1 On the economic viability of solar energy when upgrading cellular networks

2025-04-30 0 0 861.21KB 25 页 10玖币

侵权投诉

On the economic viability of solar energy when

upgrading cellular networks

Zineb Garroussi∗†, Abdoul Wassi Badirou∗†, Mathieu D’amours∗†, André Girard∗†,

Brunilde Sansò∗†

∗Electrical Engineering Department

Polytechnique Montréal, Québec, Canada

†LORLAB and Research Group in Decision Analysis (GERAD)

{zineb.garroussi, abdoul-wassi.badirou, mathieu.damours, andre.girard,

brunilde.sanso}@polymtl.ca

Abstract

The massive increase of data traﬃc, the widespread proliferation of wireless applications and the full-

scale deployment of 5G and the IoT, imply a steep increase in cellular networks energy use, resulting in a

signiﬁcant carbon footprint. This paper presents a comprehensive model to show the interaction between

the networking and energy features of the problem and study the economical and technical viability of green

networking. Solar equipment, cell zooming, energy management and dynamic user allocation are considered

in the upgrading network planning process. We propose a mixed-integer optimization model to minimize

long-term capital costs and operational energy expenditures in a heterogeneous on-grid cellular network

with diﬀerent types of base station, including solar. Based on eight scenarios where realistic costs of solar

panels, batteries, and inverters were considered, we ﬁrst found that solar base stations are currently not

economically interesting for cellular operators. We next studied the impact of a signiﬁcant and progressive

carbon tax on reducing greenhouse gas emissions (GHG). We found that, at current energy and equipment

prices, a carbon tax ten-fold the current value is the only element that could make green base stations

economically viable.

Index Terms

Solar base station, cellular networks, green energy, greenhouse gas emissions, CAPEX, OPEX, cell

zooming, carbon tax, network upgrade, network planning, energy management.

I. Introduction

Cellular operators are facing traﬃc growth due to the increasing use of mobile applications [1] and

the rapid development of wireless access technologies such as 5G and beyond. To meet this demand,

new base stations (BSs) are being added or upgraded to the next generation technologies [2].

We claim that this current and future upgrading of cellular systems provides a great opportunity

for operators to reduce their environmental impact. The COP21 Paris agreement has set a target limit

of 2°C for the global warming. In that context, the Information and Communication Technologies

sector has a role to play given that it accounts for about 4% of the global energy use [3] and has

generated approximately between 1.8% to 2.8% of the global GHG emissions in 2020 [4]. A signiﬁcant

part of this is produced by cellular networks [5] which generate the equivalent of 220 MtCO2, which

represents 0.4% of global emissions [6]. It is estimated [7] that the full deployment of 5G could have

an environmental impact up to 2 to 3 times larger. Moving to the millimeter wavebands [1] will reduce

the range of the new 5G antennas. This in turn will require a much more dense radio infrastructure [8]

with the ensuing large increase in energy use.

One solution is to wait for the grid to use energy sources that don’t emit CO2. This is a major

undertaking that will take years to be implemented. In this paper, we look at a diﬀerent approach to

reduce the emissions of the wireless networks. We posit that it is possible to reduce both OPEX and

arXiv:2210.11475v1 [cs.CE] 17 Oct 2022

CO2 emissions by carefully integrating the energy management features within the planning process

while taking into account the CAPEX cost.

The idea is to make use of technology that is either currently available or that will become more

common with 5G full deployment: sleep mode, cell zooming, user allocation and solar power. These

can be implemented gradually as the need arises and do not depend on the general greening of the

power grid. More speciﬁcally, we want to see if this evolution can be driven by market economics

where the savings provided by the reduction in grid costs are suﬃcient to cover the additional capital

cost of the equipment. In cases where this is not possible, we want to see to what extent a carbon

tax can drive the process.

For this, we propose a model to evaluate the combined technical and economic viability of future

green cellular networks. This is done by a detailed modeling of energy, communication and demand

features and by an in-depth study on how the issues of energy management, solar energy, CO2

emission, CAPEX and OPEX are interrelated.

II. Literature Review

The literature relevant to this work is very large, encompassing the use of solar equipment, energy

management, dynamic user assignment and green planning. In what follows, we go through a quick

technological and literature survey of each one of those areas.

