New developments in substations for district heating

P. Gummerus , in Advanced District Heating and Cooling (DHC) Systems, 2016

Radiator system balancing

The return temperature from a radiator system could have the potential to be as low as the room temperature, to utilize the maximum amount of heat from the DH system ( Wollerstrand and Lauenburg, 2011). Historically, the return temperatures in DH systems are significantly much higher, even during the cold part of the year. This is the result of incorrect adjustment of radiator systems and the short circuit between supply and return pipes. In order to obtain low return temperature from a radiator system, a low circulation flow rate of heat carrier must be maintained, so that stratification of the temperature in the radiator is obtained.

A heat exchanger designed for the worst case scenario (in effect the smallest temperature difference between radiator system supply and return) will also be suitable for the largest temperature difference. Normally, the mechanism that limits the use of one heat exchanger for a wide flow range is the changeover from turbulent to laminar heat transfer. This is not the case in the radiator system scenario. A modern heat exchanger designed for the secondary temperatures of 80/60   °C and the primary temperatures of 100/63   °C will result in primary temperatures of 100/34.5   °C, when secondary temperatures are 80/30   °C. Normally the design conditions have a small duration of time. When the load level is halved, the grädigkeit (temperature difference in the cold end of the heat exchanger) is also halved.

With a properly adjusted radiator or ventilation system, a standardized modern heat exchanger dimensioned for the worst case scenario, would be able to supply the required heat to the building, independent of design temperatures (normal ranges).

Read full chapter

URL:

https://www.sciencedirect.com/science/article/pii/B9781782423744000100

Temperature optimization in district heating systems

P. Lauenburg , in Advanced District Heating and Cooling (DHC) Systems, 2016

11.2.3 Heat supply

Generally, the impact of network temperature on various sources of heat supply can be summarized for a lowered return and a lower supply temperature.

For a lowered return temperature:

●

Increased efficiency in boilers with flue gas condensation, regardless of whether the boiler is heat-only or part of a combined heat and power (CHP) plant

●

Increased electricity generation in CHP plants for a given heat load – if two or more condensers are used, which usually is the case

●

Increased amount of utilized heat from low-grade heat sources, such as surplus heat from industrial processes and abundant renewable sources, such as solar and geothermal heat. This includes cases where two or more heat pumps connected in series are used to raise the temperature of a heat source.

For a lowered supply temperature:

●

Increased electricity generation in CHP plants for a given heat load

●

Increased opportunities for utilization of low-grade heat sources, such as surplus heat from industrial processes and abundant renewable sources, such as solar and geothermal heat. This includes cases were heat pumps are used to raise the temperature of a heat source.

11.2.3.1 Flue gas condensation

If district heat is supplied by a heat-only-boiler, network temperature is of little importance. The advancement of flue gas condensing boilers has offered a significant improvement in boiler efficiency. It has also helped to put focus on the pursuit of lowered return temperatures because they constitute the lower limit for how much heat can be extracted from the flue gas. Flue gas condensation increases the boiler heat output with 10–30% depending on the fuel used. Boilers using fuels with higher moisture content, typically biofuels, gain more than, for instance boilers using natural gas. If the return temperature is lowered, the gain in heat output is also larger in the former case. The increase in heat output from a 5  °C lowered return temperature is estimated to be 1–5%.

11.2.3.2 Combined heat and power

The previous section is valid also for CHP plants whenever flue gas condensation is employed, which often is the case for biomass-fired CHP plants. One thing should, however, be kept in mind: since the flue gas condenser increases the heat supply, less heat has to be supplied by the steam turbine to the condenser, which leads to less electricity being generated.

