highly detailed computational model we have in our lab to generate data and then use the RNN technique to come up with simple, elegant and computationally fast solution which can be directly used by experimental researchers in their research so that the trial and error process reduces," Prof Goyal said.

Prof Kaisare said, "The novelty of our machine learning approach is that it is able to predict the composition of the biofuel produced as a function of the time the biomass spends in the reactor. We used a statistical reactor for accurate data generation, which allows the model to be applied over a wide range of operating conditions." The IITM research team also plans to use the AI tools in different aspects of clean energy like carbon capture and the electrification of the chemical industry.