Advanced process control (APC) is a set of control techniques layered above basic regulatory control, most commonly model predictive control, that coordinate multiple variables at once to hold a process closer to its economic and operating limits.
Basic control in a DCS or PLC uses mostly single PID loops: one measurement, one valve. APC sits on top. It handles interactions between loops, dead time and constraints that single loops cannot. Techniques include feedforward, decoupling, inferential control using soft sensors, and above all model predictive control (MPC).
MPC uses a dynamic model of how manipulated variables, such as feed rate or steam flow, affect controlled variables, such as temperatures or quality, over time. Every cycle it predicts the future, solves an optimization over a short horizon while respecting constraints, applies only the first move, then repeats. The models are usually built from step tests on the live process. MPC has been used in refineries and chemical plants since the 1980s and typically writes setpoints down to the DCS loops.
APC is mainly a process-industry tool: refining, petrochemicals, chemicals, cement, pulp and paper, power. Its main limits are maintenance and trust. Models drift as equipment and feedstocks change, and if operators switch the controller off, it delivers nothing. Newer AI and reinforcement learning controllers compete with or complement MPC, but face the same test: operator acceptance and safe handling of constraints.
Questions about Advanced process control
Is APC the same as model predictive control?
No. Model predictive control is the most common APC technique, but APC also covers feedforward, decoupling and inferential control that relies on soft sensors.
What is the difference between APC and basic PID control?
Basic control in a DCS or PLC is mostly single PID loops, each pairing one measurement with one valve. APC is layered on top and deals with what a single loop cannot: loops that interact, dead time and constraints.
When does APC stop delivering value?
When operators switch the controller off: it then delivers nothing. APC's main limits are maintenance and trust, and models drift as equipment and feedstocks change. The benefit only lasts while the controller stays switched on, so its service factor is the metric to watch.
Advanced process control in real plants
- Aramco and Yokogawa Electric: Multiple coordinated autonomous control AI agents (FKDPP reinforcement learning), integrated with CENTUM VP, directly control and optimize acid gas removal operations at Aramco's Fadhili Gas Plant.
- TotalEnergies and Emerson: TotalEnergies will roll out Emerson's AspenTech Inmation data fabric across its industrial sites worldwide over a two-year period, together with Emerson advanced process control, as a base for AI use cases.
- Södra Cell and ABB: Södra Cell ordered ABB advanced process control for the bleaching process on one production line at its Mörrum pulp mill, after earlier APC projects at its Mönsterås and Värö mills.
- Covestro and Emerson: Covestro runs Emerson's DeltaV PredictPro model predictive control on one of four polyester reactors at its Barcelona plant to manage reflux flow in the esterification batch process.
- Tokuyama and ABB: Tokuyama runs ABB Ability Expert Optimizer, an advanced process control system, on the calciner, kiln and cooler of its Nanyo cement plant.
Vendors linked to Advanced process control
ABB (3) · Braincube (3) · Yokogawa Electric (3) · Emerson (2)
Related terms
DCS (distributed control system)
A DCS (distributed control system) is a plant control system in which control functions are spread across many autonomous controllers near the equipment, linked by a network to central operator stations.
Soft sensor
A soft sensor, or virtual sensor, is software that estimates a quantity that is hard or costly to measure directly by calculating it from other measurements that are available in real time.
Digital twin
A digital twin is a virtual model of a physical asset, process, or system that is kept in sync with its real counterpart through live data.
PLC (programmable logic controller)
A PLC (programmable logic controller) is a ruggedized industrial computer that reads inputs from sensors, runs a user program, and drives outputs such as motors and valves in real time.