Ammonia-based desulfurization - sample application
Gas cleaning by contact with an ammonia-containing solution can be applied to the scrubbing of acidic gases and in particular sulfur dioxide containing gases such as flue gas.
The desulfurization plants are specified for a nominal operating condition, but real-world conditions (flue gas flow, composition and temperature) change both seasonally and between day / night.
Using the simevo Process Simulation technology you can develop a "digital twin" for the gas cleaning process, using simple, empirical models or more sophisticated, custom models for the key units. The process model calculates the Key Performance Indicators (such as removal efficiency, recovery, residual sulfur dioxide content and specific utility consumption) as a function of inlet and operating conditions.
During research and development it is often required to forecast the Key Performance Indicators (KPIs) of the ammonia-based desulphurization process such as yields, efficiencies and specific utility consumptions, at conditions different from nominal:
to evaluate the return of investment (RoI)
to cross-check and validate vendor-supplied data
to compare the performance of equipment from different vendors
Once the plant is in operation, forecasting can be used for:
and predictive maintenance.
For all these purposes a steady-state, first-principle process model ("digital twin") is the way to go; just stay away from spreadsheets (because they suck at modeling !) and simulate your ammonia-based desulphurization process using the simevo process technology.
The process model calculates the KPIs as a function of inlet and operating conditions, and can be deployed as:
a small-footprint, standalone custom process simulator for desktop use
hosted in the public or private cloud for integration with other services, access via the simevo process app or via a dedicated web application
a standalone online application that communicates with the control system or with the Supervisory Control and Data Acquisition (SCADA).
In all cases you can shield the user from implementation details, modeling assumptions and hypotheses. This is useful to make the model robust and simple to use, and / or to protect your know-how.