logoGlobal Energy Interconnection

Contents

Figure(0

    Tables(0

      Global Energy Interconnection

      Volume 5, Issue 1, Feb 2022, Pages 66-76
      Ref.

      Optimal configuration of 5G base station energy storage considering sleep mechanism

      Xiufan Ma ,Qiuping Zhu ,Ying Duan ,Xiangyu Meng ,Zhi Wang
      ( School of Electrical and Electronic Engineering, North China Electric Power University, Beijing 102206, P.R.China )

      Abstract

      The high-energy consumption and high construction density of 5G base stations have greatly increased the demand for backup energy storage batteries.To maximize overall benefits for the investors and operators of base station energy storage, we proposed a bi-level optimization model for the operation of the energy storage, and the planning of 5G base stations considering the sleep mechanism.A multi-base station cooperative system composed of 5G acer stations was considered as the research object, and the outer goal was to maximize the net profit over the complete life cycle of the energy storage.Furthermore, the power and capacity of the energy storage configuration were optimized.The inner goal included the sleep mechanism of the base station, and the optimization of the energy storage charging and discharging strategy, for minimizing the daily electricity expenditure of the 5G base station system.Additionally, genetic algorithm and mixed integer programming were used to solve the bi-level optimization model, analyze the numerical example test comparison of the three types of batteries and the net income of the configuration, and finally verify the validity of the model.Furthermore, the sleep mechanism, the charging and discharging strategy for energy consumption, and the economic benefits for the operators were investigated to provide reference for the 5G base station energy storage configuration.

      0 Introduction

      Under the strategy of “New Infrastructure” and“Information Power”, the scale of 5G base station construction has gradually expanded.The backup battery of a 5G base station must ensure continuous power supply to it, in the case of a power failure.As the number of 5G base stations, and their power consumption increase significantly compared with that of 4G base stations, the demand for backup batteries increases simultaneously.Moreover, the high investment cost of electricity and energy storage for 5G base stations has become a major problem faced by communication operators.The traditional configuration method of a base station battery comprehensively considers the importance of the 5G base station, reliability of mains,geographical location, long-term development, battery life,and other factors [1].Presently, communication operators and tower companies generally configure a uniform group of 400 A·h batteries that provides a backup time of 3~4 h,for a 5G acer station based on the traditional configuration.This configuration faces the problems of idle energy storage assets, and low investment utilization rate.Additionally, in the context of carbon peak and carbon neutrality in China,the permeability of clean energy, such as photovoltaics and wind power, shall gradually increase, and the problem of peak regulation of power grids shall become increasingly prominent.Energy storage batteries, as the main flexible regulation resource in a power system [2], could effectively solve this problem.With the introduction of innovative technologies, such as the 5G base station, intelligent energy saving, participation in peak cutting and valley filling, and base station energy storage resources can be effectively activated to help achieve a win-win situation for both the power grid and the communication operators.Therefore,determining, and reasonably configuring the charging and discharging strategy for a 5G base station energy storage has become an important issue that needs to be solved urgently.

      Recently, scholars worldwide have been conducting extensive research on the optimal allocation of energy storage.In these studies, the research focus was split between the perspectives of energy storage configuration and operation.Reference [3] solely studied the configuration of energy storage, whereas [4] only studied the operation of energy storage.Currently, there is urgent need for research that comprehensively considers both the configuration and operation of energy storage.The existing models for optimal allocation of energy storage can be roughly divided into three categories: single-layer model, two-stage model and two-layer model. [5-6] established a single-layer hybrid optimization model for distribution network batteries.In [7-9], the configuration and operation of energy storage were divided into two stages.The power and capacity of energy storage were optimized first, and the day-ahead charge/discharge strategy of the energy storage was optimized after the configuration results were obtained. [10-13] studied the two-layer decision-making problem of energy storage planning and operation, and obtained optimal configuration results and optimal operation results of energy storage, through the iterative solution of inner and outer layers.Compared with the single-layer and two-stage models, the two-layer model can be used to study two different or even contradictory objectives, and better reflect the characteristics of practical problems.

