Potential of Supermarket Refrigeration Systems for Grid Balancing by

Demand Response

Tommie M˚ansson

and York Ostermeyer

Architecture and Civil Engineering Dept., Chalmers University of Technology, Sven Hultins Gata 6, Gothenburg, Sweden

Keywords:

Supermarket, Demand Side Management, Demand Response, Energy Storage, Smart Grid.

Abstract:

The environmental goals of European Union demand a larger share of renewable energy sources for electrical

energy generation. With the increasing share of renewable energy sources such as solar and wind, the utility

grids has an increasing need for energy storage and/or demand side management. With a high energy intensity

and a large thermal inertia, the refrigeration systems of supermarkets appear as an attractive actor for demand

response in such scenario. Theoretically they have the capability to absorb vast amounts of electrical energy

as stored compressor work, lowering the temperature of the food goods in the refrigerators. Alternatively

supermarkets have the capability of reducing their energy demand by allowing the food goods temperature

increase to its upper limit, reducing electrical power demand for the grid. This positioning paper will further

discuss the attractiveness as and feasibility to use supermarkets for electrical energy balancing by demand

response in a smart grid.

1 BACKGROUND

The European Union (EU) Commission launched the

European Climate Change Programme in 2000 to de-

velop strategies to implement the Kyoto protocol in

EU. In short the programme aim is to reduce the

release of anthropogenic greenhouse gases (GHG)

to the atmosphere and thereby mild the effects of

global warming and climate change. The commit-

ment was further strengthened by the Paris agreement

from COP21 in 2015 where 195 UNFCCC members

agreed to limit the global warming to less than 2

◦

while aiming at 1.5

◦

C compared to pre-industrial lev-

els.

EU has set ambitious goals towards a more sus-

tainable future and by 2050 EU aims to reduce the

GHG emissions by 80 − 95% compared to 1990 lev-

els(European Commission, 2012). The EU is pur-

suing these goals through both ﬁnancial support and

revised regulations as will brieﬂy be introduced be-

low(Union and Action, 2017). Milestones for 2020 is

a 20% reduction in GHG, 20% reduced energy usage

and a 20% increased energy efﬁciency.

The ﬁnancial support is agreed as at least 20% of

the EU’s budget for 2014 to 2020 (e 180 billion) shall

https://orcid.org/0000-0003-1536-8484

https://orcid.org/0000-0001-7285-2737

be spent on actions helping to achieve the ambitious

goals for 2050(European Union, 2018). Additionally

individual countries and company initiatives are also

funding parallel actions and projects.

The introduction of EU emissions trading system

in 2005 that limits the maximum amount of GHG that

the EU countries is allowed to release has created

a platform for businesses to trade with their GHG-

reduction. Companies and countries can trade their

rights to release GHG, which becomes an incentive to

reduce emission to increase proﬁt.

Additionally, all EU countries are required to sup-

port renewable energy generation such as solar, wind

and biomass as a part of the Green energy targets.

Following the Revised Energy Directives (REDII) the

EU should have reached an overall share of 27% of

renewable energy by 2030, with some members per-

forming signiﬁcantly better(Comission, 2018). How-

ever, an increased share of uncontrollable and slower

responding renewable energy sources creates an issue

for the utility grid to balance the supply and demand.

In a traditional electrical grid, the supply is

adapted to the demand by adjusting the electrical en-

ergy generation. The variations are often balanced

by the use of smaller gas turbines, hydroelectric etc.

that has the possibility to adjust their power outtake

rapidly. However, gas turbines most commonly are

using fossil fuels and the hydroelectric capacity is

Månsson, T. and Ostermeyer, Y.

Potential of Supermarket Refrigeration Systems for Grid Balancing by Demand Response.

DOI: 10.5220/0007749401510156

In Proceedings of the 8th International Conference on Smart Cities and Green ICT Systems (SMARTGREENS 2019), pages 151-156

ISBN: 978-989-758-373-5

151

geographically limited to regions with a satisfying

topology.

Therefore in a national, continental or global sce-

nario where the share renewable energy sources are

increased and the use of fossil fuels are to be min-

imised, these problems of balancing will become

more severe. Implementing energy storage in com-

bination with demand side management to adapt the

demand to the available supply of energy is therefore

needed (Farhangi, 2010).

