Events


> East101's Webinar Series


> Event Archive


> Webinar Archive

 


East101's Webinar Series

Introduction to Reservoir Simulation for Non-Petroleum Engineers



The worldwide effort in transitioning to renewable energy sources and achieving zero emissions present opportunities for growth in the biomass conversion sector, particularly in Malaysia, where an abundance of biomass waste sources are produced.

However, biological waste-to-energy transformation processes such as Anaerobic Digestion exhibit complex behaviours that are non-linear and stochastic in nature, requiring the use of advanced models to perform biogas production rate forecasting and development of process optimization strategy.

In this webinar, we will present how mathematical and deep learning approaches were used to overcome the aforementioned challenges. showcasing the use of five-stage hyperparameters tuning pipeline to develop an optimal Deep Neural Network (DNN) in PyTorch and PySwarms based on feed composition and biogas yield data collected from a validated kinetic model-based Aspen Plus dynamic simulation.

The performance of the DNN based on optimal parameter settings for conventional backpropagation (BP) and metaheuristic Particle Swarm Optimization (PSO) are also critically compared and discussed.



Webinar Recording and Slide Deck

Click to view




Back


© 2003-2020 East 101 Sdn Bhd. All Rights Reserved.