      Presently, there are relatively few studies on the energy storage configuration of 5G base stations.Reference [14]proposed a plan for transforming the power supply of the machine room based on existing 5G base station site resources, without considering the existing 2G/4G base station energy storage configurations.Reference [15]proposed a capacity calculation method, and configuration results of energy storage batteries for three types of 5G base station sites, namely, distributed unit (DU) centralized machine room, common machine room, and active antenna unit (AAU) remote site; however, the configuration was not optimized from the perspective of the full life cycle of the energy storage.Reference [16] proposed a refined configuration scheme for energy storage in a 5G base station, that is, in areas with good electricity supply,where the backup battery configuration could be reduced.Compared with the traditional battery configuration, the battery allocation in the whole network could be reduced by 30% on average; however, the factors considered were relatively singular.Reference [17] revealed that the 5G base station energy storage could participate in demand response,and obtain certain benefits when it meets the basic power backup requirements.Reference [18] analyzed the problems existing in the current power configuration of base stations,and proposed solutions, such as superimposed photovoltaic power generation system, remote supply and centralized power supply, peak clipping and valley filling power supply,and integration of new battery power supply; however, it did not involve a specific configuration method or model of the base station energy storage.

      In summary, currently, there is abundant research on energy storage optimization configuration.However, most of the research on the energy storage configuration of 5G base stations does not consider the factors of participation of energy storage in demand response, and the optimization models are rarely implemented.A gap in the research on energy storage optimization configuration of 5G base station combined with the sleep mechanism of base station remains.Therefore, in this article, we considered the base station energy storage in the interests of the investors and operators, and established a two-layer optimization model considering the sleep mechanism.The model added 5G acer station transmission power constraints, and other constraints ensuring reliable backup power supply, optimizing energy storage configuration, and the charging and discharging strategy, under the premise of meeting 5G communication coverage area, and backup power supply reliability.

      1 Characteristics analysis of 5G base station

      1.1 Composition of 5G base station

      In the traditional configuration mode, a 5G acer station is composed as shown Fig.1, and is mainly divided into two parts: iron tower, and equipment room.The AAU and baseband processing unit (BBU) arranged on the iron tower are connected through optical fibers, and the equipment room is equipped with batteries, air conditioning,monitoring, and other equipment.

      Fig.1 Structure diagram of 5G acer station

      1.2 5G acer base station power consumption model

      The power consumption of a 5G acer base station changes in real time according to the state of the base station, and the change in communication load.Its power consumption model [20] is expressed as follows in (1).

      where δ is a constant that represents the incremental power consumption of the 5G base station when unit transmitting power is increased.

      1) When the base station is in active state, its power loss Pactive consists of transmitting power Ptx and inherent power Pfix.With an increase in the communication load of the acer station, the corresponding transmitting power Ptx increases linearly.The inherent loss Pfix refers to the loss in data processing units, power amplifiers, cooling devices and other components, which changes negligibly with the communication load.

      2) When the base station is in sleep state, its power loss is fixed as Psleep.

      1.3 Base station sleep mechanism

      The electricity bill of 5G base stations has become the most important operating cost of operators, and three communication operators, that are China Mobile, China Telecom, China Unicom, all regard energy saving and electricity charge reduction as an important measure to save costs.The sleep mechanism of a base station refers to the intelligent shutdown of major power consumption devices,such as the AAU of the base station, when there is no load or the load is low, such that the energy consumption is greatly reduced.Base station sleep mechanism has attracted much attention because it is easy to implement, and does not require change of hardware.In August 2020, China Unicom regularly activated the deep sleep function for the AAU of three different kinds of base stations that were connected to the network under no-load state, to implement intelligent control of base station energy consumption.

      The principle of the base station sleep mechanism involves selecting base stations with little or no load,to sleep according to the dynamic changes in the communication load, and transferring the communication load of the sleeping base station to neighboring base stations, thereby reducing power consumption while meeting the requirements of communication coverage.Based on this principle, the execution steps of the sleep mechanism adopted in this article were as follows:

      1) Obtain typical daily communication load curves and power consumption curves of various base stations.

      According to the real situation, set the value of δ,Pfix, Psleep.Additionally, the communication load curve is known.Since the communication load of the base station is positively correlated with the transmission power, the transmission power of the base station under different communication load states can be obtained proportionally.Substitute the known parameters into formula (1), and the typical daily power consumption curves of various base stations before performing sleep mechanism can be obtained.

      2) Select the periods where various base stations experience no load.

      Based on the typical daily communication load curve of the base station, the communication loads of the base station in each time period are compared separately, and the time periods where the base station experiences the no load state in 24 hours are selected.When the communication load of a base station is zero, it directly enables the sleep state.

      3) Select the periods where various base stations experience light load.