The impact of network temperatures on the steam cycle is more complex. A lower supply temperature always will increase the power-to-heat ratio, i.e. more power can be produced from a given heat load. If the DH supply temperature is lowered, the turbine's condensing temperature, and consequently pressure, will be reduced, and the expansion ratio of the turbine will increase. Typical, modern, biomass-fired CHP plants employ two condensers, utilizing heat at two pressure levels. This way, more electricity can be generated. With this configuration, a reduced DH return temperature will lead to increased electricity generation in the low-pressure part of the steam turbine. The potential is, however, substantially bigger if the supply temperature can be lowered. The gain estimates are roughly a 2% increase of the power-to-heat ratio if the supply temperature is lowered by 5  °C (Falkvall and Nilsson, 2014; Johansson, 2011; Saarinen and Boman, 2012). Johansson (2011) gives a comprehensive review on distribution temperature influence on CHP plant performance.

11.2.3.3 Heat pumps

Large heat pumps generally do not have a large share of the DH supply. However, there is reason to believe that they will be more common in the future. Not least in combination with large shares of intermittent power generation, heat pumps supplying DH can contribute to the development of smart grids. As for CHP plants, the DH supply temperature has a great impact on efficiency, while the return temperature has some influence in certain configurations. Generally, a lowered supply temperature is always beneficial. The coefficient of performance (COP) will increase by typically 5% or more if the supply temperature is lowered by 5  °C (Falkvall and Nilsson, 2014; Selinder and Walletun, 2009; Zinko (ed.) et al., 2005). In order to achieve higher COPs, two (or more) heat pumps are often connected in series. Each heat pump then can operate with a smaller temperature difference. In such a case, the DH return temperature will also influence the overall COP. A lowered return temperature results in a lower condensing temperature for the first heat pump in the series.

11.2.3.4 Low-grade heat sources

There is growing interest in the DH industry for different kinds of low-grade heat sources. Without trying to define this somewhat vague concept, it is about utilizing, e.g. industrial surplus heat, solar heat, and geothermal heat, often with lower temperatures than used today in DH networks. It might also include customers that are sometimes generating heat, which is fed into the DH network (so-called prosumers), evoked by an interest among property owners to generate their own energy. In a truly smart energy system, not only will electricity be fed into the grid by end-users, but this will also be the case in thermal grids. It is quite evident that future temperature levels in DH systems are crucial for such a development. A study on the integration of prosumers is presented by Brand et al. (2014).

Read full chapter

URL:

https://www.sciencedirect.com/science/article/pii/B9781782423744000112

Thermal Efficiency Modeling in a Subtropical Data Center

N.M.S. Hassan , ... M.G. Rasul , in Thermofluid Modeling for Energy Efficiency Applications, 2016

2.4.2.3 Air Flow Paths

•

Supply and return temperature

Figure 2.13 shows the supply and return temperature contours in the data center. Return air path lines in Figure 2.13(a) show the entrainment of exhaust air through gaps between the servers. The maximum supply and return temperatures were found to be 18°C and 26°C respectively. Therefore, it is evident that path lines of return air raise the supply air temperature to a higher level than is desirable. As seen from the summary report in Table 2.1, the maximum inlet temperature to the racks is found to be 18°C. Table 2.2 shows the CRACs' performance report when all three CRACs were operating. The results shown in these two tables meet the ASHRAE guidelines [29,30]. Moreover, the inlet and exit temperatures at the racks were observed to be 18°C and 30°C, respectively. These temperature limits also meet the ASHRAE guidelines [29,30] which are shown in Table 2.3, and are also in good agreement with the experimental results. In addition, the low return temperature reduces the load on the CRACs and maintains the required supply temperatures [11].

Figure 2.13. (a) Return air temperature; (b) supply air temperature.

