CN114679344B - 5G green base station power supply optimization method considering load and meteorological influence - Google Patents
5G green base station power supply optimization method considering load and meteorological influence Download PDFInfo
- Publication number
- CN114679344B CN114679344B CN202210300788.2A CN202210300788A CN114679344B CN 114679344 B CN114679344 B CN 114679344B CN 202210300788 A CN202210300788 A CN 202210300788A CN 114679344 B CN114679344 B CN 114679344B
- Authority
- CN
- China
- Prior art keywords
- base station
- power
- power supply
- fan
- load
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 25
- 238000005457 optimization Methods 0.000 title claims description 28
- 238000004146 energy storage Methods 0.000 claims abstract description 41
- 238000004891 communication Methods 0.000 claims abstract description 19
- 238000003915 air pollution Methods 0.000 claims abstract description 17
- 238000005265 energy consumption Methods 0.000 claims abstract description 14
- 230000000694 effects Effects 0.000 claims abstract description 12
- 238000013528 artificial neural network Methods 0.000 claims abstract description 7
- 230000006870 function Effects 0.000 claims description 34
- 238000010248 power generation Methods 0.000 claims description 26
- 230000005611 electricity Effects 0.000 claims description 10
- 238000010606 normalization Methods 0.000 claims description 10
- 230000002068 genetic effect Effects 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 8
- 238000009826 distribution Methods 0.000 claims description 7
- 230000008859 change Effects 0.000 claims description 5
- 239000011159 matrix material Substances 0.000 claims description 5
- 238000012549 training Methods 0.000 claims description 5
- 238000007599 discharging Methods 0.000 claims description 3
- 230000003993 interaction Effects 0.000 claims description 3
- 239000000203 mixture Substances 0.000 claims description 3
- 238000010219 correlation analysis Methods 0.000 claims description 2
- 238000003062 neural network model Methods 0.000 claims description 2
- 230000000946 synaptic effect Effects 0.000 claims description 2
- 238000010977 unit operation Methods 0.000 claims description 2
- 239000002699 waste material Substances 0.000 abstract description 3
- 210000002569 neuron Anatomy 0.000 description 3
- 238000003860 storage Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 229920002430 Fibre-reinforced plastic Polymers 0.000 description 1
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 239000011151 fibre-reinforced plastic Substances 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L12/00—Data switching networks
- H04L12/02—Details
- H04L12/10—Current supply arrangements
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/086—Learning methods using evolutionary algorithms, e.g. genetic algorithms or genetic programming
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/12—Computing arrangements based on biological models using genetic models
- G06N3/126—Evolutionary algorithms, e.g. genetic algorithms or genetic programming
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W24/00—Supervisory, monitoring or testing arrangements
- H04W24/06—Testing, supervising or monitoring using simulated traffic
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W24/00—Supervisory, monitoring or testing arrangements
- H04W24/08—Testing, supervising or monitoring using real traffic
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Physiology (AREA)
- Genetics & Genomics (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
According to the 5G base station energy consumption prediction result based on the artificial neural network, the base station energy storage adjustable capacity is considered to be related to the base station load state and meteorological factors including air pollution index, temperature, cloud cover and humidity by limiting the charge state range of the base station energy storage charge and discharge. The method plays the economic effect of standby energy storage and peak shaving of the base station on the basis of ensuring the communication requirement of the 5G base station, and simultaneously avoids the waste of wind and light clean renewable energy sources and electric energy.
Description
Technical Field
The invention relates to the field of communication system equipment, in particular to a 5G green base station power supply optimization method considering load and weather influence.
Background
The 5G base station construction is accelerating the progress of construction as an important component of the "new infrastructure". The number of 5G mobile phone terminals connected reaches 5.2 hundred million users, which is accumulated in China to open up more than 142.5 ten thousand 5G base stations. With the rapid increase of the number of 5G base stations, the standby energy storage of the base stations is an energy storage resource with considerable capacity. The standby energy storage of the 5G base station mainly serves as emergency power supply when the commercial power is abnormal, but the energy storage capacity of a single 5G base station is about three to four times of the power consumption of the 5G base station in peak time at present, and the value of idle resources of the 5G energy storage is not fully exerted.
