FLUID SIMULATIONS FOR A AIR RECIRULATED DATA CENTER- GREENHOUSE

(1)

FLUID SIMULATIONS FOR A AIR RECIRULATED DATA CENTER-

GREENHOUSE

Philip Rudén

Civilingenjör, Hållbar energiteknik 2021

Luleå tekniska universitet

Institutionen för teknikvetenskap och matematik

(2)

Förord

Detta projekt är ett examensarbete inom programmet Civilingenjör Hållbar Energiteknik, vid Luleå tekniska universitet. Detta examensarbete omfattar 30HP och är det sista delen som ska göras av utbildningen, innan det är dags att ge sig ut i arbetslivet.

Jag vill rikta ett stort tack till alla som har varit med och bidragit från projektets start till slut. Men framförallt vill jag rikta ett stort tack till Mattias Vesterlund på RISE, som har varit till stort stöd som min handledarre och examinator för projektet, men även varit den som har tagit fram examensarbetet. Jag vill även tacka Tim Leipold från Genesis mining som har hjälpt till med att svara på frågor om datacentret och även hjälpt till med att ta fram den data som krävts för att utföra beräkningar.

Från dessa fem år så är det främst något annat jag tar med mig än det som jag lärt mig under utbildningen.

Det är att från dessa fem år har jag fått lyckan med att jobba med klasskamrater som har kommit att bli vänner för livet och under dessa år har vi stått tillsammans genom ångest, press, lycka och svårigheter.

School’s out!

Philip Rudén 2021-05-27

(3)

Abstract

This is a master thesis project is provided by ICE datacenter, RISE. With the main objective to see if it possible to use waste heat from a data center to heat a greenhouse, but also to see if it possible to recirculate the air from the greenhouse back in to the data center. The prospected greenhouse has a surface area of 300m2 and is planned to be built during the spring of 2021 and heated by a 665kW data center. In this project several actors are involved. The data center of which heat is reused for the greenhouse comes from Genesis Mining. Humid air in both greenhouses and data center is a critical parameter. Where it can causes diseases for the plants and it can damage the equipment in the datacenter.

The simulation tools Ansys and IDA-ICE where used in order to calculate parameters such as the heat demand, mass flow, temperatures and velocity’s. First a model where build in IDA-ICE, due to that the mass flow calculated there is used as an boundary condition in the Ansys simulation. To complement this, hand calculations have been made on the RH-level in both the greenhouse and the datacenter have been made.

The results from the calculations and simulations shows that it is possible to build an air recirculated system, from the greenhouse back to the data center. The heat demand and the mass flow that are requierd for the greenhouse is lower than what the data center can provide. However, there might be some concerns for the RH-level in the data center. This due to that if the air from the greenhouse is directly mixed with the outside air, the RH in the data center will be higher than wanted. Therefore, there is necessary to take the right measures to prevent this and lower the RH. One solution is to dehumidify the air stream from the greenhouse before it enters the data center.

(4)

1 Introduction

This a master thesis is provided by ICE data center (RISE), that is an company for research on data centers in Luleå, Sweden. The thesis is a continue on the project Data center - greenhouse from the course Energiteknik Huvudkurs given by Luleå University of Technology fall semester 2020.

This projects main objective is to look on the relative humidity (RH) inside the greenhouse and if it is possible to heat and recycle the air from the greenhouse back in to the data center. This will be preformed on the greenhouse in the plug and play area and heated by a data center from Genesis mining. In order to do this, the simulation tools IDA-ICE and Ansys fluent will be used to make simulations on how the greenhouse parameters behave during the winter.

1.1 data center

The data center industry is growing fast, this is due to a increasing demand to store and process data. Data centers are buildings that house equipment for information technologies (IT), and is the equipment that is used to both process and store data. The IT works 24h a day, and therefore the workload for the equipment is high.

As the workload is high the equipment is getting hot and are in need if cooling. As the IT that are working 24h a day, it stands for about 3% [1] of the worlds electricity usages. However, the number will vary depending on which source that is used, but what all of them agree on is the believe that the total electricity consumption will increase. [1]

A data center have several components, but the most common ones are the switches, the storage unit, racks and the server it self. The server is the component that is used to analyze and store the data. The IT in the data center are the divided in to three different sectors, the IT-room, IT-support and supporting area, all of them housing one or several parts of the data center. The IT-room have the components and the cable that are connected to the computer, IT-support have the power equipment, switches and cooling and the supporting ares are for example the office. [1]

1.1.1 The data center in Boden

The data center that are located in Boden are provided and owned by the company Genesis mining. The total capacity of the data center is 665 kW (Email, Tim Liepold, 30 Sep 2020). The dimensions are 6.0x2.3x2.9 m and it houses a number of miners. (Tim Liepold, 16 Sep 2020) ( Email, Tim Liepold, 3 Feb 2021).

The temperature and volume flow that will leave the data center are going to vary during the year. In the project that where preformed during the fall semester 2020, an Microsoft excel file where handed out to the group from Genesis mining. This file complimented with other information can be use to estimate the outlet temperature and volume flow during both summer and winter. However, the two variables are expected to vary between 14-33 m³/sfor the volume flow and 25-56^◦Cfor the temperature (Tim Liepold, 20 Oct 2020). Where the lower digit are for the winter time.

1.2 Greenhouse

Greenhouses have existed for a very long time and to find the first one built, a long look back in time will be needed. The first accepted greenhouse was built during the roman empire era. It was during the rule under emperor Tiberius the engineers of the time constructed a greenhouse. The emperors favourite vegetable was cucumber and he always wanted them to be near hand. During the winter period and the cold nights the cucumbers where moved in a building with walls of glazed cloth or mica. This allowed the cucumber to stay in a warm environment even though it was cold outside. [2]

The first modern greenhouse was constructed during the 16:th century in Italy. The idea with this greenhouse was to show off the tropical plants, that where taken home during their exploring trips. This type of greenhouse where named botanic gardens and the concept where spread fast around Europe. The problem with these early modern greenhouses was to supply them with sufficient heat.[2]

(6)

In the beginning of the modern greenhouse era, only rich people and university’s could afford them. The university’s mainly built them in order to study tropical fruits and plant. However they also where used for other purposes. The British used them to conserve plants and the French used it to avoid their orange threes from freezing during winter times. From the 17th:century and on wards the development of glass and greenhouse construction continued. [2]

During the 19:th century the biggest greenhouses thus far where constructed. One example of them are the kew gardens in England. However, it also had other purposes than growing. Another example are the crystal palace in London. It was constructed by Joseph Paxton, who experimented with glass and iron construction in big greenhouses. Since the 19:th century the greenhouses have continued to be developed to today’s constrictions.

