DOI: 10.18462/iir.icr.2023.0339

Stefano FILIPPINI*^(a), Umberto MERLO^(a), Dario DEMURTAS^(a), Federico VOLONTÈ^(a) Luca MOLINAROLI^(b), Ennio MACCHI^(b)

^(a) LU-VE Group

Uboldo, 21040, Italy, [email protected]

^(b) Politecnico di Milano

Milano, 20156, Italy, [email protected]

*Corresponding author: [email protected]

ABSTRACT

In recent years, concerns about the health of the Earth have increased public awareness and attention to “green” solutions. Within this framework, we present a study to investigate an innovative CO2 gas cooler which can reduce primary energy consumption. In the refrigeration industry the use of CO2 as the working fluid is increasing continuously and the opportunity to use highly efficient gas coolers is crucial for sustainable plants. LU-VE, in cooperation with Politecnico di Milano, has developed a new range of CO2 gas coolers equipped with solar photovoltaic panels which can produce a large fraction of the energy required to power the electronically commutated (EC) fans of the gas cooler. We did an intensive theoretical study, supported by an experimental campaign, to define the optimum size of the solar panels and the bank of batteries. The system is equipped with an array of batteries and a full regulation system, maximizing the number of working hours using solar energy. The paper describes the new product and the numerical model capable of predicting both the electrical energy needed by the gas cooler and that produced by the PV panels. The experimental campaign for the model validation is also illustrated. The PV panels are an integral part of the CO2 gas cooler, creating a packing device connected both mechanically and electrically to simplify installation on the work site. Finally, a techno-economic analysis is performed, to evaluate the pay-back period in different geographic areas, optimizing the configuration in terms of PV panel area and battery bank size, also considering the possibility of supplying any surplus electricity directly to the grid.

Keywords: CO2 gas cooler, air heat exchangers, solar panels, energy saving

1.      INTRODUCTION

LU-VE has always been a pioneer in the development of green solutions, aiming to lower the energy consumption of its products. Recently, a strong sustainability policy has been defined. One of its main pillars is the energy consumption reduction of ventilated heat exchangers. For this reason, LU-VE decided to study the possibility of equipping CO2 gas coolers with solar panels, aiming to self-produce at least 50% of the yearly power consumption. To evaluate the best configuration from both technical and economical side, LU-VE started a Research Project together with Politecnico di Milano.

Figure 1 Example of a LU-VE single fan heat exchanger equipped with PV panels, switch board, inverter, and battery.

In this work, the load consists of a CO2 gas cooler produced by the company, whose electrical energy demand comes from its fans and the control system. This collaboration involved the analysis of the variables related to all the components (i.e., fan speed, PV energy produced, battery state of charge, etc.), the development of a predictive model and a techno-economic analysis. This assessment will be the starting point for the development of a model coded in MATLAB®, able to predict the energy exchanges between the components of the system. At the end, some consideration about the share of self-produced energy and the economic outcomes will be presented through a yearly analysis based on available meteorological data related to the selected location for the system installation. Many system configurations will be investigated by varying the number of PV panels and batteries, to find the best ones in terms of self-produced energy and cost-effectiveness.

2.      EXPERIMENTAL FACILITY

2.1.                Plant diagram and components

The first approach with the project consists of the understanding of its components and how all the data are measured and recorded. It is necessary to have a direct and constant communication with the system to have the most complete knowledge before making all the necessary analysis. This is feasible thanks to the instruments that are included in the installed components, which are enabled to internet connection that allows a smart interface between the system and the user. Figure 2 schematically represents the system analysed in the project, highlighting the main power flows involved; in particular, the inverter handles the power flows among all the considered components (i.e., PV panels, batteries, gas-cooler, electric network).

Figure 2 Layout of the experimental facility highlighting the involved power flows: taking the inverter as a reference, PV power and grid power can only enter; load power can only exit; battery power can both enter or exit.