Disclosure of Invention
In order to solve the problems, the invention provides a 5G green base station power supply optimization method considering load and meteorological influence, and according to a 5G base station energy consumption prediction result based on an artificial neural network, the energy storage adjustable capacity of a base station is related to the load state of the base station and meteorological factors including an air pollution index, temperature, cloud cover and humidity by limiting the charge state range of the energy storage charge and discharge of the base station. On the basis of guaranteeing the communication requirement of the 5G base station, the economic effect of standby energy storage and peak shaving of the base station is exerted, and meanwhile, the waste of wind and light clean renewable energy sources and electric energy is avoided.
A5G green base station power supply optimization method considering load and weather influence comprises the following steps:
step 1, statistics of operation history data of a 5G base station, fitting of a relation curve of energy consumption and communication load of the 5G base station, and establishment of cost functions of different power supply modes;
step 2, predicting the energy consumption of a future day by using an artificial neural network method based on a time sequence according to 24-hour energy consumption historical data of the 5G base station to obtain a prediction result;
step 3, obtaining a load rate index mu of each hour based on the predicted data of 24 hours power consumption of the 5G base station and the relation curve of energy consumption and communication load loadi ；
Step 4, collecting four meteorological data of air pollution index, temperature, cloud cover and humidity and photovoltaic and fan power generation data, and carrying out normalization processing on the data by adopting initialization;
step 5, establishing a mother sequence and a child sequence of photovoltaic power generation and fan power generation required by gray correlation analysis;
step 6, giving weights of the photovoltaic and the wind turbine to be gamma according to the wind-solar output ratio of each hour 1 ，γ 2 Wherein gamma is 1 、γ 2 ∈[0，1]，γ 1 +γ 2 ＝1；
Step 7, forming a correlation sequence by calculating a gray correlation coefficient and a correlation coefficient mean value, and finally obtaining an influence weight alpha of four meteorological factors including air pollution index, temperature, cloud cover and humidity on photovoltaic power generation 1 、α 2 、α 3 、α 4 And the influence weight beta on the generating capacity of the fan 1 、β 2 、β 3 、β 4 ；
Step 8, carrying out forward normalization processing on four meteorological data of air pollution index, temperature, cloud cover and humidity generated by photovoltaic power generation and fan in a practical sense to obtain a matrix Z ij (i=1, 2,3 … 24, j=1, 2,3,4,5,6,7, 8), and calculating the total influence coefficient mu of four meteorological factors on new energy power generation per hour by combining the step 6 and the step 7 weai ；
Step 9, calculating the lower limit S of the energy storage electric energy holding capacity of the 5G base station according to the step 3 and the step 8, wherein the lower limit S is calculated according to the communication load and four meteorological factors mini ；
Step 10, according to an optimal scheduling period, taking the minimum total power supply cost of the whole day as an optimal scheduling target, comprehensively considering power balance constraint, energy storage charge-discharge power constraint, constraint that a micro-grid consisting of a 5G base station group, a fan, photovoltaic and energy storage exchanges power with a power distribution network, constraint of energy storage capacity, and combining the data obtained in the step 2 and the step 9, and establishing a 5G base station power supply system optimal scheduling model based on flexible scheduling of 5G energy storage;
and 11, solving the 5G base station power supply system optimization scheduling model obtained in the step 10, and calculating the power supply composition of the load in each period and the power supply cost of the whole day.
Further, in the step 1, the cost functions of the different power supply modes include a fan operation cost function, a photovoltaic operation cost function and an energy storage unit cost function.
Further, in the step 3, a per hour base load factor index μ loadi ∈[0，1]The larger the communication load, μ loadi The larger.