1.1 Needed Energy Storage and

Demand Response

In Germany, the implementation of renewable energy

sources has been progressing in accordance with the

goals of EU and the need of storage solutions in in-

creasing. In (Sinn, 2017) the author concludes a stor-

age capacity demand of 5, 800 GWh at 68% share of

renewable energy and 16,300 GWh with a share of

89% for the German context. In (Zerrahn et al., 2018)

the authors instead conclude a 55 GWh need at a 68%

share, and a 436 GWh need at 88% share. Already

today and since years back Germany is facing issues

with balancing the grid, resulting in negative electri-

cal energy tariff(Deloitte, 2015).

If batteries or pumped hydro is used, it has the

capability of storing electrical energy to be released

later, but it is costly and also the materials used in bat-

teries has a negative environmental impact(Sand´en,

2014).

By introducing demand side response, the

needed storage capacity will be lowered accord-

ingly(Balijepalli et al., 2011). Instead of demanding

energy when there is a deﬁcit, the energy user will

postpone the demand to a more favourable time for

the supply side.

(Haider et al., 2016) ,(Costanzo, 2015) , (Gelazan-

skas and Gamage, 2014) ,(Gelazanskas and Gamage,

2014) and (Baharlouei and Hashemi, 2013) amongst

others have discussed the beneﬁts and challenges with

demand response. They conclude one challenge to be

the communication between a vast amount of appli-

ances to schedule their loads optimally. By identify-

ing large and energy intense systems, the amount of

involved units can be reduces and this challenge is

thereby reduced.

This is where supermarkets and their refrigera-

tion systems appears as an attractive actor (M˚ansson

and Ostermeyer, 2013)(Funder,2015)(Pedersen et al.,

2014)(Hovgaard et al., 2011).

2 POTENTIAL DEMAND

RESPONSE BY

SUPERMARKETS

Supermarkets are inherently intense energy users due

to all food goods that demands refrigeration (Tassou

et al., 2011). The thermal inertia of the refrigeration

system is a great beneﬁt as it would allow the super-

market to “charge” by transforming electrical energy

into stored compressor work by temperature reduc-

tion of the food goods(M˚ansson, 2016). Addition-

ally the supermarket would be capable of reducing the

electrical power demand for refrigeration to nil within

seconds if requested by the supply side.

In Germany alone, there are about 38,000 super-

markets with an accumulated energy demand of 10

TWh (Funder, 2015), translating to an average of 30

kW of electrical power demand per market or 1.14

GW nationally. This represents about 2% of Ger-

manys electrical energy demand, which is also a rep-

resentative number for both Sweden (Arias, 2005)

and UK(Tassou et al., 2011). With refrigeration

representing approximately 50%(Statens Energimyn-

dighet, 2010) of a supermarkets energy demand it is

reasonable to estimate that the average power demand

would be correspondingly large. And most certainly

the rated power of the accumulated refrigeration sys-

tems is signiﬁcantly larger, which is advantageous

when acting as a energy sink for the grid.

2.1 Available Buffering Power and

Energy

A supermarket refrigeration system could potentially

run its refrigeration system at maximum rated capac-

ity at any given time. This action would result in all

refrigerated display cabinets decreasing in tempera-

ture at maximum rate. Several models for estimation

of this rate and the energy demand exists (Smale et al.,

2006) but are all dependent on accurate input param-

eters for the thermal properties of the food and refrig-

eration system. From the authors experience this rate

is in the magnitude of 1

◦

C/min. As the temperature

of the refrigerated display cabinets in general vary in

time with ±2

◦

C around the set point temperature, the

possible buffering capacity measured in time would

be 0− 4 minutes, depending on the individual tem-

perature levels of the cabinets in the supermarket.

Another scenario is to partially increase the com-

pressor work above the needed heat extraction rate for

the refrigerated display cabinets. This would result in

a slower decrease of temperatures in the cabinets i.e.

longer discharge time for the supply side. However,

SMARTGREENS 2019 - 8th International Conference on Smart Cities and Green ICT Systems

152

the energy buffered by the refrigerated display cab-

inets is equal for the maximum rate and partial rate

scenario. The energy storage capacity is limited to

the accepted lower limit temperature and the thermal

inertia of the food goods and refrigerated display cab-

inets.

2.2 Available Electrical Demand

Reduction

A supermarket could potentially also turn the com-

pressor completely or partially off, resulting in a tem-

perature increase of the food in the refrigerated dis-

play cabinets. The temperature increase rate is how-

ever lower than the temperature decrease rate. Fol-

lowing that the cabinets are developed to be energy

efﬁcient, the insulation capacity is high. From ﬁeld

observations and laboratory measurements an esti-

mated temperature increase rate between 0.2− 0.8

◦

per minute can be expected depending on the cabi-

net type and quality. For low temperature cabinets a

lower rate is likely.