• Source of energy, the adopted PV panels are based on the mono PERC technology with multi-busbars and half-cells. This combination of solutions ensures a higher fraction of captured sunlight and a reduction of losses, with an overall increase in efficiency, reaching values above 20%. In the test presented we have adopted 6 PV panels with the following features: size 1.04×1.76 m and nominal power 375 W each. The nominal peak power is therefore 2250 W, more than twice the peak load. [1-5]
• Energy storage system, comprising two batteries based on Li-ion technology (2400 Wh for each battery), is adopted to collect the excess energy produced by the PV panels. [6]
• Load is constituted by the four fans of the heat exchanger, a CO2 gas cooler made by a finned Their rotational speed is imposed by an internal proprietary control logic based on the ambient temperature. Regarding the software validation, both controller and simulation system have been set with two temperature bands in which fan speed, and power consumption, vary linearly with Tamb (control logic shown in chapter 3.1). The power absorbed by each fan has a nominal value of 230 W, while the controller, responsible for the fan regulation, requires a power of 110 W. The overall peak power absorbed by the heat exchanger is equal to 1030 W. It is worth noting that the benefit in terms of energy efficiency is related to both the adoption of a renewable energy source and efficient EC fans with low design rotational speed (450 rpm).
• Inverter has the role of power manager: every power flow of each component passes through it. It takes as input the power produced by the PV panels which is converted from DC to AC unless the power is stored in the Depending on the load demand, the inverter evaluates whether the PV panels and/or the batteries could supply the demand or if the intervention of the electric grid is also necessary. As this system is conceived, during the test there is no possibility of supplying the excess power produced by the PV panels to the grid, hence, when the batteries are fully charged, electricity produced by the PVs is wasted. This control philosophy was selected for this experimental campaign, but other alternatives are possible, as discussed in the second part of the paper.

Figure 3 The system installed at Politecnico di Milano: on the left, the CO2 gas cooler, and the PV panels, installed outside the laboratory; on the right, the electrical rack comprising the batteries and the inverter, located inside the lab. In the real application, all components are integrated with the CO2 gas cooler (see Figure 1)

2.2.                Data analysis

Many data from the system (e.g., currents, voltages and powers of each component, ambient temperature, insolation) are uploaded and collected on an online platform, making it possible for the user to look at the instantaneous values as well as access the past ones.

The consistency of measurements is verified by the power balance applied to the inverter:

                                               Eq. (2.1)

Net of some oscillations that can be interpreted as internal consumption of the components, during a reference week considering a nominal peak power from PV panels is 2250 W, the maximum deviation measured is lower than 50W (<2%), hence power balances can be deemed as closed. The results obtained with the data analysis also enables the calculation of the roundtrip efficiency of the energy storage system. It is analysed considering the ratio between the energy extracted from the battery during the discharging period and the energy provided to the battery during the charging period. Such energies are evaluated in an operational cycle starting and ending at the same State of Charge (starting SoC = 10%) resulting equal to 88%.

3.      SYSTEM SIMULATION

The goal is the development and tuning of a tool able to perform yearly analysis in which the performances of a real hybrid system are simulated. Once the code is run, it performs several simulations equal to all the possible combinations of PV panels and batteries. For each combination, a typical operating year is simulated, adopting the most recent one on an online database as meteorological reference and then evaluating all the power exchanges among the components.

3.1.                Load simulation

The yearly simulation of a CO2 refrigeration system involves multiple issues to be evaluated, from the variable load required by the gas cooler, passing through the variation of performance as a function of ambient temperature, finally to the definition of a control logic (of the entire cycle) in such a way that it is the most efficient. Many factors that would shift the focus of the study to other aspects. The analysis carried out in this project represents a general evaluation about the coupling between a renewable energy source and a heat exchanger¹, so, it was decided to adopt a simplified regulation procedure, disconnected from thermal performance, linking only the fan speed with ambient temperature.

the speed (equal for each fan) is defined by Eq. (3.1). In the yearly analysis has been identified a temperature range starting from the monthly average ambient temperature ± 5°C and Tmin/Tmax are the boundaries.

Eq. (3.1)

Once defined the fans speed the power absorbed is computed by Eq. 3.2:

                                                    Eq. (3.2)

3.2.                I-V method

The I-V method allows the maximum power point of a PV panel to be found, varying the voltage given the current generated by the PV panel. The solar cell can be described through an equivalent electrical circuit as described in Figure 4, so, Equation 3.3 defines the current generated by the PV cell.