      Based on the typical daily communication load curve of the base station, the communication loads of the base station in each time period are compared separately.When a base station is lightly loaded, the fourth step is implemented.

      4) Determine whether the lightly-load base station can transfer the communication load, and determine which base station to divert the associated users to.

      First, determine whether other base stations adjacent to this base station are in the sleep state during this time period.If so, this base station cannot perform communication load transfer during this time period, so it remains in the original active state.If there is at least one base station, whose neighboring base station is active, the load status is further judged, and the base station that can undertake the load transfer task and has a lighter load status, is selected for transfer.

      5) Calculate the power consumption curve of different base stations after the sleep mechanism is executed.

      The transmitting power of the base station undertaking the offloading task is superimposed on the transmitting power of the sleep base station, and substituted into the 5G acer base station power consumption model, to calculate the power consumption after the communication load is transferred in the corresponding time period.The power consumption remains unchanged except for no-load and light-load periods.Finally, the power consumption curve after executing the sleep mechanism is obtained.

      It is worth noting that the basic coverage provided by wireless communication must be guaranteed when using the base station sleep mechanism [19].The communication coverage of a base station is closely related to transmitting power, frequency, and other factors.When the frequency of a base station increases and the transmitting power decreases, its coverage decreases.As the coverage radius of a single 5G acer station is generally lesser than 500 m, their transmitting power must be limited within a certain range.

      2 Energy storage optimization configuration model of 5G base station considering sleep mechanism

      2.1 Overview of bi-level optimization model

      The bi-level decision problem is a type of hierarchical decision problem with a master-slave structure.The outer layer decision can restrict the inner layer solution, and the optimal decision of the inner layer affects the outer layer decision.Thus, a mutually restricted master-slave relationship between the inner and outer layers is formed,which can effectively improve the objectivity of the optimization results, with respect to the balance of interests of each objective function.In the optimal configuration of energy storage in 5G base stations, long-term planning and short-term operation of the energy storage are interconnected.Therefore, a two-layer optimization model was established to optimize the comprehensive benefits of energy storage planning and operation.Fig.2 shows the bilevel optimization model architecture for the energy storage configuration established in this article.

      Fig.2 Optimal configuration model architecture

      2.2 Outer layer optimization model

      From the perspective of the base station energy storage investor, the maximum net income in the life cycle of the base station energy storage system was considered as the objective function of the outer layer, and the decision variables were the total rated power, and rated capacity.The constraint conditions of the energy storage configuration in the multi-base station cooperative system included energy storage investment cost constraints, and energy storage battery multiplier constraints; the time scale was in years.

      The outer objective function, was expressed as follows in (2).

      where F is net income over the life cycle of energy storage,F1 is the arbitrage of “low charge and high discharge”,that is, the energy storage charges in the low price of electricity, and discharges in the high price of electricity, F2 is energy storage system discharge subsidy revenue from the government, F3 is recycling value of energy storage batteries, F4 is special subsidies for the construction of 5G supporting power facilities, C1 is initial investment cost, and C2 is operation and maintenance cost.

      In this article, we assumed that the 5G base station adopted the mode of combining grid power supply with energy storage power supply.In the context of time-ofuse electricity prices, the base station energy storage was regulated to be charged when the electricity price was low,and discharged to the grid when the electricity price was high, to achieve “low charge and high discharge” arbitrage.The arbitrage of “low charge and high discharge” is expressed as follows:

      To encourage the development of energy storage on the user side, energy storage is usually subsidized according to the amount of discharge.Therefore, in this article, we calculated the government electricity price subsidy income for energy storage and discharge as follows:

      where Q2 is the daily subsidy income from energy storage system, and mb is the government subsidized electricity price.

      The recycling value, initial investment cost, and operation and maintenance cost of energy storage,respectively were expressed as follows:

      where σ is the recovery factor, cp is the unit charge/discharge power cost, ce is the unit capacity cost, co is the annual operation and maintenance cost per unit charge/discharge power of energy storage system, Pmax is the rated power of energy storage system, and Emax is the rated capacity of energy storage system.

      Since the number of repetitions of charging and discharging is an important factor affecting the life of energy storage batteries, combined with the current status of energy storage applications [21], the operation cycle of energy storage batteries does not exceed 10 years in the case of frequent charging and discharging of energy storage batteries.The formula for calculating the actual life of the energy system could be expressed as follows in (10).

      where N is the cycle life of energy storage battery, and n is the charge and discharge times of energy storage in a day.

