Control and operation strategies of wind-CAES

Not only the location of the storage matters, but also how the system is operated.

Although Huntorf and McIntosh CAES plant are notably larger in scale than the system developed in this thesis, their control strategies serve as valuable information for building the model. This is largely due to the fact that the CAES systems are operated in a different manner than the conventional gas turbines (Weber 1975, 336).

Instead of regulating the power output by altering the turbine inlet temperature, in CAES the air flow rate is varied to match the required power output, simultaneously

maintaining the expander inlet temperature at a constant value. This method allows greatly improved part-load heat consumption, leading to more efficient power regulation.

CAES plants are designed to operate with either constant volume or constant pressure compressed air storage. The latter is only applicable for hard-rock caverns, in which a head of water applied by aboveground reservoir may be used to maintain the cavern at a constant pressure. In salt caverns, such operation would not be possible, as the salt would be dissolved by the flow of the water. The constant volume approach is therefore common, and governs whether compressed air is introduced to the expander train with constant or variable inlet pressure. By placing a throttle valve between the storage and the expander train, the air can be withdrawn at fixed pressure independent of the pressure variations in the storage. When operating at partial load, the inlet pressure decreases with the mass flow, which in turn reduces the power output. Although losses are caused by throttling, the method is utilized in both Huntorf and McIntosh due to increment in expander efficiency (Kaiser 2015, 3).

If the variable inlet pressure method is applied, the variation in storage pressure reflects to the expander train inlet pressure, causing deportation from the design conditions. (Weber 1975, 336; Zhao et al. 2016, 1166; Succar & Williams 2008, 29) The information in literature regarding the control strategies of the existing CAES plants is scarce. In Huntorf, certain flexibility in operation is available, as after four hours of discharging at rated power the storage is able to provide air flow high enough to produce power at an exponentially decreasing level for over ten additional hours (EPRI 2003, 15-14). Although in diabatic systems the steadily decreasing storage pressure can be compensated with increased fuel consumption, in adiabatic systems such operation is not feasibly possible. Therefore, in order to simplify the operation, the developed model is assumed to only operate between the storage pressure boundary conditions. The 110 MW rated McIntosh CAES is designed to operate mainly under part load, providing most of its yearly generation expectedly at 60 MW. The daily, weekly and continuous operating cycles also presented by Goodson (1992, 4-8 & 4-11) are not useful in the case of variable renewable integration, as the scheduled cycles have the issue of predictability. For this reason,

literature information is sought to evaluate the ramp rates and start-up times of CAES systems. Three main parameters are considered in the evaluation: the time to reach synchronous speed, the time to reach the full load and the ramp rate once the system is connected to the grid. It must be noted that the majority of systems presented in literature are diabatic, in which the demand can be met more flexibly by varying natural gas consumption, inherently allowing a faster response to the load changes than the TES. The adiabatic CAES models in literature rarely consider ramping, even then presenting values for ramp rates from the opposite ends of the scale (Wolf et al.

2012, 195; Hartmann et al. 2012, 544). Therefore, the SMARTCAES concept is selected as the primary reference, and the ramp rates are conservatively selected in accordance with the gathered information as shown in Table 2. Emergency starts, allowing faster ramp up to the load, are excluded from the consideration.

Table 2. Selected discharging and charging start-up sequence target parameters. (Bailie 2015, 2 & 4;

EPRI 2004b, A-7 & A-9; Wolf et al. 2012, 195; Hartmann et al. 2012, 544)

To synchronous speed Maximum ramp rate

Discharging Charging Discharging Charging

Selected value 3.5 min 2.5 min ±20 %/min ±20 %/min

Understanding how the storage part of the system should be operated is not enough, as the operation is largely dependent on the wind generation. Several schemes, working almost identically in terms of storage utilization, have been proposed in literature. As the role of storage is clearly to control the mismatch between the demand and wind generation, the question is when and how the storage should be utilized. In Figure 21, slightly modified scheme introduced by Arsie et al. (2007, 5–

6) is shown. The authors present a system based on simple valuation method, in which power generated by wind is primarily used to satisfy the load and the surplus power can be delivered to compressors or sold to grid depending on the margin, defined here as the difference between the price of electricity and operational costs.

If the power above that provided by wind generation is required, it can be supplied by CAES or by the grid, or both.

Figure 21. Operation strategy of a wind-CAES system. (Adapted from Arsie et al. 2007, 5)

The strategy introduced above is selected to be studied in this thesis. Although several possibilities for optimisation exist within the strategy, those are not pursued in this thesis. For example, the relative valuation of storage pressure and thermal energy should be further considered, particularly with the electric resistance heating.

Taking the simple approach, four boundary conditions according the discharge potential of the system can be distinguished:

a. Compressed air storage should be maintained at above throttle pressure b. Compressed air storage should not exceed its design pressure

c. TES should be maintained at above a certain threshold temperature d. TES cannot exceed heat transfer fluid maximum operating temperature These criteria act as the basis of the system logic, which is introduced in Chapter 4.4.