Harnessing the Power of the Sun
1D Fluid-Dynamic Study of a Molten Salts Thermal Energy Storage System
By Alberto Deponti - EnginSoft Francesco Castelletta - Eurotecnica
Eurotecnica, italy is an international engineering and contracting company covering a range of sectors from chemical products (notably melamine) to solar power and energy storage. With a core of more than 100 highly skilled employees, it has successfully carried out more than 130 projects all over the world.
A major new area for Eurotecnica is the storage of energy generated from solar power; an obvious characteristic of solar energy is that it is only available during daylight, which leads to a storage requirement more efficient than simply up-scaling standard battery technology. One solution that is often used is to store the thermal energy from the sun is in the form of a mixture of molten nitrates. Held in huge tanks, these molten salts offer a highly efficient manner by which thermal energy can be converted to electrical energy overnight or during periods of lower thermal load (bad weather for example). Heavily insulated, these tanks can hold a capacity of 10,000s of metric tons of salt at temperatures well above their melting point (typically around 131°C) for many days if required. The economic viability of large scale solar projects hinge on the reliability and performance of such systems. Thus, a thorough detailed design process is critical in order to ensure that a reliable system of appropriate scale is ultimately installed.
This paper offers an overview of part of a study, focusing on the investigation of different operating conditions of a molten salts thermal energy storage system. In particular, the emergency closure of a valve is studied at two different conditions, namely the beginning and the end of the cycle. The objective of such an exercise is to find the minimum valve closing time that guarantees the safety of the system, i.e. the minimum time for which the peak pressure is below the maximum allowable pressure for the system. The system is simulated by means of Flowmaster, which allows many different design iterations to be handled virtually, thereby offering a cost effective and robust means by which such a study can be conducted.
The system to be studied is composed by two tanks of about 15m height and 40m diameter. In each tank there is an immersed pump and a distribution torus. The two tanks are connected by a pipeline with two control valves and six heat exchangers along its length. Each valve is mounted close to a tank (Figure 1). During the day hot molten salt, warmed up indirectly by parabolic troughs via the six heat exchangers, is pumped from one tank to the other one. During the night molten salt is pumped back and releases the heat accumulated during the day through the heat exchangers. For clarity, the tank from which molten salt is pumped will be called Tank 1 and the tank in to which molten salt is pumped will be Tank 2.
Similarly, the valve near Tank 1 will be called Valve A and the valve near Tank 2, Valve B. In the present work the flow of molten salt from Tank 1 to Tank 2 is considered (the reverse flow being symmetrical) and the emergency closure of Valve B in two different operating conditions is studied. The system is studied at the beginning of the cycle when Tank 1 is full and Tank 2 empty and at the end of the cycle when Tank 1 is empty and Tank 2 full. In these simulations molten salt is at a temperature of 286°C and has a density of 1,907 kg/m3. Under these conditions the speed of propagation of a sound wave through the salt is approximately 1850 m/s. The high density and the high wave speed that characterize the molten salt have the potential to produce a severe pressure surge when Valve B closes. For this reason an accurate fluid-dynamic study is essential in order to safely diagnose and design out any serious safety issues. While Flowmaster is perfectly capable of simulating full conjugate heat transfer if required, the focus of this particular study is in the pressure surge phenomena and so heat transfer phenomena that occur in the system will not be considered here.
In Figure 1 the Flowmaster network used for modeling the molten salts thermal energy storage system is presented. Each component of the network is characterized by geometrical and performance data provided by the manufacturer. Since the heat transfer phenomena are being neglected for this investigation, each heat exchanger is modeled by means of a simple discrete pressure loss component (green rectangles in Figure 1). The distribution torus can be handled in a similar manner as its details will not significantly impact the transient response of this part of the network. The closure of Valve B is controlled by a simple tabular controller component (yellow component in Figure 1). Taking into account the elevation changes within the network and the pipe schedules and fittings specified, a maximum allowable pressure of 25.88bar is specified as the design criteria.
In order to evaluate the valve closure time that meets safety standards, two sets of parametric analysis were performed for the start of run and the end of run valve closure conditions. Such studies can be run in batch mode in Flowmaster via the 'Experiments' tab, which allows different combinations of input condition to be run automatically. The results of the two parametric analyses are presented in Figure 2. It can be noted that the maximum absolute pressure decreases significantly as valve closure time increases until about 20 seconds, after that, maximum absolute pressure decreases very slowly. The valve closing time to be used in the case of an emergency maneuver needs to be unique for the entire cycle and needs to guarantee a reasonable safety margin. A valve closure time of 20 seconds guarantees good safety margins for both start and end of run conditions.
In Figure 3 and in Figure 4 the detailed results of the simulations performed with a valve closure time of 20 seconds at the start and at the end of the cycle are presented. In particular the maximum pressure in the system, the pressure at the pump outlet and the mass flow rate at the pump outlet are presented together with the valve closure time. In both cases a strong pressure surge is established, with the largest one occurring in the end of run in the valve closure case. However, the pressure in the system never exceeds the maximum allowable value for the system with a 20 second closure.
The 1D CFD simulations performed with Flowmaster allowed the study of the detailed behavior of the system early in the design phase considering different operating conditions. Specifically, the present work allowed for the precise definition of emergency maneuvers that guarantee the safety of the system during the entire operating cycle. The precise definition of the valve closure time also allows for the identification of the appropriate motor to be used for maneuvering the control valve. This work demonstrates the importance of numerical simulation early in the design phase of a large plant in which absolute reliability is paramount.