[2]

1.2.1 The greenhouse in Boden

The greenhouse and the data center are going to be placed in the old military area AF1 in Boden. The general idea of this greenhouse is to prove that it can be possible to use the waste heat from the data center, in order to heat up the 300 m²greenhouse (10 x 30 m). The greenhouse roof will have a curved roof with the highest point at 5.5m [3]. Unfortunately, the greenhouse have been delayed to be built during 2021 (instead of fall 2020 as planned). If the project schould be proven to be successful, there is plans to increase the size of the greenhouse in order to see if the data center could supply heat to a larger one.

1.3 Humidity

Humid air can be measured in several different ways such as RH, absolute humidity and vapour pressure deficit (VPD), are some examples. Where RH is the most commonly used when referring to the humidity inside greenhouses. The humidity is needed in both data centers for cooling purposes and in greenhouses to achive a good yield of the fruits/vegetables.

1.3.1 Datacenters

A high humidity can be good for cooling of datacenters. However, it can also cause problems such as arcing or leakage for example. Although, the switching gear have arc protection that can make them stand a high level of humidity. A high humidity level can lead to a lot of other problems as well such as corrosion. [4]

Ashrae have created a thermal guidelines for data centers depending on which type that are used, where the humidity is included. This guidelines have been compiled in a Table by Ashrae. Se Table 1 [4]

(7)

The data center i Boden, targets a operate between 67-75% RH which can be reach up to an RH of 80%.

The equipment is able to handle RH at about 90% as max, but only for a short time. ( Tim Liepold e-mail 2021-01-25)

1.3.2 Greenhouses

That the humidity affect the growth for plants and vegetables have been known for a long time. Where the humidity affects for example the stomal density and the leaf conductance of the plants. The humidity can also lead to diseases among the plants, affect the photosynthesis and the quality of the yield. Different plants and vegetables need different RH ranges to not affect the yield and the plants i negative ways. [5]

1.4 Air handling systems

The air speed in the ventilation system can be disturbing. This can be caused by fans or high air speeds inside the channels, when the air passes bends, area changes branches and valves. The general guidelines for the air speed inside the ventilation channels specifies that air speeds should be between 6-9 m/s in main channels and 4-6 m/s in smaller channels.[6]

The air speed inside in a residence zone schould not be higher than 0.15 m/s. However, higher speeds can be accepted when people move during work or when it is higher temperature inside [7]. The air handling system that are going to be used for the greenhouse is shown in Figure 1.

Figure 1: The air handling system for the greenhouse

1.5 Simulation tools

Simulation tools are often referring to software that could be used in order to predict what will happen in reality.

It is mostly used to predict how a model or product will behave in the real world for example, simulations of the aerodynamics of a car are common. The feedback from the results can then be used to improve the concept.

However, simulation tools cant with a 100% certainty give the right results, as the conditions in the simulation environment often are simplified from the real case.

1.5.1 IDA-ICE

IDA indoor climate energy or as it is shorten IDA-ICE, is a simulation tool from the Swedish company EQUA.

As the name tells, the software are used in order to simulate the indoor climate and energy use of buildings.

This can be done on locations all over the world as long as the weather data is acquirable, or that weather data for an region is known is imported to the program. [8]

(8)

1.5.2 Ansys

Ansys is a simulation software from the company Ansys. Ansys have been around since it foundation at the year of 1970. As Ansys been around for a long time their simulation software could be used for many different things. For example, analysis of finite elements, design optimisations and fluent simulations [9]. There are many different simulation sections in Ansys, and in this project the Fluent will be used.

1.6 Aim and objectives

The aim for this master thesis is to see if it is possible to build a system to use a primary heat source (the data center) to heat the greenhouse, where Boden municipality will grow vegetables for their own use. Where in the long point of view, the vegetables grown in the greenhouse will be enough to supply the municipalities demand.

This instead of importing vegetables from other places in Sweden or in the world.

The aim for this master thesis is to see if it is possible to build a system to use a primary heat source (the data center) to heat the greenhouse, where Boden municipality will grow vegetables for their own use.

1.6.1 Questions

Q1: What is the maximum energy demand for the greenhouse?

Q2: Will the humidity affect the greenhouse or data center in a negative perspective?

Q3: How large can the greenhouse be made on the plug and play area.

Q4: How big is the pressure drop in the ventilation channel?

1.6.2 Limitations

• Heat transfer inside the greenhouse between the crops and air will be neglected.

• Weather data for Luleå, Kallax is used instead of Boden, because that data is easier to obtain and more reliable.

• Simulations and calculations will only be based on data for cucumbers and tomato’s.

• The 60cm high brick wall around the greenhouse are not accounted for in the simulations.

• Simulations and calculations are only for showing the results during the coldest day.

(9)

2 Theory

In this section the theory needed for the project is presented.

2.1 Flow scheme

The idea is to take the hot air from the black box (data center) and transfer it to the green box (green house), and the flow scheme for the process are seen in figure 2.

Figure 2: Flow scheme of the process.

Ambient air will be taken in to the inlet, where it will be used to cool the IT. The hot air will then exit the data center and go to the greenhouse. At this point the data center itself can recirculate some of the air back, while the rest of it goes to the greenhouse. After the hot air have heated up the greenhouse it will leave it through its outlet. There some of it will be let out to the ambient air, while the rest will be recirculated back in to the data center.

As seen in the figure above, the air will mix in the data center where it will have 3 inlets. Here, all the parameters will mix. The mass flow, RH and the Temperature. The total mass flow intake for the DC will then be:

˙

m_in= ˙m_inlet+ ˙m_regh= m_out (1)

Where ˙mout are the mass flow out of the data center and ˙mreghis the recirculation from the green house back to the data center.

2.2 Heat balance greenhouse

All of the heat losses and gains for the data center is symbolized in the equation below. [10]

Q_heating+ Q_light+ Q_solar = Qtransmission+ Qinf iltration+ Q_vent+ Qtranspiration (2) Where the Qsolar, Qtransmission, Qinf litration and Qvent are the heat gain and losses from the greenhouse through the surrounding. Qtranspirationare the heat losses inside the greenhouse between the air and he crops.

[10]

2.3 Lightning

The effect of growing lights (and lights in general) are measured in the unit µmol/W . With that means the amount of µmol that is given out by the light per watt that is used by it. For growing lights the normal number is between 1.5-2.0 µmol/W and the higher value the light have, the better is it. [11]

Another common unit for the lights are µmol/m²/s[11]. If a light then ejects of 400 µmol/s and are projected to a 4m² area. The intensity of that light will be 100 µmol/m²/s.