Table 2.1. Summary report

Model ID Nur-3rd Data Model
Mesh density Medium
Mesh count 28,627
Data center layout Raised floor with room return
Data center components
Room dimensions 8.05×7.08×2.96
Room floor area 56.99   m2
Supply plenum height 0.45   m
Number of CRAC units 3
Total flow through CRAC units 17.556   m3  s−1
Supply air cooling capacity 42,317.84   W
Number of rack rows 2
Total number of racks 10
Total rack heat load 41.12   kW
Total flow through racks 2.562   m3  s−1
Number tile rows 3
Total number of vent tile 15
Total flow through tiles 17.558   m3  s−1
Average flow through each tile vent 1.171   m3  s−1
Power
Rack heat source 41.12   kW
Wall heat transfer 1.15   kW
Total cooling load 42.28   kW
Heat density 741.8   W   m−2
Supply air cooling capacity 42,317.84   W
Temperature details
Highest temperature in data center room 29°C
Highest inlet temperature in racks 18°C
Flow through tiles
Maximum flow through a vent tile 2.433   m3  s−1
Minimum flow through a vent tile 0.092   m3  s−1

Table 2.2. CRAC performance report when three CRAC units are operating

CRAC unit name Average return temperature (°C) Average supply temperature (°C) Supply flow rate (m   s−1) Heat removal (W) Deviation from average heat removal a (%)
Down-flow units
crac_0 (A, 1) 18 15 5.852 24,650.14 74.75
crac_1 (A, 3) 17 15 5.852 14,935.51 5.88
crac_2 (A, 4) 16 16 5.852 2732.18 −80.63
Total – – 17.56 42,317.84 –
Average – – – 14,105.95 –
a
Deviation from mean value=[CRAC heat removal−ave CRAC heat removal]/ave CRAC heat removal.

Table 2.3. Rack cooling report

Rack row name Inlet temperature Average exit temperature (°C) Flow rate (m3  s−1) Power (kW) Cooling index a ASHRAE recommended inlet temperature range for class I servers
Minimum (°C) Maximum (°C) Average (°C) Recommended b (18–27°C) Allowable c (15–32°C)
rackrow_0_1_1 (D, 5) 16 18 17 29 0.225 3.46 0.839 Pass Pass
rackrow_0_2_1 (D, 6) 15 16 16 28 0.225 3.46 0.949 Pass Pass
rackrow_0_3_1 (D, 6) 15 16 16 28 0.225 3.46 0.942 Pass Pass
rackrow_0_4_1 (D, 7) 15 16 16 29 0.23 3.56 0.938 Pass Pass
rackrow_0_5_1 (D, 8) 16 16 16 29 0.225 3.46 0.935 Pass Pass
rackrow_1_1_1 (I, 5) 15 17 16 30 0.286 4.73 0.896 Pass Pass
rackrow_1_2_1 (I, 6) 15 16 16 29 0.286 4.73 0.957 Pass Pass
rackrow_1_3_1 (I, 6) 15 16 16 29 0.286 4.73 0.976 Pass Pass
rackrow_1_4_1 (I, 7) 15 16 15 29 0.29 4.82 0.986 Pass Pass
rackrow_1_5_1 (I, 8) 15 16 15 29 0.286 4.73 0.985 Pass Pass
Total – – – – 2.562 41.12 – – –
a
Rack cooling index=(ave CRAC supply temp/maximum rack inlet temp).
b
ASHRAE 2008 recommended inlet temperature range for class-I severs is 18–27°C.
c
ASHRAE 2008 allowable inlet temperature range for class-I servers is 15–32°C.
•

Air flow path in plenum

Figure 2.14 shows the air flow path lines in the plenum from the CRAC units. An enormous flowrate has been assigned to each of the CRACs in this study. Even in CRAC failure mode, when one of the CRACs is shut down, there is over three times too much air.

Figure 2.14. Air flow path in plenum.

Each of the CRACs has a supply opening that is 0.3×0.5=0.15   m2. Through each one, a flowrate of 5.852   m3  s−1 has been assigned as shown in Table 2.2. This means that the average velocity through each supply is 35   m   s−1. A parabolic profile is used to model the supply opening with the flow going to zero at the edges and about twice the average value (78   m   s−1) near the center. By zooming in on the inlets of the CRACs, it is seen that the maximum velocity only occurs near the center of the supply opening. The bulk of the flow coming from each CRAC has a velocity closer to about 35   m   s−1 or less.