With the same reasoning as above, this gives a

potential complete shut off time of 0 − 20 minutes

depending on the actual individual temperature and

quality of the cabinets. In a scenario where the su-

permarket just has been charged and all cabinets are

at their lower limit temperature and they are of highly

energy efﬁcient type, the upper ﬁgure of 40 min is

true. This ﬁgure is however also inﬂuenced by the

customer behaviours, i.e. if the doors are frequently

opened or not.

If instead of completely turning the compressors

off the supermarket keeps them at a reduced power,

the supermarket should be able to run at reduced ca-

pacity for signiﬁcantly longer times.

2.3 Example Store from Germany

To exemplify the potential, we present data from a

1,300 m

supermarket built in 2011 just outside Han-

nover in Germany. The refrigeration system serves a

total of 22 doored freezers, 15 chest type freezers and

62 doored medium temperature cabinets as well as 12

meter of refrigerated deli desk. The rated electrical

power for this system is approximately 50 kW.

In Figure 1 the electrical power demand by the

refrigeration system is shown. Here it can be seen

that the system uses on average about 25 kW over

the week. Only at a few times a week the demand

is above 30 kW and it never goes below 20 kW on the

15 min averaged data that is shown. The ﬂuctuations

in the low temperature compressors that causes the

daily peaks is a consequence of synchronised defrost

schedules for the low temperature refrigerated display

cabinets.

Figure 1: Stacked line graph showing one week of the 15

minute average electrical power demand by the 1,300 m

supermarket used in this example.

For medium temperature refrigerated display cab-

inets in the store, the setpoint temperature range is

6 ± 2

◦

C allowing for 4

◦

C variation of temperature.

For low temperature cabinets the set point is −20

◦

with the same allowed variations.

To estimate the storage potential, the temperature

decrease and increase rate must be known. The au-

thors therefore performed a test on the medium tem-

perature cabinets of the store to ﬁnd the maximum

temperature decrease rate. And a representative tem-

perature increase rate for 23

◦

C ambient conditions,

which is slightly higher than the actual store temper-

ature, making the results conservative. The results

from the experiment is presented bellow in Figure 2.

Figure 2: Temperature development in a refrigerated dis-

play cabinet during maximum cooling followed by no active

cooling. Results are from a pilot experiment by the authors.

In Figure 2 it can be seen that the temperature de-

crease from 7.7 to 6.3

◦

C occurs during approximately

100 seconds. Resulting in an temperature decrease

rate of 0.84

◦

C/min. In analogy with the decrease rate

the increase rate was calculated to be 0.24

◦

C/min.

2.3.1 Absorbing or Postponing Energy

From Figure 1 we see that the electrical power de-

mand at any given time is within the range of 20− 30

kW. Meaning that the remaining power to reach the

maximum power of 50 kW is 20 − 30 kW too. With

a temperature decrease rate of 0.84

◦

C/min, the mar-

ket would be able to absorb energy for a maximum of

Potential of Supermarket Refrigeration Systems for Grid Balancing by Demand Response

153

4.76 minutes resulting in 1.59 − 2.58 kWh of excess

energy stored.

If assuming that all components of the refrigera-

tion system is at their respective lower temperature

limit and that all cabinets have the same temperature

increase rate we can calculate the maximum post-

poned energy use. From the experiment presented

the authors concluded a temperature increase rate of

0.24

◦

C/min, meaning that it would take 16.7 min-

utes to increase the temperature by 4

◦

C. During this

time the supermarket would under normal conditions

(20 − 30 kW) have used 5.3 − 8 kWh, which is the

amount of energy that was shifted.

With the large number of refrigerated display cab-

inets it can be assumed that the distribution of their in-

dividualtemperature is evenlyspread within the upper

and lower limits. This implies that a complete shut-

down of the refrigeration system is only possible if

the refrigeration system actively has cooled down all

cabinets before the shutdown. Otherwise the warmer

cabinets will exceed the upper temperature limitation

and the food must be discarded. Running the system

on partial load, serving only the warmer cabinets is

however a more realistic scenario.

3 DISCUSSION

As shown in the example above and by the previ-

ous argumentation, supermarket refrigeration systems

could potentially be utilised for demand response in

the electrical grid. In the example the quantities of

stored energy is rather low while the available elec-

trical power is high and almost instantly accessible,

making the supermarkets attractive as a short time en-

ergy buffer for the grid. One missing part to realise

this today lay in the communication between the su-

permarkets and the electrical grid, both for the busi-

ness agreements and digital signals to allow the grid

to use the capacity.

