An ad hoc control system for managing the dynamic behaviour of a Brayton-based Carnot battery

Carnot batteries (CBs) are considered a suitable grid-scale electricity storage technology to overcome the imbalance between power demand and production in a power grids with high renewable energy sources penetration. In the literature, several studies propose different configurations, according to thermal storage technology as well as the charge/discharge cycle. Several investigations focus on the thermodynamics and optimisation of these systems. However, the literature still lacks research dealing with CBs control systems and their dynamic behaviour during transient operations. For this reason, in this work, an ad hoc dynamic model of a Brayton-based Carnot battery, along with its control system, has been developed in an OpenModelica environment. The investigated configuration is the so-called Integrated Energy Storage System (I-ESS). During the charge phase, a high temperature air flow, heated by electric heaters, releases heat to a packed-bed thermal energy storage. The discharge system is basically a modified gas turbine in which the combustion chamber is replaced by thermal energy storage. This work focusses on the discharge phase of the system, which is more complex and intrinsically characterised by dynamic behaviour. To develop the CB’s control system, the authors started from the control system of a conventional gas turbine, which usually acts on two manipulated variables: the fuel mass flow rate and the compressor variable inlet guide vanes position (VIGV). These variables are devoted to the regulation of the shaft speed and the turbine inlet temperature. In CBs, on the other hand, the turbine inlet temperature is fixed by the storage temperature, and, since there is no fuel flow, the only manipulated variable remains the VIGV. Therefore, a dedicated control system is needed. The control system developed in this work consists of a proportional-integral-derivative (PI(D)) controller that receives as input the shaft rotational speed error and gives as output the signal for the VIGV regulation. The developed controller was validated and properly tuned to achieve good results in terms of time domain specifications such as overshooting, undershooting, and settling time. The optimal proportional gain, integral, and derivative time constants are 1.1, 3.5 and 1.1, respectively. The results show that, given a power step function from 60% to 100%, a 10 MW Carnot battery can achieve a power settling time of approximately 1.2 to 3.3 seconds, with an overshoot of 2.0 to 4.6%. The maximum shaft speed undershoot is 1.4%. According to the results, a properly tuned PID controller allows Brayton-based Carnot batteries to have a dynamic response similar to that of a conventional gas turbine, with the additional value that the control system layout of a CB is much simpler and, therefore, easier to manage.