The actual unit has the same flowrate but each supply should be 0.193   m2 so, with three of them, the total supply area is 0.579   m2. This would make the average velocity about 15   m   s−1 or less, which is more reasonable. As seen from Figure 2.14, it is evident that the average velocity of the air leaving from the CRACs is around 15   m   s−1 or less.

Read full chapter

URL:

https://www.sciencedirect.com/science/article/pii/B9780128023976000026

Services

Robin Kent , in Energy Management in Plastics Processing (Third Edition), 2018

Chilled water

The chilled water process pumps should be fitted with a VSD control to control either the ΔT or the pressure of the system.

One method is to measure the return temperature of the chilled water and to control the pump speed to maintain a constant return water temperature. This reduces energy use and improves the chilled water temperature stability.

•

Tip – Depending on the system there may be a need for a second control point to ensure that the system pressure does not fall below a set level, e.g., 2.5 or 4   bar.

•

Tip – For multiple pump systems, it is possible to use VSDs to control all the pumps. These will control the multiple pumps and even rotate pump use to give even use. As an alternative, it is possible to fit a VSD to only one pump, the fixed-speed pumps then provide the base load for the system and the VSD-controlled pump provides the variable load.

The second method is to control the pump speed from the system pressure to maintain a constant system pressure. This should be as low as possible to maximise the savings (see box on the left).

Read full chapter

URL:

https://www.sciencedirect.com/science/article/pii/B9780081025079500040

District cooling, current status and future trends

S. Tredinnick , G. Phetteplace , in Advanced District Heating and Cooling (DHC) Systems, 2016

8.6.2 Additional best practices for cooling coil selection criteria

Eliminate all by-passes, since they dilute the return temperature lower. This includes shunts, three-way control valves, etc. A three-way valve at the end of the line to keep the piping cold or provide minimum flow can be replaced by a flow limiting valve that will only open once the system pressure exceeds a specific setpoint signifying that all coil valves are closing. The flow limiting valve would remain closed until actually needed; unlike a three-way valve, which would by-pass water all the time.

Provide robust control valve actuators. Typical commercial quality valve actuators are only good for 40 psid (276   kPa). Coils closer to the building or DCS plant distribution pumps will see greater pressure differentials, therefore, requiring higher shut-off pressures so the valves are not lifted off their seats due to higher system pressures and act as by-passes.

Read full chapter

URL:

https://www.sciencedirect.com/science/article/pii/B9781782423744000082

Advanced fibre-reinforced polymer (FRP) composite materials in bridge engineering: materials, properties and applications in bridge enclosures, reinforced and prestressed concrete beams and columns

Hollaway L.C. , in Advanced Fibre-Reinforced Polymer (FRP) Composites for Structural Applications, 2013

16.3.1 The influence of temperature on polymers

The influence of temperature on polymers can be separated into two effects:

•

Short-term

•

Long-term.

The short-term effect is generally physical and is reversible when the temperature returns to its original state, whereas the long-term effect is generally dominated by chemical change and is not reversible; this effect is referred to as ageing. As the temperature varies, all properties of the polymer will change; consequently, to fully characterise the temperature-dependent material, properties must be measured over a range of temperatures. To study one or more of the properties as a function of temperature, a thermal analyser (differential scanning calorimeter (DSC)) is used; it scans property change over a wide temperature range. Particular cases of the effects of temperature on polymers are:

1.

Their glass transition temperature T g and their melting point (Hollaway, 2008; Hansen and McDonald, 2007; Hollaway, 2010)

2.

Their thermal expansion (Hollaway, 2010; Hansen and McDonald, 2007)

3.

Their thermal conductivity (Hollaway, 2008)

4.

Their exposure to ultraviolet light, although this is not strictly a temperature property (Hollaway, 2010)

5.

Their resistance to fire (Mouritz and Gibson, 2007; Hollaway and Head, 2001).

