Build, commission, and maintain multivariable MPC controllers using closed-loop time-domain identification - 2 days in the classroom or 18 hours online, from step testing and model building through commissioning and long-term maintenance.
The MPC700 course trains process control engineers, DCS technicians, and supervisors to build, commission, and maintain multivariable model predictive control (MPC) systems, using PiControl's COLUMBO software to identify and refresh dynamic models directly from closed-loop, time-domain plant data. MPC700 runs as 2 days of classroom training or 18 hours online, with training slides and MPC software from the first session.
The course starts from the history and case for MPC, where MPC outperforms single-loop and advanced regulatory control, and where a simpler methodology is the better fit. It then covers conducting step tests and identifying MPC models, designing and building the MPC controller, and startup and commissioning, before moving into MPC maintenance - modifying and improving models after years of operation or after significant process changes, including automated step testing, PRBS, and other modern identification techniques. Attendees finish able to design, maintain, and troubleshoot MPC controllers, and to use modern closed-loop dynamics identification with COLUMBO to improve MPC models without new open-loop step-testing campaigns.
MPC700 moves from the fundamentals of MPC through model identification, controller design, and commissioning, and into long-term maintenance - with hands-on work on COLUMBO throughout. After completing the course, attendees can design, build, and troubleshoot MPC controllers, decide when MPC is the right tool versus APC or single-loop control, and use modern closed-loop, time-domain identification to keep MPC models accurate for years after commissioning, not just at project handoff.
The course covers the following topics.
Once these topics are complete, engineers choose the delivery format that fits their schedule and plant access.
MPC700 is built for process control engineers, DCS technicians, and supervisors who design, run, or oversee multivariable MPC controllers. Attendees should have knowledge of primary process control and PID loops, and preferably a few months of plant experience, especially on a DCS, because the course moves quickly from MPC fundamentals into applied model building and commissioning.
Learn to conduct step tests, identify dynamic models, design the MPC control matrix, and build a new MPC system from the ground up, then keep it accurate with closed-loop identification long after commissioning.
Gain the skills to observe plant trends, recognize a degrading MPC controller, run step tests correctly, and troubleshoot the controller alongside operations and control engineers.
Learn when MPC is the right choice over APC or single-loop control, how to quantify the business case for a project, and what good commissioning and maintenance practice looks like before approving one.
MPC700 can be tailored for in-plant team training with plant-specific cases, using your own plant's data, models, and DCS platform for an applied learning experience. Group participation works well when a team is responsible for MPC commissioning, DCS migrations, or advanced control upgrades. Onsite corporate training is available on request so a team trains together on its own controller.
Hands-on practice separates a working MPC course from a lecture on theory. MPC700 trains engineers on COLUMBO, PiControl's closed-loop universal multivariable optimizer, which identifies and refreshes MPC dynamic models directly from closed-loop, time-domain historian data - including data with slave PID controllers in auto or cascade mode. That removes the need for a fresh open-loop step-testing campaign every time a model needs refitting, so attendees learn to build and maintain MPC models the way they will on the job.
The course also covers conventional step testing, the MPC control matrix, LP cost optimization, and commissioning practice, so attendees can build a new MPC system as well as maintain and improve an existing one. COLUMBO is vendor-independent, so the same closed-loop identification skills apply whether the controller in the plant is Aspen DMC, Honeywell RMPCT, Emerson PredictPro, Yokogawa PACE/SMOC, or Foxboro Aveva Connoisseur. Ongoing MPC health work after the course is covered in PiControl's MPC maintenance and improvement service.
MPC700 attendees receive a PiControl MPC700 Completion Certificate, which documents completion of the course for professional development records or a resume or LinkedIn profile. Attendees also receive documentation templates, commissioning checklists, and model validation sheets to reuse when building, commissioning, or maintaining MPC systems at their own plant. Engineers who want to keep building APC and MPC skills can continue toward broader MPC certification.
Course materials include training slides and MPC software, temporary software access, and follow-up guidance after the course. PiControl also offers APC200 for engineers who want a broader foundation in advanced process control alongside their MPC skills, and remote MPC maintenance and improvement consulting for plants that need a controller refit rather than, or in addition to, training.
Short answers to the questions engineers ask most before enrolling in the MPC700 Model Predictive Control course.
Request course info for MPC700 to give your engineering team a practical, COLUMBO-based Model Predictive Control course with a completion certificate. Online, 2-day classroom, and onsite formats are available, so teams in any location or time zone can start.
Related: COLUMBO MPC software, MPC maintenance and improvement consulting, and the full PiControl training catalog. Questions: [email protected], Tel: (832) 495 6436.