Design and implementation of a digital twin for a biomass-to-gas plant

Abstract

This paper presents the design and implementation of a Digital Twin (DT) for a synthetic natural gas pilot plant, comprising a dual fluidised bed steam gasification and a fluidised bed methanation. The DT uses real-time data within multiple tools, including model predictive control (MPC) for gasification, PI control for methanation, and process monitoring via mass and energy balances in the process simulation software IPSEpro and a MATLAB-based soft sensor. Data communication is automated using a Microsoft Azure cloud solution. By automating the plant through MPC, efficiency gains of 2 % were achieved compared to manual operation, with an additional 3 % gained by optimising operating points. Furthermore, the MPC reduced the fluctuations of the product gas amount significantly. The DT also enabled the consolidation of measurement data and estimation of parameters that can only be measured offline, such as water and tar content in the gasification product gas, providing real-time estimates where online measurements are not available. The DT's scalability to demonstration-scale plants was also validated; only minor adjustments were necessary. The proposed system demonstrates the potential for fully autonomous plant operation with minimal supervision, providing significant efficiency improvements.