Design considerations of the system

Understanding the operational requirements of the system, the remaining task is to find out the features which have potential to fill the requirements. Although CAES as a technology can be considered as rather mature, the need for improvements has been

pointed out, particularly considering the compression efficiency (IEA 2014a, 42).

This chapter bases its discussion on the selected design challenges presented by De Biasi (2009, 1), mainly focusing on compressor technology.

Perhaps the most important design selection is the number of compression and expansion stages, which greatly varies between the proposed systems, as shown by Appendix II. Possessing medium complexity and low technical risk, the systems with two compression and expansion stages are the most presented option, and were found the most advantageous in the ADELE-ING project (Zunft 2015, 7; Freund et al.

2012, 8; Freund & Moreau 2012, 9). Theoretically, the cycle efficiency increases with the number of stages, but the increment is smaller with every added stage (Zhang et al. (2013, 477). Compromise between complexity and efficiency is therefore required, without neglecting the costs induced by additional components (Buffa et al. 2012, 1053). Five compression stages have been considered the optimal design choice, if heat is exchanged efficiently (Grazzini & Milazzo 2008, 2002–

2003; SedighNejad et al. 2014, 10–11).

The selection is reflected directly in the compressor discharge temperatures. When assuming constant total pressure ratio, the more stages are included, the smaller the discharge temperature becomes. This, on the other hand, simplifies the selection of TES medium and lessens the challenges in material durability (Mei et al. 2015, 602).

The current turbocompressor technology enables outlet temperatures close to 400°C, but not without downsides (Galant et al. 2013b, 264; Wolf & Budt 2014, 159; De Biasi 2009, 2). High-temperature compressors exist in gas turbine applications but often in lower pressures, and the industrial compressor trains typically are intercooled (Zunft et al. 2006, 3). Unlike the conventional compressor technology, the compressors should minimize cooling, thus allowing the maximum possible heat to be recovered and stored (Barbour et al. 2015, 813). Although intercooling does not decrease the recovery efficiency, it reduces the thermodynamic value of stored energy (Grazzini & Milazzo 2008a, 2001). Overall, the effect balances itself, as the achieved reduction in compression work is compensated by the reduction of expansion work due to lower air temperature. Therefore, the introduction of intercooling to compressor train offers the simplest design option, but the need for

specialized compressors has been pointed out for decades (Kreid 1976, 5; Bullough et al. 2004, 5; Zunft et al. 2006, 3).

With intercooling come several issues. As the ambient air is able to carry a certain amount of water vapour, the decreased temperature causes the water to condense.

Typically this occurs at the aftercooler, for which reason water separators should be installed (Wolf et al. 2009, 4). Furthermore, the issues are highlighted in two particular locations of the system: the inlets of compressed air storage and cold TES tank. This is ultimately due to heat exchanger effectiveness, which during charging limits the amount of heat transferred from air to the heat transfer fluid and vice versa during discharging. As the heat exchangers cannot operate with an ideal effectiveness, excess heat is left to the process. One may notice a contradiction in the discussion – heat is as valuable asset as the compressed air, but still considered as excess. Still, it is of great importance to remove the excess heat from the air before storing it, especially in the smaller systems, as the density of air increases with temperature. For the purpose, several authors have selected to introduce auxiliary heat exchangers, which commonly employ water as the heat transfer fluid and reject the excess heat to ambient (Hobson et al. 1981, 12; Dötsch et al. 2012, 64; Jubeh &

Najjar 2012, 86; De Biasi 2009, 3). Luo et al. (2016, 598) also acknowledged the problem with utilization of excess heat, describing it as one of the prevailing design challenges in the discharging process. The problem is opposite – the excess heat now in the heat transfer fluid is transferred to cold TES tank, from which it is dissipated to ambient over time. In both cases, the excess heat can be managed in two ways: by either increasing the heat transfer in order to minimize the amount, or utilizing the heat elsewhere. Passive heat transfer enhancements such as twisted tape inserts increase heat transfer with a cost of increased pressure losses, and have been proposed to be used with CAES (Hasanpour et al. 2014, 56; Freund & Moreau 2012, 9).

Regardless of the number of stages, the variation in storage pressure during charging creates another challenge. The compressors are forced to adapt to this variation, which leads to off-design operation. In order to achieve the highest possible profitability, the compressors should yield their entire power along the whole

charging stage. Although such ideal components do not exist according to Grazzini

& Milazzo (2012, 463), the operating curve of the components, defining the relation between the compression ratio and mass flow rate, can be modified through different means. Related to this topic are the two phenomena, which limit the operating range of compressors. In Figure 22, a simplified performance map for a multi-stage centrifugal compressor is shown. The line on the left displays the surge limit, left of which the operation is unstable, often even harmful due to momentary reversal of the flow (Mannan 2004, 633). On the right side of the map, the capacity is limited to the overload line, right of which the compressor will choke, rapidly losing its ability to produce head.

Figure 22. Simplified centrifugal compressor performance map, showing the surge limit and the overload line. (Adapted from Boyce et al. 1983, 150).

With part-load operation, start-ups, and shutdowns reducing the mass flow rate of the compressors, surge is more essential to consider than choke. This means that the capacity has to be controlled to ensure that the flow through compressor is not reduced below a certain level at a specific head. The most commonly used methods for modifying the operating curve, comprising suction throttling, variable speed control and variable guide vanes (VGV) as summarized in Table 3, are available depending on the compressor and its driver (Boyce et al. 1983, 157).

Table 3. Compressor capacity control methods and their characteristics summarized. Suitability refers to compressor types. (Lüdtke 2004, 116–120)

Control method

Part-load

efficiency Efficiency influenced by Suitability Acts on Suction

throttle Low Non-required head dissipated in

throttle valve All All stages

VGV Medium Compressor only produces the

required head Limited Pertaining stage Variable

speed High Compressor only produces the

required head All All stages

As the use of fixed-speed compression train has been a prevalent practice in both the existing and the proposed systems, variable guide vanes have been widely adopted (Dresser-Rand 2015a, 11; Budt et al. 2012, 796; Buffa et al. 2013, 1053; Freund et al.

2012, 13; Kraft 2010; Baxter 2006, 73; Pollak 1994, A.4-25). By adjusting the swirl introduced to impellers, the mass flow can be controlled in a fast and efficient way, allowing load changes through shifting the operating point (Brasz 1996, 761; Xiao et al. 2007, 473; Kappis 2013, 102; Kim et al. 2014, 4099). Compared to variable speed control, the method offers notably less efficient performance at part-load, but allows the compressor to operate at considerably higher percentage of the nominal head at reduced mass flow (Brasz 1996, 761). In reality, the integration of variable guide vanes is influenced by the selection of the compressor type, namely with respect to the number of affected stator-rotor stages (Giampaolo 2010, 25; Lüdtke 2004, 120).

As Apros does not create differentiation between the compressor types, the selection is only conceptual and does not have to be addressed.