DCS-Layer Advanced Process Control Consortium
Sparse matrix MVC/MPC in DCS to unlock the highest possible ROI
DCS-Layer Advanced Process Control Consortium
Sparse matrix MVC/MPC in DCS to unlock the highest possible ROI
Neglected Insights
- Unstable multivariable control loops identified by steady-state RGA can be closed-loop stable when dynamics (dynamic RGA) are considered. Controls of dual-end fractionators are successfully implemented in modern DCS as two separate PID loops.
- Spare matrix MVC can be more effectively implemented in modern DCS using standard DCS features without additional license or subscription fees.
- Inferential properties can be installed in modern DCS using standard DCS features.
- Open-loop feedforward is proven to be more effective and robust on Olefins/NGL trains than MPC or AI controllers.
- Hybrid AI models that combine traditional first-principle equations and deep leaning AI plant models can be used with MINLP to validate the structural design of DCS-Layer APC.
Top 10 Misinformation - Videos Coming Soon
---
---
---
1. Flaws of steady-state relative gain analysis (RGA)
---
---
---
2. Flaws of dynamic decoupling in MVC/MPC
---
---
---
3. Misrepresentation of LP solutions
---
---
---
4. Flaws of Deep Learning Spaghetti Models
---
---
---
5. Weakness of Deep Learning AI Controller
---
---
---
6. Transitions Between S.S. & Ramp
---
---
---
7. Flaws of IMC PID tuning
---
---
---
8. Hidden PID Performance Curves
---
---
---
9. Hidden MPC Characteristics in PID
---
---
---
10. Neglected Nonlinear PID Features
---
---
---
11. Neglected PID Tuning Parameters