                                                               Eq. (3.3)

¹ We considered a gas cooler in this paper, but our results could refer to other air cooled exchangers, say dry coolers or condensers

Figure 4: PV cell equivalent electrical circuit

To solve the equation characterizing the equivalent electrical circuit, hence, to carry out the PV power calculation three points of the I-V curve are required, given by the technical datasheet (i.e., short-circuit current (ISC), open-circuit voltage (VOC) and maximum power point (IMPP, VMPP). The code is able to find the terms IPV, I0 and n by solving the following system (Eq. 3.4):

Eq. (3.4)

Once the previous parameters (with subscript ref.) have been calculated, they must be updated with the instantaneous environmental conditions as in the Equations (3.5-3.8):

                                                 Eq. (3.5)

                                             Eq. (3.6)

                                        Eq. (3.7)

                                 Eq. (3.8)

The power is a function of V and I, but the MPP is not known in advance, so the idea is to vary the voltage between zero and VOC and evaluate the maximum power for every current: a matrix is obtained where the maximum of each row represents the MPP. The maximum power of a single PV panel is hence calculated as reported in Eq. 3.9, while the overall power generated by the system (PPV) is the product between the maximum power generated by the single panels (PMPP) and the number of PV panels installed.

                            Eq. (3.9)

3.3.                Simulation procedure

For every hour of the year, the difference between the mean power produced by the PV and the power demanded by the load is evaluated as:

                                                                                               Eq. (3.10)

Six main situations can be identified (Figure 5), where, depending on ΔP, the elements of the system will behave differently.

Figure 5: Main conditions in which the system could operate.

The self-production factor is calculated as the main output of the simulation (Eq. 3.11). It indicates how large the amount of energy demanded by the load is, which is produced by the PV panels coupled with the batteries.

                                                             Eq. (3.11)

3.4.                Techno-economic analysis

The technical and economic analysis of the system is based on the research of the best configuration among all the ones obtained by the simulation. Given the various rules concerning the connection of PV systems to the grid in different countries, four cases will be investigated:

• Case 0 considers just the heat exchanger, completely powered by the
• Case 1 considers the adoption of PV panels and batteries without selling the excess electricity to the grid, like the configuration installed at the Politecnico di Milano.
• Case 2 is identical to Case 1, with the difference that the sale of excess electricity is
• Case 3 refers to a system consisting of PV panels and inverter, without batteries, which is able to self-consume all the energy produced by dedicating the excess energy to the service to which the system is connected. An example of this kind of application is a supermarket, in which adoption of CO2 as refrigerant is more and more popular, and it needs a significant amount of energy to power all its internal systems.

Starting from the output of the system simulation, the two main outcomes found are the self-produced energy (Eq. 3.11) and the economic impact of the installation of PV panels and batteries. To analyse the latter, two parameters i.e., the Net Present Value (NPV, Eq. 3.12) and the Pay-Back Period (PBP, Eq. 3.13) are adopted.

                                                                         Eq. (3.12)

                                                                         Eq. (3.13)

The economic analysis has been developed adopting the following hypothesis.

The cost of the CO2 gas cooler is never considered in the initial investment because it does not change from case to case (comparative approach).

• The time frame is set to 25 years (lifetime of PV panels adopted).
• Cost of energy purchased from the grid: 0.256 €/kWh. [7, 8]
• Price of energy sold to the grid: 08 €/kWh. [7, 8]
• PV panel cost: 160 €/panel.
• Battery cost: 1000 €/battery.
• Inverter cost: calculated in its turn through a scale coefficient as in Eq 14:

where:

                                                                                      Eq. (3.14)

• Cinverter, ref is the reference cost of inverter, equal to 10000 €,
• PPV, ref is the reference peak power of the PV field, equal to 115

4.      RESULTS

4.1.                System simulation results

The system simulation results are focused both on the outcomes of the code dedicated to PV power calculation validation and the overall functioning of the code. First, a global examination of the code for PV power performances has been developed, in order to verify its reliability. Meteorological data used as input in simulations with I-V method comes from the weather station of Solar Tech Lab (Energy Department of Politecnico di Milano).