The sum of light each day, or as it is called daily light integral (DLI) is measured in the unit mol per day. This can be calculated from the following equation. [11]

DLI = Intensity × 60t²

10⁶ (3)

(10)

Where t is the time in hours that the light is on and the Intensity are in µmol/m²/s.

2.4 Humidity

The RH-range that are suitable in a greenhouse are vartying. A high enough humidity is in favour for the yield and the growth of some plants, while controlling the system so the plants don’t get desiseas. Bakkers many experiment in [5], gave many different results on how humidity and humidity treatment affected the yield of the four different crops tested. For example, the yield of cucumber where higher at the higher humidity and the yield of tomato’s where higher at a lower humidity.

It was also discovered that a higher humidity level does favour the photosynthesis of the plant, but that does not necessary favour the production of the crops. [5]

When a closer look is taken on tomato’s and cucumbers from the experiments. It is clearly shown that they are affected in different ways. However there is some similarities as both of them have higher photosynthesis and leaf conduction with an higher humidity. The biggest and most important difference is that the yield is lower for the tomato’s at higher humidity (lower than the reference case), but for cucumbers is higher. The day and night humidity levels affect them different. The total leaf area for a cucumber plant is bigger with and higher humidity, while the leaf area for the tomato plant decrease [5]. Therefore, there is of much importance to have a good range of the RH inside the greenhouse in order to get a good combined yield of the different crops that are produced.

Plants go through different growing steps during its lifetime. The start with the early growth to then in the following order enter, vegetative, flowering, fruit formation and at the end mature fruiting. [12]

For tomato’s growing in greenhouses a RH between 60-90% is condemned to be good enough. However, studies show that the optimal range for the RH is between 50-70% for the entire growing process, where being around 60% RH seems to be the best. [12]

For cucumbers the target RH is 75-85%. However the plants can aslo survive in humidity levels up to 90-95%

RH. But, with and RH that is 90% or bigger, there is a higher risk of fungal diseases on the plant.[13]

2.5 Crop transpiration

The factor that is having the highest impact on increasing the RH level inside the greenhouse are the transpiration from the crops to the air. As the root of the plants takes up water and releases it from the leaf to the air.[5]

The transpiration rate from the crops are depending on several things, such as the air temperature and the RH level. The water content that leave the leafs is depending on the pressure gradient from the leaf to the surrounding air. Which in turn are depending on the VPD of the air. The humidity of the air is then increased due to the crop transpiration. But it also affect the transpiration from the crops.[5]

When the humidity and transpiration rate are at equilibrium, are when the transpiration rate is the same as the vapour contribution from the ventilation and the condensation. [5]

One way to calculate the moisture supplement at equilibrium is with the following equation m = mg+ Aµ

φ (4)

Where m and mg are the moisture content in the green house and the incoming air, in kg/m³. A is the floor area µ is the transpiration from the crops in kg/sm² and φ are the airflow in m³/s. (Personal communication Andreas Johansson, email 2021-03-29.)

The transpiration rate can be calculated with the following equation

µ = 0.207I_in+ 16.1∆_a (5)

where Iinare the horizontal solar radiation globally and ∆a are the vapour pressure deficit. [10]

(11)

2.6 Mixing of air streams

When two or several air flows are mixed the mass flows, temperature and RH from the different streams will combine and get new values. The mixing of two streams are presented in figure 3 below.

Figure 3: Mixing of two air streams.

As of in this case two streams will be mixed inside the greenhouse, due to that the datacenters own recirculation system probably wont be needed. Then to retrieve the desired RH for the data center several calculations step will be needed, where the first one is

x_c= L_ax_a+ L_bx_b

La+ Lb (6)

where x are the water content in the air for the diffrent streams a,b and c, and L are the mixing rate for the same streams.

With the x obtained for the mix, the RH-level of the for the air can be obtained with Φ = 100x

xs

(7) where Φ is the RH in % and xs is the saturation amount of vapour in the air, at an specific temperature.

To get the mixing temperature for the two stream, the entalphy of all three streams are needed. To get the entalphy for the two mixing streams the following equation is used

h = CpaT + x(hwa+ Cp,vapourT ) (8)

Where The Cp:s are the heat capacity for dry air and water vapour and hwaare the latent heat of water vapour.

To get the mixing enthalpy the following equation is used

hc= Laha+ Lbhb

La+ Lb (9)

2.7 Dehumidification

There are three diffrent methods to dehumidify the air. Two of them are by making the water condense with, either lowing the temperature or change the total pressure of the air. Both of these methods will make the water vapour in the air to condense, due to that the saturated x value will be lower. Which means that the air will be able to carry less amount of water vapour. The third way to lower the humidity level of the air, is to let the air flow over a desiccant. Then the difference in the vapour pressure will lead to that the moisture will be taken out from the air flow. [14]

When air is cooled so that the temperature drop under its dewpoint the vapour in the air will condensate on the closest surface. This will lead to that the air have been dehumidified by removing vapour from it. How much vapour that can be removed is depending on how much it can be cooled. The higher the temperature is lowed, the dryer the air will be. [14]

(12)

Desiccant dehumidifiers works diffrent than cooling humidifiers. Instead of lowering the vapour content by cooling the air, the vapour is taken out by a desiccant material. This is done by that the material creates an area with low vapour pressure, on the desiccant surface, where the pressure for the water in the air is higher.

This means that the vapour will be transferred from the air to the desiccant an the RH in the air will decrease.

[14]

2.8 Ansys

One important setup in simulations in Ansys are the viscosity model. Depending if the flow are laminar or turbulent different models can give different results. The most common models are the of the two equation type.

One of the most common used models are the k- model. This is a model transport equations for the kinetic energy (k) and the dissipation rate ().

(13)

3 Method

In the project start a literature study where preformed in order to get a better understanding of the areas connected to the project. In parallel with that data where collected, that could be used as the boundary conditions in the simulations. This was done with Genesis mining own monitoring system and to visit the data center to measure data.

3.1 Overall method

Figure 4 below shows the two overall method use in the simulation programs Ansys Fluent and IDA-ICE.

(a) Overall method in Ansys Fluent (b) Overall method in IDA-ICE.

Figure 4: An overall view of the methods used for the simulation in Ansys and in IDA-ICE.