Read full chapter

URL:

https://www.sciencedirect.com/science/article/pii/B9780857094186500168

Cooling Towers

Stephen Hall , in Branan's Rules of Thumb for Chemical Engineers (Fifth Edition), 2012

Rating Nomenclature

To size a tower, critical parameters include: circulating water flow rate, return temperature, supply temperature, wet-bulb temperature, and dry-bulb temperature (or relative humidity).

The wet-bulb temperature design criterion must be based on the location of the tower, and should consider the time of peak load demand. According to SPX, the manufacturer of Marley cooling towers, most industrial installations that are based on the wet-bulb temperatures that are exceeded no more than 5% of the total hours during a normal summer have given satisfactory results. This is because the hours during which the wet-bulb temperature exceeds the upper 5% level are seldom consecutive, and usually occur in periods of relatively short duration. The "flywheel" effect of the total water system is sufficient to carry through the above-average periods without detrimental results [3].

Cooling towers may be defined by their approach temperature (the difference between the cooled water and entering air web-bulb temperatures) and circulating flow rate. Typical approach temperatures are 5.5   °C to 8.5   °C (10   °F to 15   °F). Figure 10-4 shows how the approach temperature specification affects the tower size.

Figure 10-4. Effect of chosen temperature approach on tower size at fixed heat load, flow rate, and wet-bulb temperature [3].

Towers that are rated in "tons" of cooling capacity use one of two conversion factors. Generally, for process cooling applications, 1 ton   =   12,000   Btu/h. However, for towers that are part of an HVAC system, where the cooling water is used by a refrigeration unit, 1 ton   =   15,000   Btu/h – which accounts for the inefficiency of the cascaded system. Engineers should clarify with the manufacturer which conversion factor is used for a particular tower.

Read full chapter

URL:

https://www.sciencedirect.com/science/article/pii/B9780123877857000104

Stirling engine systems for small and micro combined heat and power (CHP) applications

J. Harrison , E. On , in Small and Micro Combined Heat and Power (CHP) Systems, 2011

8.7.2 System design

Although the micro-CHP unit is essentially a replacement for a gas boiler, it has peculiar characteristics which mean that it cannot be treated in exactly the same way either from a design, installation or operational point of view. It may be a 'drop-in' replacement as regards the home, but the designer, the installer and the householder need to understand a number of key issues if the system is to perform at its best and to avoid disappointment for the user. One of the key requirements of a micro-CHP system as discussed earlier is to maximise the number of run hours each year in order to generate as much valuable electricity as possible. Not only does this imply a more accurately designed system, it also means that the householder may need to become accustomed to a somewhat different control approach.

For example, optimum start controls which are designed to achieve comfort conditions regardless of outdoor air temperature, are regarded with some suspicion by householders who cannot understand why the heating system comes on at different times every day (as a function of outdoor air temperature) even though they have set what they believe to be a start time for the central heating. In other words they fail to distinguish between setting the programmer to turn on the heating system at, say, 07:00 and setting the programmer to make sure the house is at the desired temperature by 07:00. in the optimised start programmer the control algorithm calculates the required start time based on the thermal inertia of the building, the necessary temperature lift (i.e. difference between current and desired temperature) and the heat loss deduced from the outdoor air temperature.

Of course, it is perfectly possible to design a micro-CHP system to perform in largely the same way as a gas boiler; the majority of micro- CHP packages now incorporate some form of supplementary burner with a significantly higher thermal output than the prime mover, so it is quite simple to use this capacity to achieve the rapid heat up which householders are accustomed to. However, this results in sub-optimal operation in which the potential value of electricity generation is squandered simply to avoid control complexities or to avoid the need to accommodate a HWC in the system.

In the case of the Remeha product, domestic hot water is provided entirely by the supplementary burner, replicating the performance of a combi-boiler, but typically losing the opportunity to generate 1000 kWh of electricity annually, worth around £100.

