Model predictive control of a phase-change-material thermal energy storage device at industrial relevant scale

• A PCM-TES device was dynamically characterized to understand how variations of heat transfer fluid mass flow affect its response. • It was observed that the state of charge of the TES device has a negligible influence on the response of the system. • Experimental data was used to design different controllers to be tested on the real equipment. • Despite its widespread adoption, a PID controller might require significant effort due to the strong non-linearities of the TES device. • MPCs guarantee better performance and easier implementation for PCM-TES devices characterized by a strongly non-linear behaviour. Thermal energy storage (TES) systems are widely used in the power generation, industrial and residential sectors, frequently coupled with concentrated solar power systems, heat pumps or dedicated heat exchangers to recover waste heat from industrial processes. They are characterized by high reliability, slow degradation and low costs, in terms of both investment and maintenance. Among the available technologies, latent heat storage systems employing phase change materials (PCMs) have the advantage of high compactness and small temperature variations. However, knowledge of operation of PCM devices is still limited, in particular regarding their dynamic response and performance of control systems devoted at regulating the thermal power exchanged. This article analyses the dynamic response of a shell-and-tube PCM-TES device, operated in laboratory but featuring industrial scale storage capacity (180 kWh) which provides cooling power exploiting the latent heat of water/ice. Experimental tests were carried out to observe its performance in various operating conditions and different states of charge, highlighting strongly non-linear behavior during transients, making the design of the controllers particularly challenging. First, a detailed system identification process was carried out, introducing new non-dimensional parameters to characterize the TES platform thermal response. Then a second order transfer function was developed to simulate the PCM-TES device and used to support the development of two control systems: the first one based on a conventional PID, and the second one developed according to a model predictive control (MPC) approach. These controllers were separately tested and compared in a software-in-the-loop setup and later installed on the actual PCM-TES device, demonstrating that (i) conventional linear control approaches might be unsuccessful with system non-linearities, causing instabilities, and that (ii) advanced control techniques, such as MPC, can compensate for system non-linearities and achieve successful regulation of the PCM-TES device.