Abstract:Numerical method is widely used for solving the mechanistic models of chromatography process, but it is timeconsuming and hard to response in real-time. Physics-informed neural network (PINN) as an emerging technology combines the structure of neural network with physics laws, and is getting noticed for solving physics problems with a balanced accuracy and calculation speed. In this research, a proof-of-concept study was carried out to apply PINN to chromatography process simulation. The PINN model structure was designed for the lumped kinetic model (LKM) with all LKM parameters. The PINN structure, training data and model complexity were optimized, and an optimal mode was obtained by adopting an in-series structure with a nonuniform training data set focusing on the breakthrough transition region. A PINN for LKM (LKM-PINN) consisting of four neural networks, 12 layers and 606 neurons was then used for the simulation of breakthrough curves of chromatography processes. The LKM parameters were estimated with two breakthrough curves and used to infer the breakthrough curves at different residence times, loading concentrations and column sizes. The results were comparable to that obtained with numerical methods. With the same raw data and constraints, the average fitting error for LKMPINN model was 0.075, which was 0.081 for numerical method. With the same initial guess, the LKM-PINN model took 160 s to complete the fitting, while the numerical method took 7 to 72 min, depending on the fitting settings. The fitting speed of LKM-PINN model was further improved to 30 s with random initial guess. Thus, the LKM-PINN model developed in this study is capable to be applied to real-time simulation for digital twin.

Petri Net Based Metabolic Network Parameters Fitting with GPU Acceleration

Modelling and Simulation of Metabolic Networks Petri Nets Approach and Perspective

Quantitative Petri Net Model of Gene Regulated Metabolic Networks in the Cell

Petri Net Models for the Semi-Automatic Construction of Large Scale Biological Networks

A Systematic Petri Net Approach for Multiple-Scale Modeling and Simulation of Biochemical Processes

Physics-informed Neural Networks to Solve Lumped Kinetic Model for Chromatography Process.

BioNetSim: a Petri Net-Based Modeling Tool for Simulations of Biochemical Processes

GPU Accelerated Simulation of Cardiac Activities.

Modeling The Krebs Cycle Based On Hybrid Petri Net

High-speed Parameter Search of Dynamic Biological Pathways from Time-Course Transcriptomic Profiles Using High-Level Petri Net.

Modeling, safety verification and optimization of operating procedures in process systems using hybrid Petri nets

Hybrid Simulation of Biochemical Systems Using Hybrid Adaptive Petri Nets

Gpu-Accelerated Evaluation Platform For High Fidelity Network Modeling

Scalable Parameter Estimation for Genome-Scale Biochemical Reaction Networks

Timed Petri Net Modeling And Simulation Of A High-Throughput Biological Screening Process

Bayesian Parameter Estimation for Biochemical Reaction Networks Using Region-Based Adaptive Parallel Tempering

Neural System Modeling and Simulation Using Hybrid Functional Petri Net

Kriging-Based Parameter Estimation Algorithm for Metabolic Networks Combined with Single-Dimensional Optimization and Dynamic Coordinate Perturbation

An advanced environment for hybrid modeling of biological systems based on modelica

Gpu-Assisted Hybrid Network Traffic Model

The design of biopathway's modelling and simulation system based on petri net