The WhisperGen unit, although configured as an integrated system boiler linked to a conventional hot water cylinder, offers users the facility to control the supplementary burner (which is triggered by the rate of increase in flow temperature, a proxy for the ability of the system to meet the instantaneous thermal demand) by means of a timed delay. The longer the delay, the less heat is provided by the supplementary burner and the more is provided by the engine, resulting in longer run hours and enhanced electricity production. However, this enhanced electricity production is achieved at the expense of longer recovery times of the hot water in the HWC for example; this can be effectively overcome by incorporating a larger capacity HWC, but this has implications for cost, space and standing losses. As always, it is a matter of compromise between theoretical efficiency and pragmatism.

As with condensing boilers, the thermal efficiency of Stirling engine micro-CHP is constrained by the return temperature which should ideally be kept at a temperature low enough to induce condensation and thus recover latent heat from the condensate in the flue gases. However, unlike heat pumps the performance of these units is not significantly compromised by the need to deliver DHW at a relatively high temperature as they are easily capable of delivering hot water at similar temperatures to a gas boiler.*

Read full chapter

URL:

https://www.sciencedirect.com/science/article/pii/B9781845697952500088

Energy Management

Sameh A. Nada , in Comprehensive Energy Systems, 2018

5.16.7 Enhancement of Data Center Energy Management Using Cold Aisle Containment

As discussed in the previous sections, the existence of hot air recirculation and cold air bypass adversely affects RTI and SHI and data center performance and energy efficiency and management. Hot air recirculation and cold air bypass can be avoided using aisle partitions and containments to make a complete physical separation between the hot and cold aisles as shown in Figs. 9 and 11.

Recently, several CFD simulations and experimental works were conducted to study the effect of using aisle partitions and containments on data center performance and energy management. Fig. 70 gives the air temperature distribution at the rear and front of the terminal and middle racks of a rack row obtained in the CFD case study considering cold aisle containment on top of cold aisles. Comparison between this temperature distribution and the temperature distributions obtained without using top containment profess that installing containments on cold aisles eliminates the cold air bypass and hot air recirculation at the middle and terminal racks, respectively.

Fig. 70. Temperature distribution at the (A) first and (B) middle racks with roof top containment.

Fig. 71 compares the performance parameters SHI, RHI, and RTI of the data centers with and without top roof containment. The figure shows that installing cold aisle top containment enhances the performance parameters; where a reduction of RTI and SHI by 18–20% occurred, and their values become within the recommended values (RTI=110% and SHI=0.2). This is attributed to the fact that using top containments eliminates the existence of hot air recirculation and cold air bypass at the terminal and middle racks, respectively.

Fig. 71. Effect of adding roof top containment of the cold aisle of layout 1 on the return temperature index (RTI), supply heat index (SHI), and return heat index (RHI) parameter.

Fig. 72 compares the performance parameters RCI of the two data centers with and without roof top containment of cold aisle. The figure reveals that adding cold aisle roof top containment enhances RCI; where 8–15% enhancing ratios of RCIHI were noticed moving the value of RCIHI of the data center to be within the recommended range (about 95%). This enhancement is also attributed to the absence of hot air recirculation and cold aisle bypass with using roof top containments of the cold aisles.

Fig. 72. Effect of adding roof top containment of the cold aisle on return cooling index (RCI).

Experiments were also conducted using the scale physical model to study the effect of adding cold aisle partitions and containments as shown in Fig. 73 on the thermal performance of data centers.

Fig. 73. Comparison of three air distribution system. (A) Typical under floor air cooling system configuration, (B) typical configuration with aisle partition system, and (C) typical configuration with aisle containment system.

For the sake of this study, three groups of experiments were carried out for a data center without partitions or containments, a data center with cold aisle partitions, and data center with cold aisle containments, respectively, as shown in Fig. 73. In the group of experiments the data center power levels was varied in the 379–1898 W m−2 by step 380W m−2. The opening percentage of the perforated tiles was set at 25% and the powers of all servers of the racks and the speed of the fans of the servers were maintained at the same value to achieve uniform server power scheme configurations.