The overall method for the Ansys simulation that are seen in Figure 4a are that the greenhouse and ventilation model are created in (NX) CAD. Then it is imported to Ansys and meshed. The boundary condition is set and the simulation are runed. Depending on the feedback from the results changes are made in the CAD model, mesh or the boundary conditions to make the simulation better.

The overall Method in IDA-ICE are ass seen in Figure 4b. The greenhouse model and zones are created, the settings are made. Simulations are made and from the feedback from the results, changes are made either in the model or in the settings.

(14)

3.2 IDA-ICE

In IDA-ICE a simplified model of the greenhouse was created. This due to that the curled roof was hard to recreate and therefore, the roof was assumed to have a constant angel, see Figure 5.

Figure 5: The simplified greenhouse in IDA-ICE, this greenhouse have a length of 30m.

As seen the greenhouse roof and walls are in glass (poly carbonate). These poly carbonate sheets are from the company Sunlite and therefore it properties are used in the simulations, see Table 2.

Table 2: Properties of the poly carbonate [15]

Solar heat gain coefficient 0.81 - Solar transmittance 0.79 -

Glazing U-value 2.9 W/m²K

The simulations on the model will be a study the heat demand to keep a desired inside temperature and humidity. In order to do this the simulation of heat demand will be made for an whole year. The set point for the temperature is between 20-25^◦C. This is due to that this is the temperature needed for growing cucumbers and tomato’s.

3.2.1 Solar

The heat gain from solar radiation will be encountered for by IDA-ICE itself. During the the visit to AF1, the orientation of the greenhouse where measured using a compass. This along with the correct weather data, will give a good estimation from the simulation results.

In the simulations weather data for Luleå, Kallax is used, from the year of 2020. This is the location that is closet to AF1 in Boden and the weather conditions for the locations are similar to each other.

3.2.2 Lightning

To get a more realistic simulation, growing lights need to be implemented in the model. As there is clear that Boden municipal want to grow vegetables, the assumption is made that the two things grown in the greenhouse will be cucumbers and tomato’s, and the lightning specifications are presented below.

For cucumbers during the plantation time, a light intensity of 250 µmol/m²/s, a DLI of 20 mol/day and light 20h/day is needed. During the winter those numbers are change to, a light intensity up to 300µmol/m²/s, DLI of 25 mol/day and 20-22 hours of lightning. [11]

For tomato’s during the plantation time, a light intensity of 200 µmol/m²/s, a DLI of 16 mol/day and light 18- 20h/day is needed. During the winter those numbers are change to, a light intensity between 250-300µmol/m²/s, DLI of 22-25 mol/day and 18h/day hours of lightning. [11]

For the winter time then, the lightning is set to be on for 19h/day as the tomatoes need 18 and the cumbers 20-22h. Then a intensity of the growing lights need to be set to somewhere between 250-300 µmol/m² .

(15)

3.2.3 Infiltration

The infiltration (Qinf litration) is also set in IDA-ICE. The setting is done in the infiltration part and the setting fixed infiltration are used. Here the value for how many times the air is changed inside the green house per hour i set.

3.2.4 Neglections

In order to simplify the greenhouse model, the heat and mass transfer from and to the crops are not taken i account (Qtranspiration= 0 in equation 2). The result will not be 100% correct, but the assumptions makes the model easier to create in IDA-ICE. Qtranspirationare not expected to be that high in comparison to other heat transfears.

Qvent from equation 2 is the heat losses from opening the windows for the greenhouse. As the simulation are gonna look on the heat demand of the greenhouse, the windows are assumed to be closed therefore, Qventis set to be 0.

3.2.5 Heating

The simulations will use a ideal heater to look at the heating demand. This ideal heater will be set to have the max capacity of the data center. When the heat demand is desired the required mass flow can be calculated with:

˙

m = Cpmean∆T

Q (10)

As the heat and temperatures are known and th Cp can be looked up in air-tables. The required mass flow for the for the peak demand can be determined and compared with the real one.

The simulations are done and redone for diffrent size of the greenhouses, the sizes are 30m and 60m on the length, and a width of 10m.

A length of 60m is set to be the max length and 30m are the length that are planned to be built. The reason 60m is set to be the max is that it probably will be harder to fit in a longer greenhouse than that. Then the width of the site is already very limited. Therefore the width of the greenhouse will be hold constant.

3.3 Ansys

The model used in this project are the same model that where used during the project in Energiteknik Huvud- kurs. The only difference is the size of the greenhouse, as is it of interest to see how large the greenhouse could be made. Therefore the model are set to a different length and that is done to the CAD file in Simens NX.

When the greenhouse length had been increased, put into Ansys and checked that it worked. After that the length of the ventilation socks is increased to match the same scale dimensions as in the real case. After that a propper simulation of the system can be done.

For the simulations the two equation turbulence model is choose, the k- model. This due to that there is risk of turbulence air flow inside the ventilation. Another reason this model i choosen is that is it one of the most commonly used for engineering simulations, as mentioned in the theory.

To check if the calculated mass flows from IDA-ICE is enough a simulation on the greenhouse size will be done with the determined mass flow and see if it can keep a desired temperature. As the only half of the greenhouse and the ventilation is simulated in Fluent the mass flow for the inlet boundary condition is set to be half, of the one obtained from the IDA-ICE simulation. The ventilation are split in to two identical part before entering the greenhouse. The mass flow is assumed to split equal between the both pipes.

The pictures for the results are taken in two planes that are inside the greenhouse. The planes were created at a height of 0.5 and 2.0 m above the floor.

(16)

3.4 Mixing

Calculations are made for the cases when the RH is at min and max both for the ambient and greenhouse air.

Where the max values are mixed and the min values are combined with each other. The assumption is made that all of the air from the greenhouse will be recirculated and that the greenhouse will have enough plants to increase the RH to a correct level. The own recirculation system for the data center are assumed to not be used in this situation.

As the temperature is known the saturated x-value can be taken from those temperatures from Tables [17].

Then as the values for RH levels for the greenhouse and the outside min and max RH is taken from SMHI Luleå Kallax measure station [18], the x can be obtained with equation 7. With the x-values the mixing x value could be determined. This is done with equation 6

To get the temperature of the mixed stream, first the entalphys of the to mixing streams need to be calculated with equation 8. Then the mixing entalphy is obtained with equation 9. Then going back to equation 8, the mixing temperature can be obtained. With the mixed temperature, the saturation x is obtained and then the new RH value.

