Modelling and control 3 min
Why having data does not mean having control
Control requires relating measurements, disturbances and actions. Historical data must support response reconstruction and model evaluation on other data.
Control
We review signals, actuators, delays and constraints before defining an advanced process control application.
When it applies
We define the decision to automate and review whether the signals make its effect observable. We identify tests and acceptance conditions before proposing control logic.
A feasibility assessment and validation scope identifying gaps in instrumentation, data and integration.
Plant trials, equipment installation and control changes require a specific scope and authorisation. This description does not replace operating procedures.
Knowledge
Modelling and control 3 min
Control requires relating measurements, disturbances and actions. Historical data must support response reconstruction and model evaluation on other data.
Instrumentation and data 2 min
An operating decision is interpreted through the sequence of disturbance, response and action. Analysis must consider delays and signals that can be compared.
Modelling and control 2 min
Validation evaluates a model on data excluded from fitting and against a baseline. Usefulness for control also requires process conditions and tests.