Figs. 74–78 gives the air temperature distribution around the racks (at front and rear of the rack) and the servers' surface temperature distribution for the different arrangements of using and not-using aisle partitions at different data center power densities. The distributions for the three arrangements are superimposed on the same figures for comparison purposes. For the three arrangements, air temperatures at the front and rear of the rack were recorded at different points along the rack height. The averages of these temperatures are used for the sake of comparison between the different configurations of using and not-using aisle partitions and containments.

Fig. 74. Comparison of the different configuration systems at 379 W m−2 power density: (A) temperature profile at front and rear of rack for and (B) server surface temperature distribution.

Fig. 75. Comparison of the different configuration systems at 759 W m−2 power density: (A) temperature profile at front and rear of rack for and (B) server temperature distribution.

Fig. 76. Comparison of the different configuration systems at 1139 W m−2 power density: (A) temperature profile at front and rear of rack for and (B) server temperature distribution.

Fig. 77. Comparison of the different configuration systems at 1518 W m−2 power density: (A) temperature profile at front and rear of rack for and (B) server temperature distribution.

Fig. 78. Comparison of the different configuration systems at 1898 W m−2 power density: (A) temperature profile at front and rear of rack for and (B) servers temperature distribution.

As discussed before, Figs. 74–78 show that for the three arrangements the air distribution around the racks and the server surface temperatures increase with the increase of the server power. The figures also show that for the three arrangements, there is a remarkable increase in the server's surface temperature with increasing its height from the floor reaching its maximum value at the top cabinet of the rack; server 4 (H=25 cm).

Figs. 74–78(A) show that using aisle partitions and aisle containments improve the rack intake average air temperature whatever the value of rack power. This can be attributed to the fact that the existence of aisle partitions as aisle containments prevents hot air from recirculating from the hot aisle to the cold aisle. For example, using aisle partitions and containment reduces the average air temperature at rack intake from 26.1 to 22.5°C at 1898 W m−2 showing an improvement of 13%. Accordingly, the surface temperature of rack servers is expected to be decreased by using cold aisle partitions and containment as shown in Figs. 74–78(B); for example, the surface temperature of the server located in the top cabinet of the rack decreased by 11% when the data center power density was 1898 W m−2. The effects of the aisle partitions and containments on the surface temperature of the server located at the bottom rack cabinet are negligible as the hot air recirculation has approximately no effect on the server located in the bottom cabinet.

Using cold aisle enclosures (cold aisle containments) to completely isolate the cold aisle from the hot aisle instead of aisle partitions further improves the rack intake temperature and the server surface temperature (see Figs. 74–78(B)); for example, the rack intake air average temperature drops from 26 to 22.1°C at a power density of 1898 W m−2 due to using cold aisle containments. The trend is the same for all power densities. This further improvement can be attributed to the fact that using cold aisle containments completely prevents the hot air recirculation and cold air bypass between the hot aisles and cold aisles. The figures also show that using cold aisle containments leads to an improvement of cooling of all the servers of the racks with an overall enhancement of about 15.5% and not only a cooling improvement of the top server as in the case of using aisle partitions.

Fig. 79 compares the data center performance and energy management as measured by SHI and RHI of the three arrangements of without using cold aisle partitions and containments, using cold aisle partitions, and using cold aisle containments. The comparison is given at different rack powers. As discussed before, the figure shows a remarkable decrease in SHI and remarkable increase of RHI occurs with increasing the data center power density, which indicates the improvements of the data center performance and energy management. The figure also shows that a small enhancement in the performance and energy management efficiency of the data center occurs after using cold aisle partition and the enhancement increases with increasing the data center power density; for example, SHI was enhanced by the range 13–62% when the data center power density varied in the range 379–1898 W m−2. Fig. 79 also shows that additional enhancement in the performance and energy management of a data center is obtained by using cold aisle containment/enclosure instead of aisle partitions. The figure shows that SHI is enhanced by 70% as a result of using cold aisles containments. The enhancement of the thermal performance of the data center due to using cold aisle enclosure increases with the increase in data center power density.

Fig. 79. Variations of return temperature index (RTI), supply heat index (SHI) with power density for the three air distribution system configurations.

Read full chapter

URL:

https://www.sciencedirect.com/science/article/pii/B9780128095973005320

Cooling towers

John Neller , Eur Ing Dennis A. Snow CEng, MIMechE, HonFSOE, HonFIPlantE, HonFIIPE , in Plant Engineer's Reference Book (Second Edition), 2002

21.4 Design requirements

The factors which should be applied at the design stage cover the water flow rate, the design wet bulb figure, the required return temperature at the design point, the cost of power and land, and the water analysis. Water flow is normally determined by the equipment that the cooling tower is serving (for example, heat exchangers). Process designers historically leave the cooling tower until last (it is, after all, the final heat sink). When water costs were negligible, this was acceptable, but with the increase in costs and, in certain cases, the restrictions on availability of water, this approach has had to be modified. Greater consideration must be given to the overall system. Experience in the last ten years has shown that economic optimization can lead to a more efficient cooling tower, with a corresponding drop in the cost of heat exchanger. This is particularly true in power generation and industrial processes.

Design wet bulbs can be determined from published meteorological data for the area concerned. The difficulty is deciding how to relate the annual coverage to the tower performance at any given time.

For some years it was common practice to quote three different figures, based on the tower's performance as a percentage of the year. For example. in air conditioning it could be shown that the tower would achieve its design for 95% of the year. Alternatively, a tower costing 15% less could obtain its design parameter for 85–90% of the year. Only the operator would know whether the 85–90% or less was acceptable, while the economists would welcome the saving of financial capital.

The tendency nowadays is to design direct for the three warmest months of the year or as specified by the requesting purchaser, or to meet the specified by the requirements of the plant supplier's necessity. The selection of a tower no longer gives the 3 wet bulb alternatives, as the selection of the final specification may influence the obtaining of an incorrectly suited tower. The economic argument therefore no longer enters into the selection process.

The frequent failures to achieve even the quoted reduced percentage figures led to a reappraisal, and current design is more accurate. In some respects this is also due to the improvement in pack designs, particularly in the European and American markets. However, it must be said again that in optimizing cooling tower selection the designer must be advised of all appropriate factors. Discussions with cooling tower designers at the outset can save time and money in the future.

Water quality is important, not only from an environmental point of view but also in relation to the type of packing to be specified. Analysis of the circulating water is simple to obtain, but it is very seldom offered to the cooling tower designer. The quality, or lack of it, will determine the type of pack to be used, the selection of structural materials and whether the tower should be induced or forced draught, counterflow or crossflow. Water treatment, in the shape of chemicals to control pH and to act as counter-corrosion agents or as biocides, all have a bearing on tower selection.

Modern film packs can be offered for a range of 'Total Suspended Solids' (TSS) levels in recirculating water i.e. typically using the most efficient pack designs. TSS concentrations should not exceed 50 mg/l.

Alternative film flow designs can be supplied for concentrations levels up to 100 mg/l and 180 mg/l. Obviously other factors may modify these parameters, but are good enough for a general rule. Over 180 mg/l splash pack designs would be required And such designs would be required and such designs are now based on plastic splash grids, rather than timber splash bars/laths.

The 'Legionella syndrome' has resulted in health authorities in the UK applying statutory regulations, which are directly reflected in terms of capital cost and tower material selection. To safeguard against this, responsible designers have already produced cooling tower designs which not only meet the regulations but anticipate future, more stringent, legislation.

The following list of information factors should be made available to any supplier so that discussions on the technical requirements can be carried out prior to optimization (see Appendices 21.1 and 21.2).

Read full chapter

URL:

https://www.sciencedirect.com/science/article/pii/B9780750644525500766