Then calculations on dehumidification where done to see the new mixing temperature and RH. The assumption is made that the dehumidifier is cooling down the temperature for the recirculation stream and a, new mixing temperature is calculated. From that a new mass flow that is requierd to cool the data center can be obtained, where the assumption is made that the cooling power will be the same as measured in Boden. Another assumption is that the outlet will have the same temperature as measured in Boden. Then the RH for the mixing stream is calculated in order to place the min mixing values at 67% RH and the max mixing values at 75% RH.

3.5 Crop transpiration

The global solar radiation where taken for February 2020 at Luleå Kallax from [18]. Then the µ is calculated with a VPD taken from [19], which are for tomato’s. The level inside greenhouse is then calculated to be between 75 or 85% RH where the VPD is changed. The min and max value that where measured for the VPD where 0.4-2.1 kP a, during a short period in March. The assumption is made that the greenhouse will be in equilibrium, in order to make the calculations easier and so that equation 4 can be used.

(17)

4 Results

In this chapter the results for the project is presented.

4.1 IDA-ICE

For the heat demand simulation the coldest day of the year was 6th of February (of the year 2020) and gave following result, see Table 3

Length [m] Heat demand [kW ] Temperature [^◦C] Outside temperature [^◦C]

30 56.3 20.0 -27.7

60 104 20.0 -27.7

Table 3: Heat demand and inside temperature for the greenhouse at the 6th of Feb.

The only thing that are changed in the results from the simulations are the heat demand.

To receive this temperature and the heat demand the mass flow of air to the greenhouse that is required are presented in the Table 4. This mass flows are for an outlet temp of 45^◦Cand a outside temperature of -30^◦C.

Length [m] mass flow [kg/s] at -30 ^◦C

30 0.75

60 1.38

Table 4: Required mass flow for each case.

In Boden the measured mass flow was 6.79 kg/s at -19^◦C outdoor temperature.

The operative temperature for the whole year in the greenhouse is represented in the Figure 6

Figure 6: the operative temperature for the greenhouse zone during the whole year of 2020, the picture where the same for both simulations.

The temperature are between 20-25 ^◦C for the winter and early fall months. During the rest of the year the inside temperature will be higher.

(18)

The max and min humidity for the greenhouse is presented in Table 5.

Min % RH Max % RH

1.58 61.8

Table 5: The min and max humidity for the greenhouse, where min and max are the lowest and the highest value for the RH inside the greenhouse during the whole year where the min and man RH for the greenhouse are the same for both simulations.

This RH are for when there is no ventilation or plants in the greenhouse, due to the set up of the IDA-ICE model.

4.2 Ansys

From the simulations it is seen that the initial flow is turbulent, see Figure 7. Then as more iterations are done the flow will become laminar, see Figure 8. The pictures are taken for the 300 m² greenhouse, but the flow is acting similar in both simulations.

Figure 7: In the start of the simulation here at iteration 6, there is turbulent viscosity in some cells.

Figure 8: At iteration 98 and 99 the flow is laminar, due to no turbulent viscosity.

(19)

4.2.1 300m² greenhouse

Figures of the results have been taken from two planes inside the greenhouse, 0.5 and 2.0m above the floor level, wit a perspective from above.

The velocity’s in the two planes are shown in Figure 9 and 10

Figure 9: Velocity in a plane 0.5m above the floor.

Figure 10: Velocity in a plane 2m above the floor.

(20)

For the lower plan the max velocity is 0.941 m/s and it is in the bend of the ventilation channel. In the higher plane the highest speed is lower, 0.575 m/s and it is in the corners of the greenhouse. Otherwise the velocity in both cases are low.

The temperatures for the same plane are shown in Figure 11 and 12

Figure 11: Temperature in a plane 0.5m above the floor.

Figure 12: Temperature in a plane 2m above the floor.

(21)

The pressure drop for the ventilation channels are shown in Figure 13.

Figure 13: The total pressure in the pipe.

The pressure is shown as total pressure, which means that the atmosphere value have to be added to see the real pressure. But as this is only suppose to the pressure drop that is not necessary. In the ventilation channel it is seen that the highest drop is in the bends of the ventilation channel, and the longer out in the ventilation pipe the air reaches.

(22)

4.2.2 600m² greenhouse

The figures for the velocity, temperature and the pressure drop where taken for this case, this are the same figures as where taken for the 300 m² greenhouse. As the geometry’s only have been scaled up, the tendencies of the profiles are quite similar to the ones in the smaller greenhouse.

The velocity’s in the two planes are shown in Figure 14 and 15

Figure 14: Velocity in a plane 0.5m above the floor.

(23)

The highest velocity occurs in the lower plane, in the bend of the ventilation. The velocity is 1.69 m/s in that area. For the other profile the highest velocity is in the greenhouse corner. Otherwise they are relative low, compared to the max velocity.

The temperatures for the same plane are shown in Figure 16 and 17

Figure 16: Temperature in a plane 0.5m above the floor.

Figure 17: Temperature in a plane 2m above the floor.

As before the highest temperature is in the ventilation where the inlet temperature is set. Otherwise the temperature are around 293K (20^◦C) in the lower plane and 296K (23^◦C) in the higher plane.

(24)

The pressure drop for the ventilation channels are shown in Figure 18.

Figure 18: The total pressure in the pipe.

The pressure drop here is also presented as total pressure. The biggest drop occurs in the bends of the ventilation channel. The pressure drop is also higher the further out in the channel the air reaches from the inlet.

(25)

4.3 Mixing

When the two streams are mixed, the properties of the streams are presented in Table 6. The RH level for the greenhouse are set for when both cucumbers and tomato’s are being produced. For this calculations the RH-level in the greenhouse are the minimum and maximum level that is accepted for growing cucumbers and tomato’s, in this particular case.

Variable size

Mass flow recycled 0.75 (300m²) and 1.38 (600m²) kg/s

Temperature recycled 20^◦C

Outside temperature -30^◦C

RH outside min and max 30 and 95 %

RH greenhouse min and max 75 and 85 %

Desired RH for DC 67-75 %

Table 6: Known data, the mass flow are for when all air is recycled.

When the recycled stream is not dehumidified, the following RH-levels inside the datacenter is seen in Table 7.

Size m² min RH % max RH%

300, full recirculation 332 394 300, half recirculation 236 310 600, full recirculation 349 412 600,half recirculation 314 387

Table 7: RH inside the greenhouse, when the recirculation stream is not dehumidified.

As it is not possible to have a RH above 100%, the actual RH-levels in the Table when the RH is above 100%

is 100%. The rest of the water vapour in the air will condensate.

As the RH will be over 100% when the recirculation stream is not dehumidified, it will require dehumidification.