      The constraint conditions of the outer model included investment cost constraint, and energy storage battery multiplier constraint [22], expressed as follows in (11) and(12), respectively.

      where Cmax is the investment cost limit, and β is the energy multiplier of energy storage battery.

      2.3 Inner layer optimization model

      From the perspective of the base station energy storage operator, for a multi-base station cooperative system composed of 5G acer base stations, the objective function of the inner model was to minimize the daily electricity cost of the system, and the decision variable was the daily charging and discharging power of the energy storage.The constraints were the power constraint, the energy constraint considering the backup power demand of the base station,and the 5G base station transmitting power constraint; the time scale was in days.

      The inner objective function, was expressed as follows in (13).

      where f is the daily net electricity expenditure of 5G operators, Z′ is the set of base stations in active state in the system at period i, is the set of base stations in sleep state in the system at period i, Pt x,z(i ) is the transmitting power of base station z at period i, and is the total power consumption of a multi-station system at period i.The nonlinear part of the product of the 0-1 variable and the continuous variable were linearized using the method stated in Appendix A.

      The constraint conditions of the inner model includes power constraints, and capacity constraints.The power constraints were expressed as follows in (14) and (15).

      The main difference between 5G base station energy storage and other ordinary user-side energy storage is that the base station must guarantee power backup.Since the communication load of the base station is different at each time of the day, the minimum reserve capacity required to ensure the reliability of the power supply of the base station is different at each time [23].Therefore, the minimum standby capacity of the energy storage system for each time period was calculated, as follows in (16).

      The capacity of each period of energy storage must be greater than the minimum standby capacity and less than the rated capacity, expressed as follows in (17).

      where is the minimum spare capacity for period i,and is the minimum storage standby time.

      Storage capacity must also ensure continuity between time periods, expressed as follows in (18).

      where E(i) and Ei(+1) are the capacities of the energy storage system at period i and period i+1, respectively, and ηch and ηdisare charge and discharge efficiencies of the energy storage, respectively.

      Additionally, to ensure the periodicity of the continuous operation of the energy storage, the energy stored at the beginning and end of the day must be consistent, expressed as follows in (19).

      Emission power constraints of 5G base station, was expressed as follows in (20).

      where Ptx,max is the maximum transmitting power of base station.

      3 Solution of two-layer optimization model for energy storage configuration

      Genetic algorithm and mixed integer programming were used to solve the bi-level model using Matlab platform.The outer model adopted genetic algorithm, transferring the initial decision variables, that is, the rated power and rated capacity of base station energy storage system, to the inner layer.The inner layer invoked the Yalmip toolbox and the Cplex solver to optimize the daily charge and discharge strategy under the initial energy storage rated power and rated capacity, and transferred it to the outer layer.Furthermore, the outer layer calculated the net income in the life cycle of the base station energy storage system,and obtained the fitness value.Subsequently, the updated configuration power and capacity of the base station energy storage system were obtained through genetic operations,such as crossover and mutation, and the optimal solution of the two-layer model was finally acquired through repeated iterations.Fig.3 shows the specific solution process.

      Fig.3 Flow chart of solution

      4 Case study

      4.1 Case description

      Considering the sleep mechanism of the base station,and the scale of the energy storage configuration, 50 5G acer base stations in a certain city were selected as a system.The stations were located in a mixed work and residence area (type 1), university boarding school area(type 2), restaurant and cafe area (type 3), residential area(type 4), and business circle workspace factory area (type 5).There were ten base stations of each type.Fig.4 shows the system structure.We used the optimized configuration model established in this paper to analyze this case.The detailed explanation of the sleep mechanism implemented in the case study are presented in Appendix B.Fig.B1 shows the power consumption curves of 5G base stations at the edges of various areas before and after performing sleep mechanism.Fig.B2 shows the power consumption curves of 5G base stations inside various regions before and after performing sleep mechanism.Fig.B3 shows the total system power consumption and reduced percentage before and after performing sleep mechanism.The configuration results only indicated the total energy storage resources of the system.The energy storage batteries of the 5G base station were arranged in a decentralized manner, and were distributed locally in the machine rooms of each 5G acer base station.