Figure 6: Comparison between PV power determined with I-V method and PV power measured by the system.

In comparing I-V and test results, it should be pointed out that I-V method used the meteorological data coming from the Solar Tech Lab of the Energy Department of Politecnico di Milano as input; it is located on the terrace on the top of the building, free of any shadow effect, while the tested PV panels were located in an area affected by shadow in some hours of the day. As highlighted in Figure 6, in February, the sun rises at very low altitude and in front of the installed system there are two high buildings, producing a penalization

due to shadows. Therefore, PIV > PPV in the morning while they are comparable in the afternoon when the influence of shadow is reduced, so, the accuracy of the estimation is higher in the period with a lower impact of the shadows. Actually, the difference between the overall energy balance obtained by I-V method and test was less than 5% by considering summer months where shadow effects were marginal.

4.2.                Techno-economic analysis for an installation in the Milan area

The aim of the techno-economic analysis is the research of the best configuration according to three criteria: the self-production, the net present value, the pay-back period. The analysis refers to the four-fan CO2 gas cooler used in the above-described test, but results can be extended to other gas coolers, by maintaining the same ratios between number of fans and number of PV panels ad batteries. The first result comes directly from the energy analysis because it is based only on the simulations performed by varying the number of PV panels and batteries, while the second and the third ones must take into consideration also the economic value attributed to energy.

Figure 7: Self-production for different number of PV panels and batteries adopted.

The label in Figure 7 characterized by a self-production of 67% indicates the system installed at Politecnico di Milano and the horizontal plane highlights the target of 50%. This criterion is more dependent on the number of PV panels than on the batteries: the target is already reached with 7 panels and no batteries, while starting from 5 panels every case with at least one battery satisfies the target. It should be underlined that even with a vast number of PV panels and batteries the self-production cannot be achieved and some energy from the grid is always required. Considering the limited number (six) of PV panels that can be mounted on the selected CO2 gas cooler, the addition of a second battery does not add a significant advantage. The combination of number of panels and batteries should also consider the possibility of creating a simple package configuration for installation of a commercial CO2 gas cooler.

Net present value is firstly evaluated between case 1 and case 0 to evaluate the benefit of the hybrid system, later the comparison between case 1, 2 and 3 is analysed for assessing the best strategy for the integration of PV panels and batteries in the plant.

Figure 8: Different NPV for Case 1 and Case 0.

From the chart reported above in Figure 8, all the combinations above the white plane (or above the dashed line in the parametric chart on the right) are convenient with respect to Case 1 meaning that in those situations it is better to install PV panels and batteries, rather than simply connect the CO2 gas cooler to the grid. Again, good values of NPV are obtained with 5/6 PV panels and 1/2 batteries. A larger number of PV panels and/or batteries would increase the investment with marginal energy saving during the whole lifetime of the system.

The pay-back period (PBP) is an economic indicator useful to determine the time required by the system to recover the initial investment, by means of the yearly money savings derived from the installation of the additional components. The following contour lines (figures 9) represent the PBPs for Case 1 and Case 2 for each configuration.

Figure 9: Variation of the Pay-back Period when changing the number of PV panels and size of battery banks.

The charts above indicates that the scenario which can sell the energy ensures faster return on the initial investment, at high values of PV panels, because the amount of energy excess that can be sold is higher. The same trend shown in case 2 is also presented in Figure 10 for the case 3, but with much lower PBP, since all energy produced is self-consumed.

Figure 10: Variation of the Pay-back Period when changing the number of PV panels in case 3

4.3.                Effect of the location

Another element that could influence the analysis is the installation location: moving to zones with different irradiation and ambient temperature, there could be different effects on the overall system, in particular the power production from PV will change significantly. For this purpose, a new location with higher irradiation has been selected: the choice fell on Catania, in Sicily.

Table 1: Economic outcomes obtained by changing the location.

Higher solar irradiance during the year ensures higher power generation by the same number of PV panels installed, so, talking about the PBPs the higher productivity in Sicily is translated into higher yearly savings, and so the investment is recovered sooner than in Milan.