In today’s supermarkets the main objective of

the control systems is to keep the food temperatures

within the temperature limits and to make the heat ex-

traction demand even for the compressors to be able

run at an energy efﬁcient level. Some system actively

plans the cooling cycles to avoid peak power events

to occur when the electrical tariff is high. Meaning

that the refrigeration system today already uses the

thermal inertia of the food goods, but for a different

purpose then demand response. But due to the low ac-

curacy of the control system the safety margins must

be high to ensure the food safety. Meaning that the

full range in allowed temperature variation is almost

never used but rather limited to a few degrees.

When designing a supermarket refrigeration sys-

tem they are made to be redundant to ensure that heat

extraction demand never is larger than the installed

capacity. This has led to that the supermarkets has

signiﬁcant amounts of spare cooling power capacity,

which most certainly would beneﬁt the electrical grid

to have as buffering capacity.

Following that the four largest supermarket com-

panies Aldi, Netto, Lidl and REWE represents 14 640

(Statista, 2013) of these stores, the sector becomes

easily addressable. Meaning that if the companies

ﬁnd incentivesto implement demand response, the ac-

cumulated effects would be very large.

With low proﬁt margins (2 − 3%)(Arias, 2005),

the supermarkets are prone to adapt cost reducing ac-

tions in their markets(Retail Forum for Sustainability,

2009). Potentially being offered by the electrical grid

companies to sell buffering capacities or beneﬁt of re-

duced energy prices might be an attractive incentive.

With the trade of Carbon-emission rights the super-

markets could claim to have lowered the GHG emis-

sions and therefore sell parts of its emission rights.

Creating a second revenue stream for the supermarket

companies.

From a technology implementation perspective,

the larger organisations are often using standardised

solutions which would mean that the demand re-

sponse technology that is necessary would have fewer

variations. Adapting the technology of one system

that is used for thousands of similar stores makes it

more attractive for developers to ﬁnd a business case.

The businesses case is however rather complex

and must involve beneﬁts for two actors, both the

supermarket and the energy producer. Fiscal and

economical incentives for increased energy efﬁciency

and lower climate impact for the energy producers are

obvious. But these beneﬁts must also be transferred to

the supermarkets to motivate the initial investment in

systems allowing for the demand response to be im-

plemented.

An interesting aspect is that the driving force that

makes supermarkets attractive for demand response is

their high energy intensity. Meaning that any energy

efﬁciency measures that lowers the energy demand

for refrigeration will negatively affect its capacity as a

resource for demand response. Yet it will lowerthe lo-

cal energy demand and thereby the energy bill. There-

fore, ﬁnding a balance between the incentives here is

crucial.

Another aspect of the utilisation of the refrigera-

tion system for demand response is the fact that the

thermal mass is highly valuable food. The accumu-

lated monetary value of the stored food in the cabinets

is monumental. This demands that the implemented

SMARTGREENS 2019 - 8th International Conference on Smart Cities and Green ICT Systems

154

technologies are safe against failures and the highest

priority must always be the food safety.

Depending on the business model of the super-

market companies and the incentives provided by the

electrical grid, the installation of additional cold ther-

mal storage for the refrigeration system might be ben-

eﬁcial(Ochieng et al., 2014). As presented earlier

the duration for which the refrigeration system can

be turned off or ran at maximum capacity is limited

to the thermal inertia of the food. If installing addi-

tional thermal energy storage units, the duration and

stored energy could be increased. Additionally de-

pending on on chosen technical solution for thermal

energy storage, this would also allow for complete

shutdown by allowing the valuable food goods to cool

directly from the thermal energy storage instead of in-

creasing its temperature. The optimal sizing of such a

storage depends on the energy demand by the super-

market and the business agreements with the utility

grid. However, the storage is one directional as the

stored compressor work cannot be converted back to

electricity. Meaning that there must be a balance be-

tween the supermarkets heat extraction and the stored

electrical energy overthe chosen storage period which

could be stretching from minutes to seasons.

4 CONCLUSION

With the above presented arguments and discussion

we conclude that utilising supermarket refrigeration

systems as a part of the grid balancing mechanism

is feasible. The technology is there and the incen-

tives could be created via the right business model.

In theory almost any centralised refrigeration control

system would be capable of providing this buffering

capacity to the grid, the main barrier is the business

models and control system accuracy and communica-

tion.

ACKNOWLEDGEMENTS

The authors would like to thank the main funding

body, Climate-KIC, providing not just the necessary

funds but an invaluable network and inspirational

community.

REFERENCES

Arias, J. (2005). Energy Usage in Supermarkets - Modelling

and Field Measurements. PhD thesis, Royal Institute

of Technology.

Baharlouei, Z. and Hashemi, M. (2013). Demand Side

Management challenges in smart grid: A review. In

2013 Smart Grid Conference, page 96, Tehran,Iran.

IEE.

Balijepalli, V. S. K. M., Khaparde, S. A., Shereef, R. M.,

and Pradhan, V. (2011). Review of Demand Response

under Smart Grid\rParadigm.pdf. IEEE PES Innova-

tive Smart Grid Technologies.

Comission, E. (2018). European Comission - Renewable

energy directive.

Costanzo, G. T. (2015). Demand Side Management In The

Smart Grid. PhD thesis, Technical University of Den-

mark.

Deloitte (2015). European energy market reform Country

proﬁle : Germany Contents. Technical report, Deloitte

Conseil, Zurich.

European Commission (2012). Roadmap 2050. Policy,

001(April):1–9.

European Union (2018). Energy in ﬁgures – Statistical

pocketbook. Publications Ofﬁce of the European

Union.

Farhangi, H. (2010). The path of the smart grid. IEEE

Power and Energy Magazine, 8(1):18–28.

Funder, T. (2015). Supermarkets as an Important Smart

Grid Application. In 16th European Conference, Tech-

nological Innovations In Refrigeration And In Air

Conditioning, Milan. Danfoss.

Gelazanskas, L. and Gamage, K. A. A. (2014). Demand

side management in smart grid : A review and pro-

posals for future direction. Sustainable Cities and So-

ciety, 11:22–30.

Haider, H. T., See, O. H., and Elmenreich, W. (2016). A

review of residential demand response of smart grid.

Renewable and Sustainable Energy Reviews, 59:166–

178.

Hovgaard, T. G., Halvgaard, R., Larsen, L. F. S., and

J{\o}rgensen, J. B. (2011). Energy Efﬁcient Refrig-

eration and Flexible Power Consumption in a Smart

Grid. In Proceedings of Ris{\o} International Energy

Conference, pages 164–175.

M˚ansson, T. (2016). Energy in supermarkets. Licentiate

thesis, Chalmers University of Technology.

M˚ansson, T. and Ostermeyer, Y. (2013). The potential of

thermal energy storage in food cooling processes in

retail markets for grid balancing. In Nordic Sympo-

sium on Building Physics 2013, pages 1–6, Lund.

Ochieng, E. G., Jones, N., Price, a. D. F., Ruan, X., Egbu,

C. O., and Zuofa, T. (2014). Integration of energy ef-

ﬁcient technologies in UK supermarkets. Energy Pol-

icy, 67:388–393.

Pedersen, R., Schwensen, J., Biegel, B., Stoustrup, J., and

Green, T. (2014). Aggregation and control of super-

market refrigeration systems in a smart grid. IFAC

Proceedings Volumes (IFAC-PapersOnline), 19:9942–

9949.

Retail Forum for Sustainability (2009). Issue Paper on the

Energy Efﬁciency of Stores. Retail Forum for Sustain-

ability, 1(1):1–9.

Potential of Supermarket Refrigeration Systems for Grid Balancing by Demand Response

155

Sand´en, B. (2014). Systems Perspectives on Renewable

Power 2014. Chalmers University of Technology,

G¨oteborg, v1.0 edition.

Sinn, H. W. (2017). Buffering volatility: A study on the

limits of Germany’s energy revolution. European Eco-

nomic Review, 99:130–150.

Smale, N., Moureh, J., and Cortella, G. (2006). A review

of numerical models of airﬂow in refrigerated food

applications. International Journal of Refrigeration,

29(6):911–930.

Statens Energimyndighet (2010). STIL 2 - Ener-

gianv¨andning i handelslokaler. Technical report, En-

ergimyndigheten, Stockholm.

Statista (2013). Retail Week. (n.d.). Number of stores of

key food and grocery retailers in Germany in 2013, by

store count.

Tassou, S. A., Ge, Y., Hadawey, A., and Marriott, D. (2011).

Energy consumption and conservation in food retail-

ing. Applied Thermal Engineering, 31(2-3):147–156.

Union, E. and Action, C. (2017). Energy Union. THE EU

AND series of the European Commission, 1(February

2017).

Zerrahn, A., Schill, W. P., and Kemfert, C. (2018). On

the economics of electrical storage for variable re-

newable energy sources. European Economic Review,

108:259–279.

SMARTGREENS 2019 - 8th International Conference on Smart Cities and Green ICT Systems

156