      Fig.4 Schematic diagram of system

      Since China uniformly implements general industrial and commercial electricity prices for 5G base stations, the general industrial and commercial peak and valley time-ofuse electricity prices of a certain city were selected, as listed in Table C1.The inflation rate was 2%, and the discount rate was 8%.Considering the operation and maintenance time of the battery, the battery usage days in a year was assumed to be 345 days.Three types of energy storage batteries were selected: lead-carbon batteries, brand-new lithium batteries, and cascaded lithium batteries.Table C2 lists the specific parameters of the energy storage batteries.The energy multiplier of an energy storage battery was 2.74.Based on the actual situation, the minimum storage backup time was set to 3 h, the battery recovery factor was 0.5, the energy storage discharge subsidy was 0.3 CNY per kW·h,and the investment cost limit was 2 million CNY.

      4.2 Analysis of optimized configuration results

      The 5G base station energy storage optimization configuration double-layer model was solved using the Matlab platform, and Table 1 lists the optimization configuration results obtained for the three types of batteries.

      Table 1 Optimal configuration results of 5G base station energy storage

      Battery type Leadcarbon batteries Brandnew lithium batteries Cascaded lithium batteries Pmax/kW 648 271 442 Emax/(kW·h) 1,775.50 742.54 1,211.1 Battery life/year 1.44 4.97 4.83 Life cycle cost/104 CNY 194.70 187.99 192.35 Lifetime earnings/104 CNY 200.98 203.05 201.23 Net income over the life cycle/104 CNY 6.28 15.06 8.88

      Different battery types affected the size of the system’s energy storage configuration, battery life, and net income during the full life cycle.It is seen from Table 1 that in the 5G macro base station system, the net income of the brand-new lithium battery was the highest, followed by the cascaded lithium batteries, and the lead-carbon battery.The brand-new lithium battery had the advantages of a long service life, green environmental protection, and light weight.In the case of frequent charging and discharging of energy storage, they exhibited longer life, higher efficiency, and the largest net benefit in the full life cycle.However, they had the disadvantages including high cost,and poor security.Since the current cascaded utilization of lithium batteries had a high hidden cost, and the model built in this article did not consider the environmental benefits of using cascaded lithium batteries, the resulting economy was not ideal.Lead-carbon batteries had a lowcost advantage similar to that of traditional lead-acid batteries, thus under the same investment cost constraints,their configured capacity was relatively larger; however on account of their low energy density, they were not suitable for use in the communication base station where the energy storage battery is required to be light in weight, and small in volume.Additionally, the lead-carbon battery had a short life when it was charged and discharged many times,thus resulting in a low net profit over the full life cycle.In summary, since the relevant technical conditions for battery echelon utilization were not sufficiently mature, the 5G acer base station system was most suitable to be equipped with a brand-new lithium battery, with an optimal configuration power of 271 kW, and an optimal configuration capacity of 742.54 kW·h.

      Table C1 Time-sharing electricity price for general industrial and commercial users

      Time period Time Electricity prices/(CNY/kW·h)Flat 7:00—8:00、12:00—19:00 0.6578 Critical peak 19:00—21:00 1.1183 Peak 8:00—12:00、21:00—23:00 0.9867 Valley 23:00—7:00 (the next day) 0.3289

      Table C2 Energy storage battery related parameters

      Battery types Ce/[CNY/(kW·h)]Cp/[CNY/(kW)]Co/[CNY/(kW·a)]η/% N/times Lead-carbon 650 1200 25 88 3000 Brand new lithium 2000 1120 97 90 12000 Cascaded lithium 1150 760 48 75 5000

      4.3 Analysis of optimized operation results

      It is seen from Fig.B3 that the percentage reduction in system power consumption of the 5G base station was up to 23.45% after the sleep mechanism was implemented,and the sleep mechanism demonstrated an evident effect on reducing system power consumption when the base station communication load was less.

      Fig.5 shows the daily electricity bill, and the reduction in electricity bills under the four combinations.

      Fig.5 Daily electricity rate of base station system

      It is seen from Fig.5 that when energy storage was only used as a backup, the daily electricity bill was not influenced by the configuration plan; however, it was directly related to the load curve.After the implementation of the sleep mechanism, the daily electricity expenditure was reduced from 1,877.81 CNY to 1,863.87 CNY, which was a small reduction.Generally, under the three different battery configuration schemes, the daily electricity cost for energy storage with “low charge and high discharge” was lower than that when energy storage was only used for standby.After the sleep mechanism was implemented, the system incurred the lowest cost under the condition of energy storage with “low charge and high discharge”, wherein the lead-carbon battery exhibited the most evident effect on the reduction of electricity cost.This was attributed to the configuration capacity of this kind of battery being much larger than that of the other two kinds of batteries; thus,providing more space for the arbitrage, such that the effect on daily electricity cost reduction was more prominent.