5.      CONCLUSIONS

The aim of this project is the techno-economic evaluation of how a CO2 gas cooler, considered as an electric load, could improve its energy and economic performance when included in a hybrid solar system, instead of being simply connected to the grid. Knowing a typical yearly consumption curve is enough to potentially simulate every kind of configuration. The results achieved with the activities presented in the previous paragraphs are summarized below.

• The installed system is correctly operating, and the power balances are closed and coherent with what was expected.
• Thanks to the synchronization with the meteorological data from the Solar Tech Lab of the Energy Department of Politecnico di Milano, it is possible to verify the productivity of the installed PV panels and to compare it with the one forecasted by the code for the PV power evaluation.

• From an energy perspective, starting from 7 PV panels and zero batteries, or from 5 PV panels and at least one battery, every configuration shows a self-production higher than 50%.
• From an economic point of view, despite the initial investment for the additional components, any hybrid system (Case 1 to 3) is advantageous economically in terms of NPV. Moreover, the hybrid systems can show interesting performance in terms of pay-back period. This depends on the ability to produce renewable energy, reducing the dependency on the grid.
• Case 2 has better economics than Case 1, but it approaches Case 1 for small numbers of PV panels. However, Case 2 involves several bureaucratic aspects related to the sale of energy, which are avoided in Case 1, where the energy is just imported from the grid and never exported.
• Case 3 has better economics than Case 2, proving that the onsite self-consumption, which reduces the electricity import from the grid at high purchasing prices, is very advantageous.

Generally, a smart CO2 gas cooler has a good techno-economic potential. A realistic high number of PV panels (about one PV panel/fan) leads to satisfactory performance, while in contrast the number of batteries is limited; selling electricity is interesting, and onsite consumption is even more interesting.

ACKNOWLEDGEMENTS

Special acknowledgement is dedicated to the Energy Department of Politecnico di Milano, in particular to the professors and researchers who contributed with their knowledge and experience to the good success of the activity presented in this paper.

NOMENCLATURE

∆CFy                         Difference  in  yearly  cash

flows with respect to case 0 (€/y)

∆Ebattery                 Difference         in         battery

capacity between the end and the start simulation (Wh)

PBP                        Pay Back Period (y)

Pcontroller                          Power absorbed by the controller

CFt                              Yearly cash flows (€/y)                                  Pfan, nom                            Power absorbed by each fan at

nominal speed

Ebattery                      Energy       stored      in       the batteries (Wh)

Pgrid                                     Power extracted form electrical grid (W)

Eexcess                      Yearly energy excess (Wh)                          Pload                                    Overall Power absorbed by load

(W)

Eg                     Energy gap between valence and conduction band(eV)

Eload                           Yearly     energy     consumed (Wh)

PPV                        Power produced by PV panels (W)

RPM                      Fan speed

EPV                              Yearly energy produced (Wh)                    RPMnom                         Nominal fan speed

G                       Solar irradiance (W/m2)                               SoCbattery                       Battery state of charge (Wh)

REFERENCES

• Söderströma, , Papetb, P., Yao, Y., and Ufheilc, J., 2015. Smartwire connection technology. Meyer Burger.
• Green, M.A., 2015. The Passivated Emitter and Rear Cell (PERC): from conception to mass production. Solar Energy Materials & Solar Cells 143, pages 190–197.
• Blakers, A.W., Wang, A., Milne, A.M., Zhao, J., and Green, M.A., 1989. 22.8% efficient silicon solar cell. Phys. Lett. 55, pages 1363–1365.
• Green, M.A., 2009. The path to 25% silicon solar cell efficiency: History of silicon cell evolution. Progress in Photovoltaics: Research and Applications, 17:183–189, 5.
• Swanson, M., 2005. Approaching the 29% limit efficiency of silicon solar cells. Pages 889–894.
• Preger, Y., Barkholtz, H.M., Fresquez, A., Campbell, D.L., Juba, B.W., omàn-Kustas, J., Ferreira, S.R., and Chalamala, B., 2020. Degradation of Commercial Lithium-Ion Cells as a function of Chemistry and Cycling Conditions. Journal of The Electrochemical Society, 167:120532, 1.
• ARERA, 06-
• GSE, 06-
• it, 06-2022.