Further, in step 5, the parent sequence is the output of the photovoltaic and fan for 24 hours, and the child sequence is the value of the four weather factors normalized for 24 hours. Here, x i (k) The value of the kth hour representing the ith meteorological factor is 1 for air pollution i, 2 for temperature i, 3 for cloud i, and 4 for humidity i.
Further, in the step 6, γ 1 ，γ 2 The output of the fan is respectively proportional to the output of the photovoltaic and the fan.
Further, in step 9, the lower limit S of the stored energy and the stored energy is considered in the case of communication load per hour during discharging mini And mu loadi The relation of (2) is:
S mini ＝0.5*(S 0 -Soc min *6)
wherein S is 0 Is the initial charge capacity of energy storage, soc min ＝μ weai *min{μ loadi ，Soc down }，Soc down Is a lower limit value for avoiding overdischarge, an upper limit S for storing energy and electric power max Take 95% of the battery capacity.
Further, in the step 10, the optimal scheduling objective function is the sum of the running cost of the photovoltaic unit, the fan unit and the energy storage unit and the electricity purchasing cost of the micro-grid to the power distribution network; constraint conditions considered by the power supply system optimization scheduling model comprise active power balance constraint, power constraint allowing interaction between a micro-grid and a power distribution network and energy storage unit operation constraint.
Further, in the step 11, a genetic algorithm is used to solve an optimization model; the genetic algorithm firstly generates an initial population with a random scale of N, and then calculates the fitness according to a set fitness function (namely the total power supply cost for 24 hours of the base in the future set in the step 10); the fitness function is set as the total power supply cost based on 24 hours in the future, and selection, crossover and variation are carried out on the basis of obtaining the fitness, so that a population of the next generation representing a power supply strategy is generated until the average change of the fitness function value is smaller than Tolfun (for the preset precision in the optimization function, the relative function value of the fitness function value change is smaller than the preset precision to indicate the optimization is finished) or the iteration number is larger than the maximum value, and the most economical power supply scheme is obtained.
The beneficial effects of the invention are as follows: on the basis of guaranteeing the communication requirement of the 5G base station, the economic effect of standby energy storage and peak shaving of the base station is exerted, and meanwhile, the waste of wind and light clean renewable energy sources and electric energy is avoided.
Drawings
Fig. 1 is a flowchart of a 5G green base station power supply optimization method in an embodiment of the present invention.
Fig. 2 is a schematic diagram of a 5G base station power supply system according to an embodiment of the present invention.
FIG. 3 is a converging curve based on a genetic algorithm in an embodiment of the invention.
Fig. 4 is a schematic diagram of 24 hours of power supply of a 5G base station (for example, 50 base stations) on a certain day when MATLAB is operated in an embodiment of the present invention.
Detailed Description
The technical scheme of the invention is further described in detail below with reference to the attached drawings.
Examples of micro-grid for 5G base stations include fans, photovoltaics, energy storage units, etc., as shown in fig. 2. For the micro-grid system example, the specific steps for implementing the optimized scheduling by the method of the invention are as follows:
step (1): and counting operation history data of the micro-grid, and establishing cost functions of all controllable power supplies in the micro-grid.
The 24 hour fan operating cost function is described as follows:
the photovoltaic 24 hour operation cost function is described as follows:
here, P Fi Fan power, P, for the ith time period Si Is the photovoltaic power, ω, of the ith time period F Is the power generation price of the fan, omega S Is the price of photovoltaic power generation.
The cost function of the energy storage unit is described as follows:
here, P Ci Is the battery power, ω, for the i-th period C From charging to dischargingElectricity costs per kilowatt-hour.
And linear interpolation is used according to the energy consumption and communication load conditions of the 5G base station, so that a relation curve of the energy consumption and the communication load conditions is obtained.
Step (2): and predicting the power consumption of the base station in the future 24 hours according to historical data of past one-week operation of the 5G base station, and predicting by using an artificial neural network method.
