Distribution generation (DG) technology based on a variety of renewable energy technologies has developed rapidly. A large number of multitype DG are connected to the distribution network (DN), resulting in a decline in the stability of DN operation. It is urgent to find a method that can effectively connect multienergy DG to DN. photovoltaic (PV), wind power generation (WPG), fuel cell (FC), and micro gas turbine (MGT) are considered in this paper. A multiobjective optimization model was established based on the life cycle cost (LCC) of DG, voltage quality, voltage fluctuation, system network loss, power deviation of the tieline, DG pollution emission index, and meteorological index weight of DN. Multiobjective artificial bee colony algorithm (MOABC) was used to determine the optimal location and capacity of the four kinds of DG access DN, and compared with the other three heuristic algorithms. Simulation tests based on IEEE 33 test node and IEEE 69 test node show that in IEEE 33 test node, the total voltage deviation, voltage fluctuation, and system network loss of DN decreased by 49.67%, 7.47% and 48.12%, respectively, compared with that without DG configuration. In the IEEE 69 test node, the total voltage deviation, voltage fluctuation and system network loss of DN in the MOABC configuration scheme decreased by 54.98%, 35.93% and 75.17%, respectively, compared with that without DG configuration, indicating that MOABC can reasonably plan the capacity and location of DG. Achieve the maximum tradeoff between DG economy and DN operation stability.
In recent years, the awareness of environmental protection has gradually gained popularity. The proportion of distributed generation (DG), which is dominated by renewable energy sources such as photovoltaic (PV) and wind power generation (WPG), is gradually increasing. Reasonable access to DG can improve the voltage distribution of DN by changing the power flow distribution of the distribution network (DN), reducing the system’s active network loss, and reducing voltage fluctuations [
Aiming at the research on improving the stability of DN by accessing DG, Reference [
The above studies are all singleobjective planning models of DG. However, the location and capacity determination of DG needs to achieve the optimal balance between the economy of DG investment and the stability of DN. The traditional singleobjective model cannot achieve the optimal balance between the economy of investors and the stability of DN. Reference [
All the above studies have made some contributions to the field of DG planning, but there are still shortcomings. This paper proposes a new method for DG siting capacity determination, and its main contributions are as follows:
(1) In this paper, four typical DG types such as PV, WPG, fuel cell (FC) [
(2) In this paper, the multiobjective artificial bee colony algorithm (MOABC) based on Pareto is adopted to solve the DG location fixedcapacity model, and MODA, MOGOA and MOPSO were used as comparison algorithms. The results show that MOABC has better optimization ability, and stability, and can obtain the optimal DG location capacity determination scheme.
(3) In this paper, an improved grey target decision scheme based on the entropy weight method is adopted, which can effectively solve the weight subjectivity of multiobjective results obtained from singleobjective weighting, making the model solution results more objective and effective.
This article considers four types of DGs: PV, WPG, FC, and MGT.
where
where
From
The capacity determination problem of DG location is a multidimensional, multiconstraint and multiobjective optimization problem. In this paper, the life cycle cost (LCC), voltage deviation of DN, voltage fluctuation, network loss minimization, minimum tieline power deviation, meteorological index weight and pollution emission index of four types of DG are used as objective functions to establish a multiobjective optimization model. The reason for choosing the above objective function is that when DG accesses DN, whether the construction of DG is economical and feasible should be considered first. Therefore, this paper takes the LCC of DG as one of the objective functions. Secondly, when DG is connected, the stable operation of DN is particularly important. Therefore, the objective function is to minimize the voltage deviation, voltage fluctuation and network loss of DN, and minimize the power deviation of the tieline. Finally, it is necessary to consider the degree of environmental pollution after some types of DG (FC and MGT) are put into operation, so the pollution emission index is taken as one of the objective functions. Since the amount of PV and WPG power generation output largely depends on meteorological conditions, to install wind turbines and photovoltaic systems in areas rich in wind and light resources, maximize the absorption of scenic energy and make full use of renewable energy, this paper innovatively sets meteorological indicators as the objective function.