      Fig.6 shows charge and discharge control strategies for the three different configuration schemes, after the execution of the sleep mechanism.When the charging and discharging power is positive, it denotes charging; when the charging and discharging power is negative, it denotes discharging; when the charging and discharging power is zero, it denotes the battery in a floating state, that is, neither charging nor discharging.

      Fig.6 Charging and discharging strategies under three configuration schemes

      As seen in Fig.6, for the three configuration schemes,three different charging and discharging strategies were obtained with the minimum daily electricity cost of the energy storage system as the optimization goal.There were some differences in the charging and discharging power, the charging and discharging period, and the times of charging and discharging.Therefore, in the bi-level optimization model established in this article, the outer layer energy storage configuration was closely related to the inner layer energy storage operation.

      Fig.7 shows the optimized results of charge and discharge after the implementation of the sleep mechanism,based on the results of the brand-new lithium battery configuration.

      Fig.7 Optimized results of energy storage charging and discharging

      It is seen from Fig.7 that there was a strong correlation between the charging and discharging strategy of energy storage and the time-of-use electricity price curve.Energy storage was charged when the electricity price was low,and discharged when the electricity price was high.After the original load curve was superimposed on the charge and discharge power, the composite load characteristics were found to be inversely related to the peak and valley of the electricity price.Therefore, when the electricity price was at its peak, the base station system had a low power load and would discharge to the grid in part of the time.Conversely, when the electricity price was at its low, the base station system had a high power load.This was a concrete embodiment of the 5G base station playing its peak shaving and valley filling role, and actively participating in the demand response, which helped to reduce the peak load adjustment pressure of the power grid.

      5 Conclusion

      In this article, we established a bi-level optimization model for a 5G base station energy storage configuration considering the sleep mechanism, taking into account the time-scale difference of the two-layer model, the “low charge and high discharge” arbitrage during the full life cycle of the 5G base station energy storage, and factors such as special subsidies for the construction of 5G base stations,full life cycle costs, and the cost of daily electricity charges for 5G base stations.The analysis results of the case study showed that:

      1) The effectiveness of the bi-level model established in this article was effectively verified.The model comprehensively considered both the 5G base station energy storage operation and planning issues.On the premise of ensuring energy storage and backup power, it greatly improved the income during the life cycle of the energy storage.

      2) The optimized configuration results of the three types of energy storage batteries showed that since the current tiered-use of lithium batteries for communication base station backup power was not sufficiently mature, a brandnew lithium battery with a longer cycle life and lighter weight was more suitable for the 5G base station.

      3) The base station sleep mechanism could reduce the power consumption of the base station, while meeting the communication coverage requirements.There was a strong correlation between the charging and discharging behavior of the base station energy storage and the time-of-use electricity price curve.It could fulfill the principle of “low charge and high discharge” to a great extent.Furthermore, it could alleviate the problem of high electricity bills for 5G base stations, and reduce the pressure on power grid peak shaving.

      The optimization configuration method for the 5G base station energy storage proposed in this article, that considered the sleep mechanism, has certain engineering application prospects and practical value; however,the factors considered are not comprehensive enough.Further research will be conducted in the follow-up on the collaborative optimization of the energy storage configuration of 5G base stations, distributed photovoltaics,and wind power.

      Appendix A

      The nonlinear part of the product of the 0-1 variable,and the continuous variable in (13), that is, Pc h(i)Rc h(i ) and Pd is (i) Rd is (i ), were linearized using the following method:

      Add constraints as follows:

      where M is a large positive integer.

      Appendix B

      The detailed explanation of the sleep mechanism implemented in the case study of this article is as follows:

      In the base station system, only a few base stations that were located at the edge of the area (the coverage area overlaps with other types of base stations) had the possibility of achieving communication load transfer; whereas, base stations located inside the area (all nearby base stations of the same type) could not transfer the communication load,only when the communication load is zero can it go to sleep.Therefore, it was assumed that 2 of the 10 base stations of each type were located at the edge of the area (the coverage area overlaps with other types of base stations), and 8 base stations were located inside the area (all nearby base stations of the same type).

