Real-time state of charge estimation in thermal storage vessels applied to a smart polygeneration grid

In thermal grids and district heating, thermal storage devices play an important role to manage energy demand. Additionally, in smart polygeneration grids, thermal energy storage devices are essential to achieve high flexibility in energy demand management at relatively low cost. In this scenario, accurate evaluation of state of charge of storage vessels based on available measurements is critical. The aim of this paper is to develop and compare three different models for state of charge estimation in stratified water tanks (discrete temperature measurements) and the related application in an experimental polygeneration grid with a real-time management tool. The first model is based on the empirical calculation of the state of charge considering the thermal power difference between generation and consumption, and afterwards correction based on measured temperatures. The second model is a mathematical approach considering a pre-defined temperature shape fitted with experimental data. The latter model is based on a 1-D physical approach using a multi-nodal method forced on the basis of the measured temperatures. The models were compared considering an experimental test performed in the polygeneration laboratory by the Thermochemical Power Group (TPG). As a result of the comparative analysis, the first model was selected for applications in complex poly generation grids, due to its good compromise between accuracy and computational effort. Several tests were carried out to demonstrate the performance of the empirical approach selected for the thermal storage model and the economic benefit related to the utilization of this vessel. The experimental plant, constituted by two different prime movers (a 100 kW microturbine and a 20 kW internal combustion engine) and a thermal storage tank, was able to demonstrate the performance of a real-time management tool. For this reason, special attention was devoted to the variable cost comparisons. The novelty of this work lies in the development of the real-time management tool coupled with a thermal storage model by considering the simplified modelling approach. This is an essential requisite for complex polygeneration grids including hundreds or thousands of prime movers and thermal storage devices. Additionally, it is important to state that in such cases the required real-time performance could be difficult to obtain. The results, produced with the innovative and flexible experimental rig, demonstrate the positive impact of thermal storage as well as the effective management performance of this quite simple dispatching approach. Another important novel aspect regards this experimental assessment considering both specific 3-h tests and extended conditions typical of a possible real application.