A. Solar Equipment

Solar panels have been obvious candidates for green networks for more than 10 years (See references

in [9]). The main advantage of solar energy is the fact that the cost per kW has been decreasing steadily

over the last 10 years and is currently the lowest of the more frequently used green sources [10]. Solar

panels are particularly useful and coveted for power stations in remote areas not covered by the

power grid and that use non-renewable resources. An extreme case is oﬀ-grid rural areas where base

stations are just diesel powered.

For instance, the technique proposed in [11] decreases both OPEX and greenhouse gas emissions for

remote rural base stations in Malaysia using solar photovoltaic/diesel generator hybrid power systems.

The economical and environmental viability of PV/diesel/battery hybrid system conﬁguration for a

BS in Nigeria was examined in [12]. Due to the large amount of solar energy available, this system

oﬀers an alternative power source for a BS by reducing operational costs and emissions of greenhouse

gases. A solar PV/Fuel cell hybrid system under the software HOMER is proposed in [13] to power a

remote base station in Ghana. The objective is to reduce both greenhouse gas emissions and lower the

levelized cost of electricity (LCOE). In this hybrid system, the LCOE is reduced by 67% compared to

diesel power. Similarly, the authors of [14] use the HOMER software to simulate PV-battery-diesel to

power a BS during 24 hours under South African climate. They minimize operation costs, emissions,

and power use. The potential of photovoltaic to power BS is studied in [15] for Kuwait where the

HOMER software is used to determine the number of PV, batteries, and converters for an oﬀ-grid

solar PV system with diesel generator while minimizing the net present cost in order to power a

cellular BSs.

There have also been some urban implementations. The work of [16] studies a case in urban areas of

South Korea where both on- and oﬀ-grid sites use standalone solar batteries to power a macro LTE

cellular base station. A multi-objective optimization algorithm is proposed in [17] for the optimal

sizing of a standalone PV/battery to power a BS under the climate of Sydney. The objective is the

minimum annual total life cost. An optimal decision of demand-side power management for a green

wireless base station is proposed in [18] to minimize the power cost and provide ﬂexibility under

traﬃc load, grid power price, and renewable source uncertainties.

These studies, rural or urban, all have something in common: they refer to a single base station

and, in most cases, they lack accounting for the replacement, degradation and installation costs for

the solar equipment.

B. Energy Management and Dynamic User Assignment

Given that 80% of the energy used by a cellular network comes from the radio infrastructure [19],

radio energy management must become an integral part of the base station operation. Even though

the existing management tools are performance-oriented, they make it possible to design management

strategies that can also be used to reduce energy use, based on the fact that real-world data shows

that most base stations are underutilized during low traﬃc periods [20].

In fact, it has already been shown that dynamic management may potentially save between 20%

to 30% of energy [21] in cellular networks if the system is planned from the start with enough base

stations. These savings are obtained either by completely turning oﬀ some base stations or by reducing

their power during key times of the day. Even though these methods are nothing new, they will be

more easily implemented in 5G networks and beyond using the notion of cell zooming.

The actual operation of cell zooming can be viewed in two ways. In [22], the total available power

can be reassigned arbitrarily among radio blocks which in turn can be allocated to diﬀerent down-

link transmissions. It is also stated in [23], [24] that a cell can extend its range through zooming. An

extreme case of cell zooming is putting some base stations in sleep mode [25], [26], [27], [28], or even

turning them oﬀ completely whenever user demand is lower. Diﬀerent objective metrics have been

used such as coverage [29], user demand [30] or the actual power usage [31]. A multi-criteria model

is presented in [32] to minimize both the energy use and the user drop probability.

A nonlinear model to minimize energy use subject to quality of service constraints is proposed

in [33]. The technique described in [34] uses solar power and on-oﬀ operation to minimize energy cost

by assigning users to diﬀerent base stations and choosing solar or grid power during the day.

One consequence of sleep mode or cell zooming is that one must re-assign users to diﬀerent base

stations during the day whenever a user can no longer be served by its base station due to a power

reduction. This has been examined in detail in [35], [36], [37]. The conclusion was that dynamic user

assignment can reduce the network cost signiﬁcantly when it is integrated into the long-term planning

model. These results will not be repeated here in order to conserve space.