When then the air is dehumidified to a certain temperature. The mixing temperature (temperature inside the data center) for the mixing streams for all cases an all dehumidification cases are presented in Figure 19.

Figure 19: The graph displays the mixing temperature for the mixed stream, for the four cases when the recycled flow are max and half for both greenhouses. For when the 600 m² greenhouse with half flow and for the 300 m² greenhouse

with full flow, two of the points overlap so that the orange point is under the grey.

The amount of ambient air needed to mix with the recycled air, in order to still cooled the datacenter with the same cooling demand, which is 438 kW . The mass flow are seen in Figure 20. The cooling demand are based on data collected in Boden on the 14th of February.

(26)

Figure 20: The amount of ambient air needed to cool the datacenter for all four cases. This is to keep the cooling power the same when its was calculated from the measures in Boden.

the RH it is dehumidified to is shown in 21 and 22.

To get the desired RH-range in the data center the recirculation stream need to be dehumidified to the value that are seen in Figure 21 and 22. The RH of the stream that leave the greenhouse are needed to be lower in between the points in the Figures for the same temperature the stream is dehumidified to (5,10,15 and 20^◦C).

Then the RH for the greenhouse will be 67% for the min and 75 % for the max. The trend lines then represent the linear trend, and can be used to approximate the RH-level needed at other dehumidification temperatures.

(a) The RH if all flow is recirculated. (b) The RH if half of the flow is recirculated.

Figure 21: For the 300m² greenhouse when the recirculation flow is full (left picture) and when only half of the is recirculated. The RH for the recirculation stream is calculated in order to reach a RH in the datacenter for the min an

max value. For the mean a RH of 67% is reached and for the max it is 75%.

(a) The RH if all flow is recirculated. (b) The RH if half of the flow is recirculated.

Figure 22: For the 600m² greenhouse when the recirculation flow is full (left picture) and when only half of the is recirculated. The RH for the recirculation stream is calculated in order to reach a RH in the datacenter for the min an

max value. For the mean a RH of 67% is reached and for the max it is 75%.

(27)

4.4 Crop transpiration

To reach an desired RH level inside the greenhouse, if only tomato’s are grown. The VPD that is needed are seen in Table 8

Greenhouse VPD min [kP a] VPD max [kP a]

300m² 0.45 0.56

600m² 1.12 1.33

Table 8: Where the min VPD is to reach an RH of 75% and the max to reach 85%.

Calculations have been made when the dehumidification is the most and the least, for the recirculating stream.

that will say that the max VPD is when the maximum humidity streams are mixed and then crop transpiration take it to 85% RH. The VPD min is when the min humidity’s for the streams are increased to 75%.

(28)

5 Discussion

The results from IDA-ICE shows that is possible to have a greenhouse that is 60m long and 10m wide, from the Table 3 it is seen that the heat demand for that length is 104kW . As the heat capacity of the datacenter is 665kW , which is more then enough. As the space at the site at AF1 is limited the only the length of the greenhouse was increased, it will also be hard to fit a longer greenhouse.

The mass flow that are required are seen in Table 4. That the mass flow that are required to reach the heat demand are lower than the one measured in Boden. This is not surprising due to that the heat demand is lower then the heat capacity. Therefore, the mass flow of the datacenter will be enough to supply the greenhouse with heat. However, it has to been noted that the measured mass flow is for an outside temperature of -19 ^◦ C. When the outside temperature is lower the mass flow from the datacenter probably will also be lower. This is due to that it will require less air to cool the datacenter. If it is discovered that the mass flow will be to low for lower temperatures, a fan in the ventilation channel between the datacenter and the greenhouse can be installed to solve that problem.

The operative temperature inside the greenhouse are shown in Figure 6. The operative temperature will be the same during the year for both of the simulations. This is also due to that the heat demand is lower than the capacity. The ideal heater in IDA-ICE will be only providing enough heat to meat the set points that have been set (20-25^◦ C). If the temperature would e higher no heat is supplied and if it is lower the heater will supply enough heat to meet the minimum requirement. From the same Figure it is seen that heating is required from beginning of October to end of March on a 24 hours basis. During the rest of the year some heating is required alltought, it hard to tell during the summer months if the heating is requierd for the lowest temperatures. During the other months heating is probably requierd during the evening and nights, when they are cold.

The RH levels in the greenhouse where also the same for both simulations, see Table 5. It also seen that the RH level does not reach the requierd set points. However this results schould not be considered valid due to the neglections and how the simulation model was built up.

One of the reasons that the RH level is not reached is because how the heat demand is chosen to be simulated.

There is no air handling unit system (ventilation) that is used to look at the heat demand. Instead the easier method is chosen were an ideal heater with the heating capacity as input. When the simulations then are done there is no external air added or taken away from the greenhouse, the model will only look on how much heat to supply to try to reach the diffrent set points target. As the RH humidity is depending on the temperature, where the saturated data for x is taken at a specific temperature, equation 7. As then the temperature had a limit to be between 20-25 ^◦ C if possible, the indoor climate will not reach the state where both the right temperature level and RH is acquired.

There is also some neglections that will affect the value for the RH so it would have been higher. The simulations are not encountering for crop transpiration in the model The simulation where chosen to done with out them because it will simplify the model as the amount of vegetables that will be grown is unknown.

In order to get a more reasonable result on the RH from the IDA-ICE, simulations with a propper air handling unit instead of a ideal heater should be used. Then it will regulate the mass flow from the air handling unit to meet the set points. In this case the ambient that will have a humidity that will increase the humidity inside.

The air handling unit can then be set up to have diffrent in-parameters and regulate other conditions. For example, the humidity and temperature can be controlled at the same time, if the right air handling unit is choosen and set up. Also to get a even better results the crops schould be added to the simulation and, the last thing is to connect the air handling unit so that air is recycled.

The humidity that is calculated from the recirculation case, so that the mixing stream will have a RH at min and max level (when it enters the datacenter), which is 67 and 75%. However the recirculation stream will most likely will have the need to be dehumidified. If the humidity is not lowed, the RH after the two steams have mixed will be over 100%, see Table 7. When the RH is above 100%, it means that the x-value is bigger tan the saturated x-value, which means that the air is consisting of more water vapour than it is able to carry.

This is not possible, instead the water vapour will condensate inside the datacenter until the x is matching the saturated x-value. Which means that the RH of the mixed streams will be 100%. Condensation inside the datacenter is unwanted because the water can damage the equipment. Therefore, one of the mixing streams

(29)

When the recirculated stream is dehumidified it is possible to reach an RH of either 67 or 75%, in the datacenter.

