Machine Learning Tool for Prediction Amine Emission from Carbon Capture Technology

Statement: Machine learning (ML)-based software will be developed for the efficient and accurate analysis of amine emission data, which will play a pivotal role in the rapid development of carbon capture and storage (CCS) technologies. The tool will revolutionize the utilization of emissions data, providing valuable insights and enhancing monitoring capabilities for carbon capture plants. Its ability to enable real-time monitoring and prediction of amine emissions empowers plant operators to implement timely interventions and optimize operations, thereby minimizing environmental impacts. General statement of how this problem is being addressed: We will develop the PREMAM module, utilizing deep learning techniques like recurrent neural networks (RNN) to forecast amine emissions in carbon capture plants. It will involve constructing a robust forecasting model integrating time-dependent process and emission data, employing techniques such as Stacked LSTM, Bi-directional LSTM, and Convolutional LSTM. By prioritizing causal impact analysis, we will assess emission influences under different conditions and explore mitigation strategies using "what-if" scenarios, leveraging data from the CESAR1 solvent testing campaign at Technology Center Mongstad (TCM) under different parametric tests. What is to be done in Phase I: We aim to develop PREMAM, a data-centric module, for forecasting amine emissions in carbon capture plants. It integrates historical and present operational data and employs deep learning techniques to construct a robust forecasting model under various operational scenarios. We also plan to integrate the developed module into our in-house data analytics platform, MatVerse. Commercial Application and Other Benefits: The envisioned inclusive, intuitive, and machine learning-powered data analysis tool, MatVerse, equipped to integrate and analyze amine emissions datasets, will prove highly beneficial for carbon capture and storage (CCS) technology, as well as industrial process monitoring. It serves a variety of industries, such as oil and gas, coal and biomass power plants, iron and steel, chemicals, and others. Notably, the software's competitive edge lies in its accessibility, allowing non-technical experts to leverage advanced analysis tools through handheld devices or desktop computers.