The BP neural network is a multi-layer feedforward network, and consists of forward propagation and error back propagation. The topology structure of the device comprises 3 layers: input layer, output layer and hidden layer, wherein hidden layer can have more than or equal to 1. Neurons of each layer have an reachable path only with neurons of immediately adjacent layers, and neurons of the same layer do not correlate. A great deal of practical experience has shown that the BP network of two hidden layers can represent any function of any graph. The input propagates from the input layer, after the function processing of the hidden layer, the output process from the output layer is called forward propagation, the difference between the true value and the predicted value becomes an error, the network propagates from the output layer to the hidden layer to the input layer by reversely giving weight, and the error is equally divided to each layer unit.
The hidden layer of BP neural network of the prediction model is set as 6 layers, the adopted activation function is a unipolar S-shaped function, the learning rate is 0.01, and the training stop criterion is set when the mean square error reaches 0.00001 or the iteration times reach 10000.
And dividing the training set and the testing set of the selected data by 7:3, and carrying out normalization processing on the training data. After the synaptic weight and threshold matrix of the initial network are set, forward propagation and error reverse propagation calculation of data are carried out, the weight is updated, iteration is carried out, and forward propagation and error reverse propagation are repeated by using new samples until the stopping criterion is met. And obtaining an ideal training BP neural network model. And finally, carrying in the historical power consumption data of the 5G base station, and predicting the power consumption of the base station for 24 hours in the future.
Step (3): obtaining a load rate index mu of each hour base station according to 24-hour power consumption prediction data of the 5G base station and a relation curve of energy consumption and communication load loadi . According to the 5G base station power consumption and the communication base station load obtained in the step (1)Rate index mu loadi The load rate index mu of each hour base is deduced loadi ，μ loadi The size is according to the following grading rule: mu (mu) loadi 1 at full load and 0.5 at 50%.
Step (4): let the original sequence of four meteorological indexes be x i (k)′。x i (k) ' represents the kth value of the ith meteorological factor, the first factor is the air pollution index, x 1 (1) ' represents the value of the air pollution index at the first hour, x 1 (2) ' is the value of the second hour, x 2 (1) ' is the value of the temperature in the first hour, x 3 (1) ' is the value of cloud in the first hour, x 4 (1) ' is the value of humidity at the first hour, and so on. Because the meteorological factors considered by the invention are indexes of different things, the numerical values can be greatly different due to different dimensionalities, and the dimensionless meteorological factors need to be subjected to dimensionless treatment, and the sequence after dimensionalization is x i (k) A. The invention relates to a method for producing a fibre-reinforced plastic composite Dimensionless, namely normalization operation, and comprises the following specific processes: the normalization adopts an initial value method, the data of a certain index is uniformly divided by the first data of the index, and the values can be sorted to be about 1 order of magnitude by dividing by the initial value.
Step (5): taking photovoltaic power generation as an example, the parent sequence is the output of 24 hours of photovoltaic power, and the child sequence is the numerical value of 24 hours normalized by four meteorological factors. Here, x i (k) The kth value representing the ith meteorological factor, the first factor being the air pollution index, x 1 (1) Indicating the value of the air pollution index in the first hour, x 1 (2) Is the value of the second hour, x 2 (1) Is the value of the temperature in the first hour, x 3 (1) Is the value of cloud in the first hour, x 4 (1) Is the value of humidity in the first hour, and so on. The invention uses x 0 (k) The parent sequence and the child sequence i.gtoreq.1, i.e.the sequence of the element to be analyzed.
