Validating a process model with experimental data is a key part in ensuring that the model is both accurate and predictive across a range of scales. PSE has recently augmented its state-of-the-art model validation technology, significantly streamlining the validation workflow. It has always been possible to process steady state and dynamic experimental, pilot or operational data from a variety of sources, estimate a large number of parameters simultaneously, and receive a detailed statistical analysis to quantify the predictive accuracy of the model. The new developments make it much easier to do so. They include easy-to-use data-import and processing, the ability to select some of the data for fitting and some for independent ‘blind test’ validation, plus significantly improved reporting and visualisation of results. In addition, the model validation activity can take advantage of PSE’s new high-performance computing capabilities, with full parallelisation to significantly reduce time for parameter estimation.

What this webinar covers

• Model validation workflow
• Data import
• Parameter fitting
• Blind tests
• Reporting
• Parallel computing

Presenter(s)

Mayank Patel, Process Systems Enterprise

Mayank Patel has been at PSE since 2012. He has over 13 years’ experience on application of advanced process modelling (APM) and optimisation, with an emphasis on reaction, adsorption and control applications. Mayank’s current focus is supporting the growth of the energy and chemicals business within PSE by driving adoption of APM capabilities and methodologies within organisations. He has recently conducted a model-based engineering project to suggest improvements for production within EO/EG plant operations. He has a particular interest in the area of adsorption, and assisting organisations worldwide in maximizing the use of their assets. Publications include technical papers on adsorption, application of model predictive control to small-scale reactors and optimisation of industrial EO reactors.