Advanced Process Control for the Process Industry

The Multivariable Model Predictive Control (MPC) Technology is the most widely adopted Advanced Process Control technology in the Process Industry. Almost all major process units in the Refinery and Petrochemicals industry such as Crude Distillation Unit, Fluidized Catalytic Cracking Unit, Reforming Unit, Ethylene, Aromatics, Fertilizers, Methanol, etc. have implemented Advanced Process Control using MPC Technology. The main reasons why MPC technology is popular are:

  •  Intuitive; operator and engineer can easily understand the technology without special training
  • Progressive; the technology can be easily layered on top of the base-layer PID controller
  • Robust; the technology is Robust in that it has the ability to handle disturbances caused by changes in operation mode and/or feed characteristics. The application can also run in crippled mode; that is some of the Manipulated Variables or Controller Variables are taken off-line due to base-layer or process problems
  • Attractive ROI; while it may cost a few hundred thousand dollars to implement the MPC technology on a process unit, the return of the investment is about 3 to 12 months only. The Technology generates the quick return by pushing the high-value-add and/or high through-put process units to operate closer to constraints, to maximize high value products yields, higher through-put and lower energy consumption.

In last decade, the industry focused on implementing large applications that control hundreds of Controlled Variables and Manipulated Variables. This is due to the fact that a large control application is better than a few smaller controllers, because it is able to push the processes to operate closer to its true optimum. For example, the Vacuum Distillation Unit (VDU) which is receiving Reduced Crude from the Crude Distillation Unit (CDU) has a Heater Skin Temperature constraint. If the two columns are controlled by two independent controllers, it will be difficult to develop applications to make the following decisions:

  • Relax separation at CDU to leave some Gasoil in residue which can be recovered in VDU more efficiently, subjected to heater constraint
  • Reduce VDU heater load by shifting the load to CDU heater
  • Moderate CDU charge rate subject to heater constraint

Many Refinery CDU’s and VDU’s are therefore controlled by one large controller, to take advantage of the controller ability to push the Process Unit to operate closer to true constraints.  Though large applications serve the purpose of achieving overall operational objective, an operator may be at a loss to understand the control behavior as the controller structure is not transparent to him. Hence it is convenient, from the operator’s perspective, to define the Crude Unit control application in terms of operation group’s – for example Crude Heater, Crude Distillation Column, Vacuum Heater and Vacuum Distillation Unit.  

Similarly, there are large MPC applications that control across operation groups such as Diesel Hydrotreating Units in on-site operation groups, and Diesel Blenders in off-site operation group controlled by one large application. In fact, such control strategy is so popular that most of the Low-Sulphur diesel blender with continuous run-down are controlled using such an application.

In the Petrochemical industry, a similar approach has been made to control large process units such as Ethylene, Aromatics, Ammonia, and Methanol with one single large controller. While the Ethylene unit typically has two operation groups; Hot Section and Cold Section, it is very common for both to be controlled by one large application.

While this large control application trend will continue, especially in upgrading those old applications developed in the 80’s and early 90’s, another trend that is emerging is to implement MPC on much smaller process units, that have a much lower return on investment. Smaller process units include power plant boilers, digesters in pulp mills, and paint shops in the automotive industry. In a Refinery and Petrochemicals complex, the incremental cost to deploy new MPC application on these smaller units is very small, and therefore even with small economics as low as USD25/hr, it is viable to install an MPC application – at the same time, advances in MPC technology have significantly reduced deployment time and cost.

Are you enjoying this article?? Follow us on LinkedIn and Facebook.. 

The key break-through comes on 3 fronts; firstly the Open System such as OPC (Ole for Process Control), and advancement in software technology has made it possible to eliminate man-months of Engineering to interface the MPC Software to the DCS (Distributed Control System). Secondly, the advancement in Automatic Step-Testing significantly reduces the Engineering hours required excite the process for dynamics response. Automatic Step-Testing technology is not only beneficial for a new project but also for periodic control model updates and revamp projects. Thirdly, advancement in control technology makes it possible to embed a Quality Estimator that infers Product Quality, in real time, inside the MPC Controller, resulting in tighter integration and improved control performance.

Controller benefits are sustained by measuring the Control performance online and triggering Automatic Step-Testing, in case of plant-model mismatch, to update the Controller model.  

Going forward, we can expect MPC technology applications for not just large complex process unit, but also for smaller process units. 



Read about how effective disturbance handling with APC  led to higher yields in Delayed Coker Units here, and download the full whitepaper on this topic by clicking the link below: