gPROMS Operational Excellence Solutions (gOES) are model-based automation solutions that allow high-fidelity models to be deployed within, for example, plant automation or purchasing decision support systems for the first time.
They exploit the power of the gPROMS platform's robust and rapid equation-oriented solution to provide advanced process monitoring, optimisation and decision support capabilities as a core component of intelligent manufacturing systems.
A typical example is gPROMS Utilities, a set of energy optimisation solutions that provide planning, monitoring and operator advisory information for minimising the day-to-day cost of operating complex plant utilities systems.
The following gOES have been deployed at customer operating sites:
gPROMS Olefins Cracking Monitor is a powerful ‘virtual multisensor’ that uses your plant data in conjunction with a high-fidelity model of the cracking process to provide accurate, up-to-the minute yield prediction, current coking state and a wealth of other furnace operational information in real time.
Model-based automation overview
PSE's model-based automation architecture is shown in the diagram below. gOESs can comprise some or all of the functionality in the blue boxes:
The PSE components interface with standard plant automation system components, including Advanced Process Control (APC) elements such as Model Predictive Controllers (MPCs), via PSE's OPC client.
The key components of the model-based automation approach applied within gOESs are:
- Dynamic state estimation. This uses plant data in conjunction with the model to provide reconciled values for all key model parameters as well as an up-to-date plant model that can be used for many other activities.
- Monitoring and diagnosis. The reconciled values provide a wealth of soft-sensed information for monitoring the plant state, including – for example – equipment fouling, furnace coking and catalyst deactivation, and diagnosing potential problems.
- Decision support tools. The up-to-date model can be used on-demand to provide decision support – for example, by using dynamic simulation to determine the effects of proposed feedstock or product grade changes
- Non-linear model-predictive control. gPROMS's dynamic optimisation capabilities can be used to develop next-generation non-linear model predictive control based on high-fidelity dynamic models
- Model reduction and linearisation. The gPROMS dynamic model can be used to generate up-to-date linear models at any operating point for use in the plant MPCs, bypassing the need for extensive plant perturbation and ensuring that the underlying MPC models closely reflect the current operation
- Setpoint and trajectory optimisation. The gPROMS steady-state and dynamic optimisation capabilities can be applied to the up-to-date plant model to generate truly optimal steady-state setpoints to provide to the MPC, as well as the optimal setpoint trajectories required to achieve these.
gPROMS Operational Excellence Solutions bring next-generation capabilities to real-time optimisation (RTO) based on rapid and robust high-fidelity plant models that can be continually maintained up-to-date, irrespective of whether the plant is at steady state or not.
This brings many advantages and benefits:
- Accurate up-to-date values can be provided for every variable in the plant, providing a wealth of key performance indicator (KPI) information to operators as well as soft-sensed values for use in control.
- Accurate linear models can be generated for use in MPC at any state of plant operation, without the need for lengthy and disruptive plant perturbation.
- gPROMS's powerful optimisation capabilities can be used to generate optimal setpoints for steady-state operation based on accurate current plant state, without the need to wait for steady state. This avoids the errors implicit in the steady-state assumption, providing optimised setpoint values much closer to the real optimum.
- Dynamic optimisation can be used to generate optimal non-linear trajectories, in order to minimise transition times on feedstock, product grade and other transitions. This can help avoid significant periods of off-spec or non-optimal production.
- Because the dynamic state estimator does not require the plant to be at steady-state (a typical requirement of most existing real-time optimisation applications) before performing reconciliation, the benefits of optimisation can be realised continually rather than only when the plant is determined to be at steady state.
PSE typically provides gOESs as packaged solutions that include implementation, integration within real-time/online systems and/or workflows, maintenance and annual software licences.
Deployment includes a comprehensive turnkey service that starts with a consultation on your requirements and ends with an implementated and tested application.
Operational Excellence Solutions are configured for and deployed on specific assets, meaning that they can be transferred on sale of the asset.