To reach one of those humidity levels the RH needs lowered for the stream that is coming from the greenhouse, so it follows the trend lines in the Figures 21 and 22 depending on which flow that are used. When the min level are then mixed with the min outside RH, an inside RH of 67% is reached. The same goes for the max streams, but then the RH is 75%. This shows that it is possible to recycle the air from the greenhouse back in to the datacenter, during the conditions when the minimum RH levels is combined in order to reach the minimum RH for the data center, and when the maximum streams are combined to reach the max level for the data center.

This is also calculated when the outside temperature is set to be -30^◦C. That means that it cant for sure be said that the greenhouse stream always need to be dehumidified. For example it is not known if the stream need to be dehumidified (or increase the humidity), when the spring arrives.

What is seen in Figure 21 and 22 is that the more the recycled stream is lowed in temperature, the less the streams humidity have to be lowed from its initial value, the only case that says otherwise is in the max humidity in Figure 21b. This can be due to that there is a small flow coming from the greenhouse, compared to the data center inlet. The outside temperature and humidity is then much lower than the onces coming from the greenhouse. Then as seen in Figure 19 is that the 300 m² will reach the lowest mixing temperature in the data center. Which means that the air can carry lees vapour.

The recirculation flow can not be designed as a closed system. This due to that the mass flow will be to low to cool the datacenter and that if the system is closed, it will increase the humidity of the system to 100% RH.

Therefore air will at some point need to be let out in the atmosphere instead of being recycled. But as the mass flow just to keep the greenhouse heated is to low and will not be enough for cooling the datacenter, ambient air need to be taken in to the process. The air that then is needed to get the same outlet temperature with the same cooling demand is shown in Figure 2. Here the least ambient air is needed when the 600 m² greenhouse have full recirculation flow and most when the 300 m² greenhouse have half of the air recirculated. The other two cases requires almost the same is because the mass flow from the recirculation stream is almost the same.

The mass flow seen in the Figure 2 is the requierd mass flow of ambient air that need to be taken in by the datacenter in order to cool it. Therefore, a system is needed controlling that the right amount of air will enter the datacenter inlet. The total mass flow out of the datacenter will then be higher then the amount needed for the greenhouse. Where a system is needed to let out the overflow of air.

When the (in this case) colder outside air is mixed with the hotter recirculated air, the mixing temperature will be low. This is due to that much more ambient air is put in to the system. As then the outside air is much colder, than the recycled air. The mixing temperature will be low, as seen in Figure 19. The less of ambient air that is used in the mix, the closer the mixing temperature will be to the outside temperature.

In order to reach the desired RH-level inside the greenhouse the increase of humidity from the plants by crop transpiration have to be enough. The data that have been taken from a earlier experiment, shows that tomato’s have a VPD of 0.4-2.1 kP a and with a mean of 0.8, for some days in March. For thees numbers it is seen that it is possible to reach an RH of 75-85 % RH because, what is seen in Table 8 is that the VPD is in between what the tomato’s can give for one short period. However, all of the numbers in the Table are above the mean value that have been measured on the tomato’s.

Even that the crop transpiration seems to be enough to increase the humidity there are some concerns. Be- cause, VPD that comes from the tomato’s seems to fluctuate. Therefore, it is hard to tell exactly how much transpiration the plant will contribute with. One concern is that for the period of time that the mean value of the VPD is lower than the VPD that is needed to increase the RH to the desired level. Whit that means that the transpiration from the crops will not be enough to increase the RH to the right level. Another concern is that in the smaller greenhouse, the VPD is low, as seen in Table 8. Then as this is at the minimum range of tomato’s seems to contribute with. Which means that the transpiration from the crops can be to high and that can lead to an high of humidity level. When RH is not in a good range for the plants it can lead to several consequences, such as lower yield, smaller vegetables and diseases of the plants etc.

(30)

The mass flows that have been simulated by IDA-ICE are sufficient enough to keep the desired temperature inside the greenhouse. The mass flow calculated was to keep the inside temperature at 20 ^◦C, and the target is to have an inside temperature between 20-25 ^◦C. The temperature profiles for both greenhouses at the ventilation inlet height. In the Figure 11 and 16 it is then seen that the temperature here is at 293 K which corresponds to 20 ^◦C. However, the other two temperature profiles that are placed in a plane a bit higher in the greenhouse, shows a temperature around 297 and 298 K which corresponds to about 25 ^◦C see Figure 12 and 17. This schould not be a problem because this is not exceeding 25^◦C limit. However, the mass flow was set to be to have a uniform temperature of 20 ^◦C inside (with only a little deviation). The reason that the temperature is higher is laying in what is simulated. In the simulation time is set to be steady.

When the time setting is set to be steady, it means that the simulation is not accounting for time. Therefore, the simulation results shows the results for the steady state conditions. The heat is then entering the greenhouse and starts to get cooled by losses through the greenhouse shell and by the temperature difference inside the greenhouse. Therefore, as seen in the temperature profile Figures 11, 12, 16 and 17, several heat layers will be formed. As the hot air enter the greenhouse and starts to get colder, it will rise in height and push the colder air down, forming the diffrent heat layers. The steady state in Ansys shows that it will be a higher temperature than the mass flow is set to give. This is probably because the two models are set up diffrent in IDA-ICE and in Ansys. However, the around the mass flows that are presented in Table 4, are good enough to supply enough heat to the greenhouse to keep it at the right temperatures, during the winter. This solution can seen to be suitable for winter when it is colder for a long period of time.

Another critical parameter is the air speed inside the ventilation and greenhouse. To high of air speed in the ventilation can lead to disturbing noises for the workers. If the air speed then should be to high inside the greenhouse it can also be experienced for the workers, and to high air speeds can damage the plants. However, this will not become a problem. Because the velocities in the whole greenhouse and ventilation system will be to low. As seen in Figure 9 and 14 the max velocities are 0.94 and 1.69 m/s respectively. Both of these velocities are in the bend(s) in the ventilation system, which seems to be the critical point of the air speed in the system. The velocities in this bend are lower than the maximum recommended value from arbetsmiljöverket, so the potential noise from this are okay. For the other two velocities profiles that are seen in Figure 10 and 15, the velocities are under 1 m/s. However, there is velocity’s that are over the 0.15 m/s, this scould not be a problem as the greenhouse probably not go as a residence zone . Therefore, exclusions of the air speed can be made and it will be allowed to be higher. The velocity’s are then not that high that it will damage the plants or bother the workers.

The other critical parameter is the pressure drop in the ventilation. Because if the pressure drop is to high the ventilation wont work. The pressure drop for the smaller greenhouse are seen in Figure 13 is barley nothing and are the biggest in the bends, it because here the pressure loss is a one-time loss. The same goes for the bigger greenhouse, where the loses is biggest in the bends, see Figure 18. Here the pressure drop is bigger at the end of the ventilation sock compared to the smaller greenhouse, it s because the friction losses are bigger due to that the ventilation are longer. However in both cases the drop is to low to be of concern.

For the simulation, the viscosity model k- where choosen. This model favours cases when the flow is turbulent.

From Figure 7 it is seen that in the start of the viscosity is turbulent, so the viscosity model schould be the right one here. However, later in the simulation there will be now turbulent viscosity, see Figure 8. Therefore, it could be better to have another viscosity model in here. But as only one model can be chosen for the whole simulation, a turbulent model where chosen. This can affect the results, but schould not have a large impact.

This project are working towards several of the 17 Sustainable Development Goals. The ones that are affected in a positive way is the goals number 7,9,12,13 and 17. The work shows that it is possible to heat greenhouses with excess heat from data centers. Then it is possible to change from heat systems that are using a oil and pellets pan, and use cleaner energy and reduce the climate impact, which will lead to a sustainable industry.

Then on a social level, this information can be shared in order to show that it is is possible to use excess heat to heat greenhouses.

(31)

6 Conclusion

The results from IDA-ICE, Ansys and the calculations gives that it is very much possible to build both the smaller and the larger greenhouse. They will most likely not be needed to be heated up during the warmer months of the year, the datacenter are able to provide more then enough heat for the greenhouse. However, there is some concerns about the humidity. As it is probably going to be to high for the datacenter if it is not dehumidified, or crops that requires lower humidity is grown. The other concern of humidity is that the crop transpiration seems to fluctuate a lot and that can lead to to high or low RH for the greenhouse. However, the recommendation is given to proceed as planed with the project. Build the smaller greenhouse and make test runs on it and measure the parameters of interest.

Q1: The maximum heat demand for the greenhouse is 104 kW , that are seen in Table 4.

Q2: The humidity can be regulated so that it neither will affect both the data center or the greenhouse. If the right precautions are made.

Q3: Due to geographical reasons, the greenhouse limit is about 60 m. However, as seen in the max heat demand. It can be made larger if the terrain would have allowed it.

Q4: The pressure drop in the ventilation channel will not be of concern.

6.1 Continued work

• Set up an more advance model with the ventilation system in IDA-ICE, that can be used to controlling both the humidity and temperature level.

• Build and test the 300m² greenhouse, a recirculation system from the greenhouse to the data center.

• Make a more in dept analysis and calculations on the humidity for the mixing and inside the greenhouse.

For example, look on the humidity when it is warm outside.

(32)

References

[1] Pei Huang et al. “A review of data centers as prosumers in district energy systems: Renewable energy integration and waste heat reuse for district heating”. In: Applied Energy 258.January (2020). doi: https:

//doi.org/10.1016/j.apenergy.2019.114109.

[2] New world Encyclopedia. Retrived 2021-01-28. 2017. url: https://www.newworldencyclopedia.org/

entry/Greenhouse.

[3] Stina Mattsson. Driva växthus med spillvärme från datacenter. Retrieved 2020-01-29. 2020. url: https:

//bodenbusinesspark.com/sv/nyheter/95--driva-vaxthus-med-spillvarme-fran-datacenter. [4] Ashrae. Data Center Power Equipment Thermal Guidelines and Best Practices. Retrieved 2021-02-16.

2016. url: {https://www.ashrae.org/file\%20library/technical\%20resources/bookstore/

ashrae_tc0909_power_white_paper_22_june_2016_revised.pdforshorturl.at/qKSW1}.

[5] J.C.Bakker. Analysis of humidity effects on growth and production of glasshouse fruit vegetables. Retrived 2021-01-20. 1991. url: https://library.wur.nl/WebQuery/wurpubs/fulltext/206443.

[6] Catarina Warfvinge and Mats Dahlblom. Projektering av VVS-installationer. Lund: Studentlitteratur, 2010. isbn: 978-91-44-05561-9.

[7] Anna Varg. Arbetsplatsens utformning. 2020.

[8] EQUA. Retrieved 2021-01-27. 2020. url: https://www.equa.se/en/ida-ice.

[9] Ansys. Retrived 2021-01-27. 2020. url: https://www.ansys.com/about-ansys.

[10] Hampus Markeby Ljungqvis. Retrived 2021-02-04. 2021. doi: https://doi.org/10.1016/j.energy.

2020.119169.

[11] Karl Johan Bergstrand. Modern växthusbelysning. Retrieved 2021-02-15. 2015. doi: 978-91-576-8910-8.

[12] R.R. SHAMSHIRI et al. Review of optimum temperature, humidity, and vapour pressure deficit for micro- climate evaluation and control in greenhouse cultivation of tomato: a review. Retrieved 2021-02-18. 2018.

doi: 10.1515/intag-2017-0005.

[13] J.B.Parker et al. Greenhouse cucumber production. 2019 edition. 2019.

[14] Lewis G. Harriman III. The dehumidification handbook. USA: Munters Corporation, 2002. isbn: 0-9717887- 0-7.

[15] SUNLITE. Sunlite techical guide. Retrieved 2021-02-01. url: https : / / vicwest . com / wp - content / uploads/2019/07/SUNLITE_En_Technical_Guide_61353_Web_1.pdf.

[16] cmteknik. The led fixture. Retrieved 2021-02-15. url: https://www.cmteknik.se/files/ProductGroup/

3768/LED_FL_300_UK.PDF.

[17] Mohnsen Soleimani-Mohseni, Lars Bäckström, and Robert Eklund. Formelsamling i energiteknik. 2nd edition. 2018.

[18] SMHI. Ladda ner metrologiska observationer. Retrieved 2021-03-08. url: https : / / www . smhi . se / data / meteorologi / ladda - ner - meteorologiska - observationer # param = airtemperatureInstant , stations=all,stationid=162860.

[19] B.J.Bailey O.Jillet. “The effect of climate on tomato transpiration in greenhouses: measurements and models comparison”. In: Agricultural and Forest Meteorology, 58 (Aug. 1991). Retrieved 2021-04-01, pp. 43–

62.

FLUID SIMULATIONS FOR A AIR RECIRULATED DATA CENTER- GREENHOUSE