      Assuming Ptx,max = 200 W, δ = 15, Pfix = 1000 W, and Psleep = 600 W, when the communication load of the base station in a certain period of time was lower than 6% of the peak communication load of this type of base station in one day, it was regarded as light load.

      · The steps for 5G base stations located at the edges of various areas to perform the sleep mechanism are as stated in 1)-5), in Section 1.3 of the main text.Fig.B1 shows the power consumption curves before and after performing the sleep mechanism.

      Fig.B1 Load curves of 5G base stations at the edges of various areas before and after sleep mechanism

      · The steps for 5G base stations located inside various areas to perform the sleep mechanism are as stated in 1)-2), in Section 1.3 of the main text.Fig.B2 shows the power consumption curves of 5G base stations before and after the sleep mechanism.

      Fig.B2 Load curves of 5G base stations inside various regions before and after sleep mechanism

      The power consumption of the five types of base stations located at the edge of the area, and the inside of the area were superimposed to obtain the total power consumption curve of the multi-base station cooperative system before and after executing the sleep mechanism, as shown in Fig.B3.

      Fig.B3 Total system power consumption and reduced percentage before and after sleep mechanism

      Appendix C

      Acknowledgements

      This work was supported by the State Grid Science and Technology Project (KJ21-1-56).

      Declaration of Competing Interest

      We declare that we have no conflict of interest.

      References

      1. [1]

        Xu H, Tian Z Y, Zhao Z H (2020) Research on base station power system innovation based on 5G network.Telecom Power Technology, 37(02): 23-26+29 [百度学术]

      2. [2]

        Jiang H Y, Du E S, Zhu G P, et al.(2020) Review and prospect of seasonal energy storage for power system with high proportion of renewable energy.Automation of Electric Power Systems,44(19): 194-207 [百度学术]

      3. [3]

        Alharbi H, Bhattacharya K (2017) Stochastic optimal planning of battery energy storage systems for isolated microgrids.IEEE Transactions on Sustainable Energy, 9(1): 211-217 [百度学术]

      4. [4]

        Feng B, Guo Y Z, Chen Y, et al.(2021) Charging and discharging strategy of cloud energy storage based on GRU multi-step prediction technology.Automation of Electric Power Systems,45(09):46-54 [百度学术]

      5. [5]

        Xiang Y P, Wei Z N, Sun G Q, et al.(2015) Life cycle cost based optimal configuration of battery energy storage system in distribution network.Power System Technology, 39(01): 264-270 [百度学术]

      6. [6]

        Liu W X, Yang M Y, Wang J, et al.(2021) Energy storage planning for incremental power distribution systems based on the intellectual generating method of operational strategies.Proceeding of the CSEE, 41(10): 3317-3329+3658 [百度学术]

      7. [7]

        Guo J Y, Liu Y, Guo Y L, et al.(2020) Configuration evaluation and operation optimization model of energy storage in different typical user-side.Power System Technology, 44(11): 4245-4254 [百度学术]

      8. [8]

        Zhao Y T, Wang H F, He B T, et al.(2020) Optimization strategy of configuration and operation for user-side battery energy storage.Automation of Electric Power Systems, 44(06): 121-128 [百度学术]

      9. [9]

        Chen H H, Du H H, Zhang R F, et al.(2018) Optimal capacity configuration and operation strategy of hybrid energy storage considering uncertainty of wind power.Electric Power Automation Equipment, 38(08): 174-182+188 [百度学术]

      10. [10]

        Ding Y X, Xu Q S, Lv Y J, et al.(2019) Optimal configuration of user-side energy storage considering power demand management.Power System Technology, 43(04): 1179-1186 [百度学术]

      11. [11]

        Ma X F, Chen Jing, Yu S Y, et al.(2020) Research on user side energy storage optimization configuration considering capacity market.Transactions of China Electrotechnical Society, 35(19):4028-4037 [百度学术]

      12. [12]

        Li J L, Guo B Q, Niu M, et al.(2018) Optimal configuration strategy of energy storage capacity in wind/PV/storage hybrid system.Transactions of China Electrotechnical Society, 33(06):1189-1196 [百度学术]

      13. [13]

        Dvorkin Y, Fernandez-Blanco R, Kirschen D S, et al.(2016)Ensuring profitability of energy storage.IEEE Transactions on Power Systems, 32(1): 611-623 [百度学术]

      14. [14]

        Gong G Y, Ding Y (2019) Solution to power supply reform of 5G base station.Telecom Power Technology, 36(03): 106-108 [百度学术]

      15. [15]

        Lin T L, Tan L J, Dong B (2020) Discussion on power matching solutions for 5G site construction.Designing Techniques of Posts and Telecommunications, 2020(04): 89-92 [百度学术]

      16. [16]

        Chen D X (2019) The impact of 5G on power supply matching and its ountermeasure.Information and Communications Technologies, 13(04): 32-37 [百度学术]

      17. [17]

        Liu J H, Guo P, Li H J, et al.(2020) Feasibility study on energy storage configuration and demand response of 5G base station.2020 Excellent Scientific and Technological Papers Review Conference of Henan Society for Electrical Engineering, 28 August, Zhengzhou Henan, China:6 [百度学术]

      18. [18]

        Huang H, Tong E, Wang W B, et al.(2020) Discussion on base station power demand and solutions in 5G deployment.Information & Communications, 2020(11): 177-179 [百度学术]

      19. [19]

        Li D (2018) Research on base station sleep mode technology under 5G dense heterogeneous nerwork.Dissertation, Beijing University of Posts and Telecommunications [百度学术]

      20. [20]

        Liu Y B, Wang Q, Zeng Q, et al.(2021) Key technologies and Prospects of energy consumption management for 5G network in background of energy internet.Automation of Electric Power Systems, 45(12): 174-183 [百度学术]

      21. [21]

        Sun Z, Tian H P, Wang W X, et al.(2021) Research on economy of echelon utilization battery energy storage system for user-side peak load shifting.Acta Energiae Solaris Sinica, 42(04): 95-100 [百度学术]

      22. [22]

        Chen L J, Qu T T, Liu H B, et al.(2019) Demand management based two-stage optimal storage model for large users.Automation of Electric Power Systems, 43(01): 194-200 [百度学术]

      23. [23]

        Yong P, Zhang N, Ci S, et al.(2021) 5G communication base stations participating in demand response: key technologies and prospects.Proceeding of the CSEE, 41(16): 5540-5552 [百度学术]

      Fund Information

      Author

      • Xiufan Ma

        Xiufan Ma received the B.S.and M.S.degrees in Electrical Engineering from Northeast Electric Power University, Jilin, China, in 1992 and 1995, respectively, and a Ph.D.degree in Electrical Engineering from North China Electric Power University, Beijing, China, in 2013.She is currently an Associate Professor with the School of Electrical and Electronic Engineering, North China Electric Power University.Her current major research interests include distribution network planning and operation, and electric vehicle planning and operation.

      • Qiuping Zhu

        Qiuping Zhu received the B.S.degree in Electric Engineering from Shanghai University of Electric Power, Shanghai, China, in 2019.She is currently working towards a Master’s degree at the School of Electrical and Electronic Engineering, North China Electric Power University, Beijing, China.Her research interests include the optimal configuration of energy storage, and the power market.

      • Ying Duan

        Ying Duan received the B.S.degree in Electric Engineering from North China Electric Power University, Beijing, China,in 2018.She is currently working towards a Master’s degree at the School of Electrical and Electronic Engineering, North China Electric Power University, Beijing, China.Her research interests include the optimal configuration of energy storage, and the power market.

      • Xiangyu Meng

        Xiangyu Meng received the B.S.degree in Electric Engineering from Minzu University of China, Beijing, China, in 2018.She is currently working towards a Master’s degree at the School of Electrical and Electronic Engineering, North China Electric Power University, Beijing, China.Her research interests include energy storage optimization scheduling, and the power market.

      • Zhi Wang

        Zhi Wang received the B.S.degree in Electric Engineering from North China Electric Power University, Beijing, China, in 2019.He is currently working towards a Master’s degree at the School of Electrical and Electronic Engineering, North China Electric Power University, Beijing, China.His research interests include energy storage optimization scheduling, integrated energy systems, and the power market.

      Publish Info

      Received:2021-12-08

      Accepted:2022-01-26

      Pubulished:2022-02-25

      Reference: Xiufan Ma,Qiuping Zhu,Ying Duan,et al.(2022) Optimal configuration of 5G base station energy storage considering sleep mechanism.Global Energy Interconnection,5(1):66-76.

      (Editor Yajun Zou)
      Share to WeChat friends or circle of friends

      Use the WeChat “Scan” function to share this article with
      your WeChat friends or circle of friends