LCC is the sum of all costs in the entire life cycle of a product. The whole life cycle cost is a management strategy with the whole life cycle cost theory as the core. The calculation of LCC is usually carried out every year, which is 365 days [
As an economic indicator for evaluating DG LCC mainly includes initial investment cost (IIC), maintenance cost (MC), and recovery cost (RC). The LCC of DG is calculated as follows:
where
(1) Initial Investment Cost
where
(2) Maintenance Cost
where
(3) Recovery Cost
where
(1) Voltage deviation index
To ensure that DG can effectively improve the voltage distribution of DN, this paper considers minimizing the voltage deviation of each node in DN as the objective function. The voltage deviation index can be described by the following formula [
where
(2) Voltage fluctuation index
Due to the changes in the distribution of power flow in DN after DG is connected, the voltage fluctuation of DN is significantly increased. Therefore, the standard deviation of voltage within one day is selected to define the voltage fluctuation of DN, as follows:
where
(3) System network loss index
where
This article introduces pollution emission indicators that include carbon dioxide, sulfur dioxide, and nitrogen oxides to measure the total amount of pollutants emitted by DG during operation time in light of countries worldwide committed to building a lowcarbon society and using green energy more efficiently to decrease pollution.
where
DG unit  

FC  0.502  0.5216  
MGT  3.445  
PV  –  –  – 
WPG  –  –  – 
As can be seen from the above table, photovoltaic systems and wind turbines in addition to the construction period will produce pollution emissions, after the completion of power generation will not produce pollution emissions, belong to clean electricity. However, fuel cells and microgas turbines not only produce pollution emissions during construction, but also emit a certain amount of pollution emissions in the environment after the completion of power generation. Therefore, this paper takes the pollution emission index as one of the objective functions, to maximize environmental friendliness.
This paper proposes an objective function considering the annual average wind speed
where
Due to the intermittent nature of the new energy output, large power fluctuations will occur when it is connected to the grid. In this paper, the power stability of the grid is considered in the DG sizing and capacity planning, which is expressed in terms of the daily power deviation of the power tieline as follows:
where
(1) Transmission line power constraint.
where
(2) Node voltage constraint.
where
(3) DG configuration power constraint.
where
(4) Node power balance.
where
Let ABC be presented as a multiobjective intelligent optimization algorithm to mimic the process of bees collecting nectar in nature. It consists of a food source, leading bees, and follower bees. The solution process is as follows:
(1) The population is initialized based on upper and lower limits, followed by the calculation of initial fitness function values.
(2) The leading bees will continuously update their food sources to ensure the freshness of their food. After updating the food source, bees will update the fitness values and optimal food source according to the new food source. The process of updating the food source is as follows:
where
(3) After updating the food source, the leading bees will share information with the follower bees. The follower bees will then allocate to the food sources.
(4) The bee colony will allocate food based on the information provided by the follower bees, which can be calculated as
where
(5) If the food source falls into a local optimum, it will be abandoned. The scout bees will then generate a new food source based on
where
(6) Repeat steps (2) to (5) until the end of the cycle to obtain the optimal food source and fitness value.
Multiobjective optimization problems are different from singleobjective ones, as it is impossible to obtain a solution that minimizes all objectives at the same time, only a set of Pareto optimal solutions can be obtained [
(1) An external archive is added to MOABC to store nondominated solutions, with capacity limited to a fixed number.
(2) A solution update mechanism based on Pareto nondominated ranking.
(3) Calculate the crowding distance of all populations, and rank populations with the same level based on the size of the crowding distance.
In order to avoid the influence of subjective decision on the final result, the grey target decision scheme based on the entropy weight method (EWM) [
(1) Build a sample matrix
The normalized fitness function F of all the solutions is taken as one of the evaluation indexes, and the sample matrix is established. In this paper, we consider adding two related indexes to the sample matrix, one is the Euclidean distance (ED)
In the formula,
(2) Computational bullseye
The operator
where
The decision matrix
The selected bullseye is
(3) Establish the weight and Mahalanobis distance
Based on EWM, the weights of each evaluation index can be obtained objectively, and the optimal compromise solution can be selected from Pareto nondominated solution set. The weights
Each MD to the bullseye can be expressed as:
The nondominated solutions are sorted according to MD. Each set of solutions in the archive set is considered an independent decision scheme. The solution closest to the bullseye is chosen as the optimal decision solution.