C. Green Planning

The techniques described above can be viewed as network management and operate on a short-

term horizon of minutes or hours. Network planning, on the other hand, where one has to decide on

the installation of base stations and the equipment to go with it, works on a time scale of many years.

This is by itself a complex procedure since one needs to combine cells of diﬀerent sizes, including

macro, micro, pico, and femto cells. The size of base stations depends on several factors, including

the site location and the antenna position. Small stations serve areas with high traﬃc density using a

lower amount of energy [38], [39] but need to be combined with larger base stations for lower-density

areas.

Because of the large diﬀerence in time scales, the process was often split into two independent

parts: plan the equipment upgrades every year and manage the resulting network on a shorter scale

using the equipment available.

Unfortunately, network planning and network management are tightly coupled because one cannot

use a network management technique, such as solar cells, if the equipment has not been installed. In

fact, it has been shown that integrating the technology choice with the long-term planning process

produces networks signiﬁcantly cheaper than a two-step procedure [35], [36], [37]. There is thus a

deﬁnite need for an integrated model that takes into account the eﬀect of short-term management

techniques when deciding on equipment upgrades.

The problem of minimizing the total CAPEX and OPEX cost over a 10-year horizon has been

examined, albeit for a single base station, in [40]. They make a detailed model of the energy interac-

tions and study the trade-oﬀ between solar panels, a diesel generator or grid power. A similar energy

approach was presented in [34] when minimizing the energy cost of a given network. The variables are

the allocation of users to base stations and the energy source used by each base station during each

one of a given number of time intervals. Even though both [40] and [34] study the energy planning

of the network, neither treats the networking planning problem, in particular the decision of where

and when to locate new green base stations over the planning horizon.

Some limited work has been done on related topics. The allocation of users over a single day to

maximize the network operator’s revenue has been examined in [41].

The work of [42] explores the long-term network planning with green energy harvesting of solar

or wind sources. The problem is to select a subset of candidate BSs and to assign users to the

available base stations subject to a minimum SINR requirement in order to minimize the base station

installation and the user connection costs and the cost of electric grid.

The ﬁrst attempt at integrating the short-term beneﬁts of solar energy and network planning has

been presented in our previous work [37]. The goal was to minimize the total operating plus capital

cost of the network. The users can be re-assigned and the base station antennas can be switched oﬀ

depending on the demand at diﬀerent times of day.

Some work has also considered a marketing-oriented approach. A ﬁnancial analysis of network

upgrade is described in [43] to optimize the trade-oﬀ between the generated revenue produced by

the upgrade and its cost. A dynamic programming model is proposed and a fast heuristic solution

is used to compute solutions. A similar approach is used in [44] where a case study is presented to

evaluate the impact of energy trading either between base stations or through an energy broker and

also with spectrum sharing between operators.

D. Our contribution

Diﬀerently from the above body of work, this paper proposes a comprehensive approach to study

not just the technical or the economic viability of using solar equipment to update cellular networks,

but the interaction between the two. The strength of our contribution to the state of the art comes

from a detailed model of energy management and networking planning. It is precisely this level of

detail what allows us to clarify the technical and economic features of network upgrade.

1) We integrate the short-term network and energy management with the long-term network ex-

pansion over many years into a single model to provide a more realistic view of how the planning

and operation are inter-related.

2) The planning determines decisions on diﬀerent base stations sizes and types, including solar

panels, that are needed to update an existing network in an urban area.

3) The modeling of network and energy operation is very detailed and includes sleep mode, cell

zooming, user re-allocation, solar, batteries, inverters, controllers to reduce the energy use of

cellular networks

4) Very detailed and realistic costs are considered for solar equipment and battery purchase and

replacement. We also model the degradation of the eﬃciency of batteries and solar equipment

while taking into account the time value of money.

5) Realistic evolution of user demand, energy usage and illumination proﬁles are considered as well

as regulatory or environmental constraints on base station installations.

6) Co2 emissions and possible taxes are integrated into the model.

We want to use this model to answer some speciﬁc questions such as:

•How large is the cost reduction provided by allowing installations over the whole horizon as

opposed to a model where the installation decisions can be taken only at the beginning of the

planning horizon?

•Are the OPEX savings provided by solar energy large enough to justify the added CAPEX?

•Does cell zooming make solar energy more economical?

•Do the beneﬁts of cell zooming add up to those of solar?

•Is it possible to use cell zooming to reduce the CAPEX by delaying the installation of base

stations?

•What is the impact of a carbon tax on the reduction of greenhouse gases?

III. Mathematical Model

As can be seen from the previous discussion, planning a network is a complex task, from long-term

market considerations to technological choices and real-time network management.

In this work, we want to see how prices and taxes can be used to reduce carbon emission using

diﬀerent energy-saving techniques. For this, we need to compare the eﬀect of these techno-economic

options to some base case. If this is to be meaningful, problems have to be solved to optimality.

Therefore, the model’s complexity should not prevent us from ﬁnding optimal solutions. As a result,

in the model presented below, we have just kept enough level of detail for the study to be meaningful.

The approximations and assumptions will be clearly introduced as the model description progresses.

A. Assumptions

We now brieﬂy review some of the more important simpliﬁcations and assumptions that were made

and explain to what extent they are realistic.

1) Network Structure: In our model, the users are aggregated into so-called test points which can

be viewed as real concentrators or simply as a collection of users close together. We are given a set

of available test points for the whole planning horizon, some of which may be currently inactive.

The choice of technology is not modeled by separate decision variables. Instead, we introduce the

notion of a base station type which contains a description of the technical features of the base station,

e.g., whether it has solar panels or not, its size, e.g., pico, micro, etc. The test cases can then be run

with a small set of given types.

2) Demand: Network growth is driven by the increasing number of users and new applications

and services and the quality of service they require. This involves economic issues such as market

forecasting and technical issues such as power or bandwidth management. Because our focus is on

energy, we assume that these diﬀerent kinds of demands are converted in an energy requirement from

each test point. This can be done by the engineering or the traﬃc department of the operator and

is outside the scope of this paper. A simple example of such a procedure is given in the Appendix.

Demand growth is modeled by activating these test points at some future time and computing

their energy requirement as given by the demand forecast.

3) Forecasting: In practice, the results of a long-term planning model depend of the growth forecast

provided by the marketing department. Because the forecast accuracy decreases for later years, one

can use the model’s results for the coming year to decide whether to install new equipment for that

year only. The model can then be run each year with new forecasts and technology options. This

approach is more realistic than doing a one-year planning since it does take into account the future

demands and technology as they are known at the time when a decision has to be made.

4) Solar Energy Model: An important feature of solar power is that the energy production can

vary widely over short periods of the order of an hour or less and these variations are unpredictable.

A minimum amount of solar equipment is thus required to power a green base station. The main

components are the solar panels and the battery bank. We also model the charge controllers needed

to protect the batteries from overﬂowing and increase their lifetime. Finally, an AC-DC converter is

also required to power the base station with DC current from its battery bank. Figure 1 shows the

storage of solar energy into electrical energy as a back-up energy source to the electrical grid.

We use a solar proﬁle to estimate the amount of electricity produced during the day since replacing

the daily variation of sunlight by a single average value for the whole planning horizon would leave

out the daily changes in sunlight. This proﬁle is computed for an average day and assume that this

is representative of the operation for the whole year.

文档加载中……请稍候！
如果长时间未打开，您也可以点击刷新试试。

下载文档到电脑，查找使用更方便

10 玖币 0人已下载

立即下载

摘要：

1OntheeconomicviabilityofsolarenergywhenupgradingcellularnetworksZinebGarroussiy,AbdoulWassiBadirouy,MathieuD'amoursy,AndréGirardy,BrunildeSansòyElectricalEngineeringDepartmentPolytechniqueMontréal,Québec,CanadayLORLABandResearchGroupinDecisionAnalysis(GERAD){zineb.garroussi,abdoul-wassi.badir...

展开>> 收起<<

1 On the economic viability of solar energy when upgrading cellular networks.pdf

共25页,预览5页

还剩页未读，继续阅读

声明：本站为文档C2C交易模式，即用户上传的文档直接被用户下载，本站只是中间服务平台，本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间，仅对用户上传内容的表现方式做保护处理，对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私，请立即通知玖贝云文库，我们立即给予删除！

1 On the economic viability of solar energy when upgrading cellular networks

相关推荐

开通VIP享超值会员特权

作者详情

相关内容

热门标签

举报选择: