Digital process operation involves generating value by combining current or historic plant data with the deep process knowledge contained in high-fidelity predictive models, as part of Operational Excellence, Smart Plant, Industry 4.0 and other similar initiatives.
Digital process twins configured for specific purposes can be used for a variety of applications, from long-term health monitoring to real-time optimisation, that add daily value to operations.
Digital process twins
The digital process twins use the combination of prior knowledge in the predictive process model and the up-to-date information in the plant data to generate valuable new information:
A typical application is shown below:
- a long-term health monitoring application determines the values of key parameters subject to drift over the operation of the plant – for example, catalyst activation state in a catalytic reactor, or amount of coking in a furnace coil – using the predictive plant model combined with current and historic plant data.
This provides valuable information on the current plant state to Operations and Maintenance. It is also used to update a predictive master model that can then be used in digital process twins for many other applications:
- Run-length prediction, to determine the expected end-of-run date under different operational scenarios. This can be used to improve maintenance scheduling, or to determine the most profitable operation mode for the remainder of the run.
- Real-time soft sensing, to provide reliable current values of KPIs such as yields, conversion/severity, coking rate, as well as equipment internal variables that cannot normally be measured. This provides valuable information for real-time monitoring of operation, or use in enhanced process control.
- Operations decision support tools, for what-if analysis of steady-state and dynamic operating scenarios. This allows operators to visualise and understand the consequences of their decisions.
- Real-time optimisation, to determine set points for economically optimal operation taking account of plant constraints. This makes it possible to maximise the economic performance of the plant from hour to hour, and react rapidly to disturbances and upsets.
Typical benefits of such applications include:
- Better operations through enhanced, up-to-the-minute decision support information
- Improved maintenance scheduling through run length prediction
- Improved economics from real-time optimisation
- Improved asset integrity from better health monitoring.
Digital process operation technologies are already being used to drive next-level productivity enhancements to operations in the Chemicals & Petrochemicals, Oil & Gas, Refining, Pharmaceutical, Food & Beverage and Water industries.
Digital process twins are implemented online using the gPROMS Digital Applications Platform (gDAP), a comprehensive framework for implementing high-fidelity predictive models within the plant automation system. Developed in conjunction with major chemical companies, the gDAP takes care of data exchange, data processing and validation, co-ordination of different calculations, and transfer of results between calculation modules.
Siemens AG and PSE recently announced a collaboration agreement that brings the power of PSE’s gPROMS technology to Siemens' automation and digitalisation offerings for the process industries.