Heat pumping technology systems are getting ever more complex. And with them, their controller systems. This development means that more components must be controlled, also in various configurations and modes of operation. To secure this, there is a growing demand for precise and cost-efficient testing.
Testing could be done through software or prototyping. But software testing is too far from real operation, and prototype testing is expensive, and difficult to adapt to new criteria. The Controller-in-the-Loop (CIL) concept might be the solution to this. With this approach, a physical controller is used to control a virtual system. Specifically, a virtual (part)-model of the refrigeration circuit is mapped in a simulation tool and coupled with a real controller. With this set-up, the controller behaviour can be checked for correct function before the actual initial operation. The behaviour of the system and the control is investigated through simulation of different test scenarios.
With CIL, two different levels are considered: communication and simulation. At the communication level, data is transmitted via a connection between controller and computer. This connection could be either digital or analogue. At the simulation level, various tools might be used for simulation building and component simulation. The simulated model is then connected to the controller.
A demonstration of the CIL approach is carried out. It models a brine-water heat pump system used for domestic hot water and space heating. The control mechanism is provided from the hardware controller. After successful modelling, the virtual sensors and actuators, which are needed for the subsequent control, are defined as inputs and outputs of the model. And some errors were detected. One example is abruptly reduced volume flows, leading to a significant number of compressor starts in a short amount of time. This, in turn, leads to reduced compressor lifetime.
In conclusion, the CIL approach could very well be useful. And the technology should not be an obstacle: there is good availability of different protocols and coupling with a PC is quite uncomplicated. If the signals are analogue they can be converted into digital signals. Also, simulation programs are widely available. Thus, through coupling of a real controller with a virtual plant model, errors could be identified and corrected without prototypes and without the risk of damaging the plant.
Andreas Sporr, Austrian Institute of Technology GmbH
Michael Lauermann, Austrian Institute of Technology GmbH