Step (6): the weight given to the photovoltaic and the wind turbine is gamma according to the wind-solar output ratio per hour 1 ，γ 2 Wherein gamma is 1 、γ 2 ∈[0，1]，γ 1 +γ 2 ＝1。
Step (7): taking photovoltaic power generation as an example, according to a gray correlation coefficient formula (wherein ρ is an adjustable coefficient and is (0, 1) for adjusting the difference of the output results, and in this embodiment 0.5) for calculating the gray correlation coefficient and the different weather factors ζ i (k) Mean value zeta of 24-hour correlation coefficient of (2) i (i=1, 2,3, 4), and finally obtaining the influence weight alpha of four meteorological factors of air pollution index, temperature, cloud cover and humidity on photovoltaic power generation 1 、α 2 、α 3 、α 4 Likewise replace the parent sequence x 0 (k) For the output of the fan for 24 hours, the influence weight beta of four indexes on the generating capacity of the fan is calculated 1 、β 2 、β 3 、β 4 . Here, taking the calculation of 4 parameters for the influence weight of photovoltaic power generation as an example, +.>β i And the same is true. )
Step (8): the four meteorological data of the air pollution index, the temperature, the cloud cover and the humidity generated by the photovoltaic power generation and the fan are processed in a forward and normalization way by combining the practical meaning to obtain a matrix Z mn (m=1, 2, 3..24, n=1, 2,3,4,5,6,7, 8), the first four columns being photovoltaic power generation correlation matrices and the last four columns being fan power generation correlation matrices.
In combination with practical significance, the same index is an index in different directions for photovoltaic power generation and fan power generation, and forward processing is carried out on the forward index:reverse index is reversely processedAnd then to Z mn ' normalization processing to obtain Z mn The normalization adopts an initial value method, the data of a certain index is uniformly divided by the first data of the index, the values can be arranged to be about 1 order by dividing by the initial value), the step (6) and the step (7) are combined, and the total influence factors of four weather factors in the ith hour on new energy power generation are calculated>
Step (9): combining 5G base station hourly load rate index mu loadi Total influence coefficient mu of four meteorological factors on new energy power generation weai And (5) making upper and lower limits of energy storage and electric power storage. Here, because the standby energy storage of a single 5G base station holds about 19.2kwh of electricity, and the electricity consumption peak value of the 5G base station is about 3kwh, a flexible scheduling space of 6kwh is planned, the peak electricity consumption of about four hours is reserved, and the lower limit S of the energy storage and the electricity storage under the condition of communication load per hour is considered during discharging mini And mu loadi The relation of (2) is:
S mini ＝0.5*(S 0 -Soc min *6)
wherein S is 0 Is the initial charge capacity of energy storage, soc min ＝μ weai *min{μ loadi， Soc down }，Soc down Is a lower limit value for avoiding overdischarge, an upper limit S for storing energy and electric power max Take 95% of the battery capacity.
Step (10):
1) The total power supply cost target expression of the whole day is:
min W＝W C +W N +W F +W S
here, W is C For 24 hours energy storage operation cost, W N For total exchange cost of all-day micro-grid and power grid, W F For 24 hours running cost of the fan, W S For a photovoltaic 24-hour operating cost,wherein W is N The expression of (2) is:
here, ω Gi Is the electricity purchase price in the ith time period omega Mi Is the electricity selling price of the ith time period. In the formula, if electricity is sold to a commercial network, x is i =0; if electricity is purchased from the commercial network, x is i ＝1。
Here, P Ni Is the power exchanged with the main network in the ith time period, P Li Is the load power of the i-th period.
2) Active power balance constraint:
P Li +P Ci ＝P Ni +P Fi +P Si
3) Energy storage charge-discharge power constraint:
-1.2≤P Ci ≤1.2
4) Constraint of exchange power of micro-grid and distribution network:
-30≤P Ni ≤30
5) Energy storage and storage capacity constraint:
S mini ≤SP Ci ≤S max
SP Ci the whole is a variable representing the battery capacity (kVA) at the ith time.
Step (11): and solving an optimization model by using a genetic algorithm. The optimization algorithm flow chart is shown in fig. 1. The genetic algorithm firstly generates an initial population with a random scale of N, and then calculates the fitness according to a set fitness function (the fitness function is the total power supply cost of the base station in the future 24 hours set in the step (10)). And selecting, crossing and mutating on the basis of obtaining the fitness, generating a population of the next generation representing the power supply strategy until the average change of the fitness function value is smaller than Tolfun (Tolfun is the preset precision in the optimization function, the relative function value of the fitness function value change is smaller than the preset precision to indicate that the optimization is finished) or the iteration number is larger than the maximum value, and obtaining the most economical power supply scheme. As shown in fig. 3 and fig. 4, fig. 3 is a convergence curve of the genetic algorithm, fig. 4 is a power supply composition per hour for 24 hours in the future, in which the vertical axis of the daily load curve is kw, which represents the energy consumption requirement of the 5G base station per hour, the vertical axis of the GRID operation plan is kw, which represents the interaction between the power supply system and the mains supply, the vertical axis of the BA operation plan is kw, which represents the charge and discharge power per hour, the vertical axis of the BA energy storage condition is kVA, and which represents the electric quantity holding condition of the energy storage free dominant part of all the 5G base stations per hour.
The above description is merely of preferred embodiments of the present invention, and the scope of the present invention is not limited to the above embodiments, but all equivalent modifications or variations according to the present disclosure will be within the scope of the claims.
Claims (10)
1. A5G green base station power supply optimization method considering load and meteorological influence is characterized in that: comprises the following steps:
step 1, statistics of operation history data of a 5G base station, fitting of a relation curve of energy consumption and communication load of the 5G base station, and establishment of cost functions of different power supply modes;
step 2, predicting the energy consumption of a future day by using an artificial neural network method based on a time sequence according to 24-hour energy consumption historical data of the 5G base station to obtain a prediction result;
step 3, obtaining a load rate index mu of each hour based on the predicted data of 24 hours power consumption of the 5G base station and the relation curve of energy consumption and communication load loadi ；
Step 4, collecting four meteorological data of air pollution index, temperature, cloud cover and humidity and photovoltaic and fan power generation data, and carrying out normalization processing on the data by adopting initialization;
step 5, establishing a mother sequence and a child sequence of photovoltaic power generation and fan power generation required by gray correlation analysis;
step 6, giving weights of the photovoltaic and the wind turbine to be gamma according to the wind-solar output ratio of each hour 1 ，γ 2 Wherein gamma is 1 、γ 2 ∈[0，1]，γ 1 +γ 2 ＝1；
Step 7, forming a correlation sequence by calculating a gray correlation coefficient and a correlation coefficient mean value, and finally obtaining an influence weight alpha of four meteorological factors including air pollution index, temperature, cloud cover and humidity on photovoltaic power generation 1 、α 2 、α 3 、α 4 And the influence weight beta on the generating capacity of the fan 1 、β 2 、β 3 、β 4 ；
Step 8, carrying out forward normalization processing on four meteorological data of air pollution index, temperature, cloud cover and humidity generated by photovoltaic power generation and fan in a practical sense to obtain a matrix Z ij (i=1, 2,3 … 24, j=1, 2,3,4,5,6,7, 8), and calculating the total influence coefficient mu of four meteorological factors on new energy power generation per hour by combining the step 6 and the step 7 weai ；
Step 9, calculating the lower limit S of the energy storage electric energy holding capacity of the 5G base station according to the step 3 and the step 8, wherein the lower limit S is calculated according to the communication load and four meteorological factors mini ；
Step 10, according to an optimal scheduling period, taking the minimum total power supply cost of the whole day as an optimal scheduling target, comprehensively considering power balance constraint, energy storage charge-discharge power constraint, constraint that a micro-grid consisting of a 5G base station group, a fan, photovoltaic and energy storage exchanges power with a power distribution network, constraint of energy storage capacity, and combining the data obtained in the step 2 and the step 9, and establishing a 5G base station power supply system optimal scheduling model based on flexible scheduling of 5G energy storage;
and 11, solving the 5G base station power supply system optimization scheduling model obtained in the step 10, and calculating the power supply composition of the load in each period and the power supply cost of the whole day.
2. A 5G green base station power optimization method taking into account load and weather effects as defined in claim 1, wherein: in the step 1, the cost functions of different power supply modes include a fan operation cost function, a photovoltaic operation cost function and an energy storage unit cost function.
3. A 5G green base station power optimization method taking into account load and weather effects as defined in claim 1, wherein: in the step 2, after the synaptic weight and the threshold matrix of the initial network are set, forward propagation and error reverse propagation of data are calculated, the weight is updated, iteration is carried out, forward propagation and error reverse propagation are repeatedly carried out by using a new sample until a stopping criterion is met, an ideal training BP neural network model is obtained, finally, historical power consumption data of a 5G base station are brought in, and base station power consumption for 24 hours in the future is predicted.
4. A 5G green base station power optimization method taking into account load and weather effects as defined in claim 1, wherein: in the step 3, the base station load rate index mu is calculated every hour loadi ∈[0，1]The larger the communication load, μ loadi The larger.
5. A 5G green base station power optimization method taking into account load and weather effects as defined in claim 1, wherein: in the step 5, the parent sequence is the output of the photovoltaic and fan for 24 hours, and the child sequence is the numerical value of the four weather factors normalized for 24 hours; here, x i (k) The value of the kth hour representing the ith meteorological factor is 1 for air pollution i, 2 for temperature i, 3 for cloud i, and 4 for humidity i.
6. A 5G green base station power optimization method taking into account load and weather effects as defined in claim 1, wherein: in the step 6, gamma 1 ，γ 2 The output of the fan is respectively proportional to the output of the photovoltaic and the fan.
7. A 5G green base station power optimization method taking into account load and weather effects as defined in claim 1, wherein: in the step 7, the gray correlation coefficient is calculatedAndSolving and calculating grey correlation coefficient and different meteorological factors zeta i (k) Mean value zeta of 24-hour correlation coefficient of (2) i I=1, 2,3,4, ρ is an adjustment coefficient, the value is (0, 1), and the influence weight alpha of four meteorological factors including air pollution index, temperature, cloud cover and humidity on the photovoltaic power generation amount is obtained based on the gray correlation coefficient 1 、α 2 、α 3 、α 4 Likewise replace the parent sequence x 0 (k) For the output of the fan for 24 hours, the influence weight beta of four indexes on the generating capacity of the fan is calculated 1 、β 2 、β 3 、β 4 。
8. A 5G green base station power optimization method taking into account load and weather effects as defined in claim 1, wherein: in the step 9, the lower limit S of the stored energy and the stored energy is considered under the condition of communication load per hour during discharging mini And mu loadi The relation of (2) is:
S mini ＝0.5*(S 0 -Soc min *6)
wherein S is 0 Is the initial charge capacity of energy storage, soc min ＝μ weai *min{μ loadi ，Soc down }，SoC down Is a lower limit value for avoiding overdischarge, an upper limit S for storing energy and electric power max Take 95% of the battery capacity.
9. A 5G green base station power optimization method taking into account load and weather effects as defined in claim 1, wherein: in the step 10, the optimal scheduling objective function is the sum of the running cost of the photovoltaic unit, the fan unit and the energy storage unit and the electricity purchasing cost of the micro-grid to the power distribution network; constraint conditions considered by the power supply system optimization scheduling model comprise active power balance constraint, power constraint allowing interaction between a micro-grid and a power distribution network and energy storage unit operation constraint.
10. A 5G green base station power optimization method taking into account load and weather effects as defined in claim 1, wherein: in the step 11, a genetic algorithm is used for solving an optimization model; firstly, generating an initial population with a random scale of N by a genetic algorithm, and then calculating the fitness according to the total power supply cost of the base station within 24 hours in the future set in the step 10; the fitness function is set to be the total power supply cost based on 24 hours in the future, selection, crossover and variation are carried out on the basis of obtaining the fitness, a next generation of population representing a power supply strategy is generated until the average change of the fitness function value is smaller than the preset precision Tolfun or the iteration number is larger than the maximum value, and the most economical power supply scheme is obtained.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210300788.2A CN114679344B (en) | 2022-03-25 | 2022-03-25 | 5G green base station power supply optimization method considering load and meteorological influence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210300788.2A CN114679344B (en) | 2022-03-25 | 2022-03-25 | 5G green base station power supply optimization method considering load and meteorological influence |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114679344A CN114679344A (en) | 2022-06-28 |
| CN114679344B true CN114679344B (en) | 2023-09-29 |
Family
ID=82074980
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210300788.2A Active CN114679344B (en) | 2022-03-25 | 2022-03-25 | 5G green base station power supply optimization method considering load and meteorological influence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114679344B (en) |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112968456A (en) * | 2021-03-01 | 2021-06-15 | 华北电力大学 | 5G base station energy storage regulation and control method considering base station communication load state |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111275240B (en) * | 2019-12-27 | 2023-06-09 | 华北电力大学 | Load prediction method based on multi-energy coupling scene |
-
2022
- 2022-03-25 CN CN202210300788.2A patent/CN114679344B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112968456A (en) * | 2021-03-01 | 2021-06-15 | 华北电力大学 | 5G base station energy storage regulation and control method considering base station communication load state |
Non-Patent Citations (1)
| Title |
|---|
| 光储系统储能电站的经济优化调度策略;顾飞;黄陈蓉;张建德;王鑫琪;;电工技术(第14期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114679344A (en) | 2022-06-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2023274425A1 (en) | Multi-energy capacity optimization configuration method for wind-solar-water-fire storage system | |
| CN111030188B (en) | Hierarchical control strategy containing distributed and energy storage | |
| CN109190802B (en) | Multi-microgrid game optimization method based on power generation prediction in cloud energy storage environment | |
| CN111401604B (en) | Power system load power prediction method and energy storage power station power distribution method | |
| CN113437752B (en) | Operation control method of comprehensive energy system containing hybrid energy storage | |
| Hong et al. | Interactive multi-objective active power scheduling considering uncertain renewable energies using adaptive chaos clonal evolutionary programming | |
| CN108197766A (en) | A kind of active distribution network Optimal Operation Model for including micro-capacitance sensor group | |
| CN116667325B (en) | Micro-grid-connected operation optimization scheduling method based on improved cuckoo algorithm | |
| CN112508325A (en) | Multi-time-scale electric energy scheduling method for household micro-grid | |
| CN115374692B (en) | Double-layer optimization scheduling decision method for regional comprehensive energy system | |
| CN203933038U (en) | From the grid-connected mixing photovoltaic power generation control system of net | |
| CN115114854A (en) | Two-stage self-organizing optimization aggregation method and system for distributed resources of virtual power plant | |
| CN112836849A (en) | Virtual power plant scheduling method considering wind power uncertainty | |
| CN117748622A (en) | Micro-grid polymorphic coordination control method and system | |
| CN107609690A (en) | A kind of method of load active management decision optimization | |
| CN114679344B (en) | 5G green base station power supply optimization method considering load and meteorological influence | |
| CN111178642A (en) | Micro-grid optimization method based on simulated annealing particle swarm algorithm | |
| CN116799856A (en) | Energy control method, controller and system for multi-microgrid system | |
| Perera et al. | Supply and demand planning of electricity power: a comprehensive solution | |
| CN114707704A (en) | Electric vehicle charging station microgrid energy management method, system and platform | |
| Kai et al. | Optimization for PV-ESS in Distribution Network Based on CSBO | |
| CN111932401A (en) | Multi-party trust service interaction method based on block chain | |
| CN112270440B (en) | Power distribution network load prediction method based on capsule neural network | |
| CN118249420B (en) | Light storage direct-soft system power scheduling method and device based on model prediction | |
| Chinnadurrai et al. | Energy Management of a Microgrid based on LSTM Deep Learning Prediction Model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |