Engineer-to-engineer writing on closed-loop system identification, PID tuning, APC and MPC maintenance, and control-loop monitoring — straight from the desks that tune the loops. No hype, just the control math that holds up on the DCS.
The step test isn't just disruptive — it quietly biases your models. Here's how closed-loop identification reads true process dynamics from data you're already collecting, and why the loops you thought were "fine" usually aren't.
The two objectives pull the controller in opposite directions. Decide which one the loop is actually for before you touch a gain.
When an AspenTech DMC or Honeywell RMPCT starts cycling, you rarely need a full rebuild. Closed-loop data tells you which sub-models moved.
Oscillation index, valve travel, setpoint crossings, time-in-auto — a field guide to the numbers behind a control-loop performance score.
Square waves on a flow loop, sawtooth on level — the shape of the oscillation tells you whether to tune the controller or fix the valve.
A plain-language walk through how PITOPS extracts transfer functions from normal operating data while every controller stays in automatic.
Flow, level, temperature, pressure, and a nasty integrating loop — the practice sequence we use before anyone touches a live DCS.
Multivariable predictive control without a dedicated APC platform. What the deployment looks like, step by step, on existing hardware.
How a Gulf Coast refiner re-tuned 40 interacting loops from closed-loop data alone, and what the production numbers did afterward.
A fast inner loop fixes more disturbances than a clever outer one. The order you tune in decides whether the cascade helps or fights you.
One technical email a month — system identification, PID tuning and APC, written for engineers who run the loops. No marketing.
Unsubscribe anytime · We never share your address.