A model-based controller for thermo-hydraulic circuits with variable transport delays

The work, published in the journal Energies, addresses a control challenge that is common in laboratory testing environments and industrial processes: when a mixing valve and a variable-speed circulator operate in the same circuit, their actions are not independent. Changing one affects the other. At the same time, the physical distance between sensors and actuators introduces transport delays that vary with flow rate, making standard PID control increasingly inadequate as operating conditions change.

The proposed controller combines several established control techniques, including a Smith predictor configuration, actuator decoupling, feed-forward compensation, and flow-rate-dependent delay management, integrated into a coherent architecture designed around the physics of the system rather than around linearisation assumptions. The controller was first tested in a virtual environment modelled in OpenModelica, then implemented in the laboratory's real-time control software and validated experimentally. Under the test protocol used in the study, the model-based controller achieved a reduction in root mean square error of up to approximately 80% compared to the standard PID controller in use at the laboratory.

"The primary aim of the work was to experimentally assess the performance of the implementation of the previously patented control method," explains Anton Soppelsa, researcher at Eurac Research and lead author of the study. "A secondary aim was to explain the details of the implementation more clearly, avoiding patent-specific jargon and providing a more accessible and readable description for those interested in understanding why and how the controller works."

The architecture of the controller draws on components individually known in the control literature, but their combination follows an original logic. The controller consists of several components, (Smith predictor configuration, decoupling, feed-forward compensation scheme, flow-rate-dependent delay line, and flow-rate-dependent anticipation operator) all of which are theoretically known in the literature or relatively straightforward to formulate" says Soppelsa. "However, they are combined in a peculiar and original way to improve the performance of the controllers currently used in our laboratories. The paper also provides a physical interpretation of why the controller works. This differs significantly from the black-box approaches commonly used in the field of process control."

The method is applicable to a class of thermo-hydraulic circuits beyond the specific laboratory setup described in the paper, including applications in thermo-mechanical and chemical process regulation.

The original article is available here