Executive Summary : | Real-time process monitoring and fault diagnosis are one of the major components of the fourth industrial revolution (Industry 4.0), which ensures the operational safety, economic optimality, and environmental viability of modern chemical/petrochemical plants. Faults are abnormal process deviations, tend to degrade the process performance initially, and further may lead to hazardous incidents causing huge damage to plant personnel and the environment. Therefore, there exists an urgent need for the development of efficient online methodologies for early detection of process faults, to bring the process back to normal operating conditions without causing any damage. ` Artificial intelligence (AI) and machine learning (ML) techniques have been gaining a lot of attention recently, found to be promising in recognizing the underlying patterns of large amounts of data resulting from different application domains including finance, climate, bioinformatics, medical industry, etc, and the application of AI methods in chemical process plants is very limited. Moreover, recent advancements in instrumentation and control systems of modern chemical plants lead to the availability of huge process data in the form of sensor measurements, which necessitates the need for the development of efficient AI modules for real time monitoring of chemical plants. Catalytic reforming is one of the most widely used processes for the large-scale production of H2 and Syngas. Especially in petrochemical industries, reforming is performed in fixed and fluidized bed reactors to produce high octane gasoline from naphtha and it involves several cracking, dehydrogenation, and isomerization processes simultaneously. Managing abnormal situations (failures of pumps, sensors, actuators, controllers, catalytic deactivation, etc) only by human intervention in such large, complex, and high dimensional processes is not an easy task as the disturbances spread plant-wide and the task becomes more complex with the increase in plant size. Therefore, it is proposed to develop efficient AI methodologies for real-time monitoring of complex catalytic reforming processes and demonstrate their application experimentally. The interdisciplinary (catalysis, chemical engineering, and modeling) team of investigators has adequate expertise in developing novel AI methodologies for online motoring, catalytic systems for reforming processes, and computational & experimental facilities. Although, multinational companies such as GE, Honeywell, Shell, etc supply similar modules, which are highly expensive and mostly non-affordable and there exists no indigenous technology for online monitoring and fault diagnosis of complex chemical/petrochemical plants. The proposed work fulfills the make in India initiative of our government and contributes to the advancement of the modern Indian chemical industry. |
Co-PI: | Dr. Soujanya Yarasi, CSIR-Indian Institute Of Chemical Technology, Hyderabad, Telangana-500007, Dr. Nakka Lingaiah, CSIR-Indian Institute Of Chemical Technology, Hyderabad, Telangana-500007 |