In order to verify the effectiveness of the multiobjective optimization algorithm proposed in this paper [
MOABC  

Maximum iterations  100 
Population size  100 
Quantity of food resources  50 
Maximum number of updates of food resources  5 
Quantity of objective function  7 
Archive size  100 
Optimization problem dimension  13 
Node  Node  Branch impedance  Load  Node  Node  Branch impedance  Load 

0  1  0 + j0  0 + j0  17  18  0.7320 + j0.5740  90 + j40 
1  2  0.0922 + j0.047  100 + j60  2  19  0.1640 + j0.1565  90 + j40 
2  3  0.4930 + j0.2511  90 + j80  19  20  1.5042 + j1.3554  90 + j40 
3  4  0.3660 + j0.1864  120 + j80  20  21  0.4095 + j0.4784  90 + j40 
4  5  0.3811 + j0.1941  60 + j30  21  22  0.7089 + j0.9373  90 + j40 
5  6  0.8190 + j0.7070  60 + j20  3  23  0.4512 + j0.3083  90 + j50 
6  7  0.1872 + j0.6188  200 + j100  23  24  0.8980 + j0.7091  420 + j200 
7  8  0.7114 + j0.2351  200 + j100  24  25  0.8960 + j0.7011  420 + j200 
8  9  1.0300 + j0.7400  60 + j20  6  26  0.2030 + j0.1034  60 + j25 
9  10  1.0440 + j0.7400  60 + j20  26  27  0.2842 + j0.1447  60 + j25 
10  11  0.1966 + j0.0650  45 + j30  27  28  1.0590 + j0.9337  60 + j20 
11  12  0.3744 + j0.1238  60 + j35  28  29  0.8042 + j0.7006  120 + j70 
12  13  1.4680 + j1.1550  60 + j35  29  30  0.5075 + j0.2585  200 + j600 
13  14  0.5416 + j0.7129  120 + j80  30  31  0.9744 + j0.9630  150 + j70 
14  15  0.5910 + j0.5260  60 + j10  31  32  0.3105 + j0.3619  210 + j100 
15  16  0.7463 + j0.5450  60 + j20  32  33  0.3410 + j0.5362  60 + j40 
16  17  1.2890 + j1.7210  60 + j20  –  –  –  – 
Node  Node  Branch impedance  Load  Node  Node  Branch impedance  Load 

0  1  0 + j0  0 + j0  3  36  0.0044 + j0.0108  26 + j18.55 
1  2  0.0005 + j0.0012  0 + j0  36  37  0.064 + j0.1565  26 + j18.55 
2  3  0.0005 + j0.0012  0 + j0  37  38  0.1053 + j0.123  0 + j0 
3  4  0.0015 + j0.0036  0 + j0  38  39  0.0304 + j0.0355  24 + j17 
4  5  0.0251 + j0.0294  0 + j0  39  40  0.0018 + j0.0021  24 + j17 
5  6  0.366 + j0.1864  2.6 + j2.2  40  41  0.7283 + j0.8509  1.2 + j1 
6  7  0.3811 + j0.1941  40.4 + j30  41  42  0.31 + j0.3623  0 + j0 
7  8  0.0922 + j0.047  75 + j54  42  43  0.041 + j0.0478  6 + j4.3 
8  9  0.0493 + j0.0251  30 + j22  43  44  0.0092 + j0.0116  0 + j0 
9  10  0.819 + j0.2707  28 + j19  44  45  0.1089 + j0.1373  39.2 + j26.3 
10  11  0.1872 + j0.0619  145 + j104  45  46  0.0009 + j0.0012  39.2 + j26.3 
11  12  0.7114 + j0.2351  145 + j104  4  47  0.0034 + j0.0084  0 + j0 
12  13  1.03 + j0.34  8 + j5.5  47  48  0.0851 + j0.2083  79 + j56.4 
13  14  1.044 + j0.345  8 + j5.5  48  49  0.2898 + j0.7091  384.7 + j274.5 
14  15  1.058 + j0.3496  0 + j0  49  50  0.0822 + j0.2011  384.7 + j274.5 
15  16  0.1966 + j0.065  45.5 + j30  8  51  0.0928 + j0.0473  40.5 + j28.3 
16  17  0.3744 + j0.1238  60 + j35  51  52  0.3319 + j0.1114  3.6 + j2.7 
17  18  0.0047 + j0.0016  60 + j35  9  53  0.174 + j0.0886  4.35 + j3.5 
18  19  0.3276 + j0.1083  0 + j0  53  54  0.203 + j0.1034  26.4 + j19 
19  20  0.2106 + j0.069  1 + j0.6  54  55  0.2842 + j0.1447  24 + j17.2 
20  21  0.3416 + j0.1129  114 + j81  55  56  0.2813 + j0.1433  0 + j0 
21  22  0.014 + j0.0046  5 + j3.5  56  57  1.59 + j0.5337  0 + j0 
22  23  0.1591 + j0.0526  0 + j0  57  58  0.7837 + j0.263  0 + j0 
23  24  0.3463 + j0.1145  28 + j20  58  59  0.3042 + j0.1006  100 + j72 
24  25  0.7488 + j0.2475  0 + j0  59  60  0.3861 + j0.1172  0 + j0 
25  26  0.3089 + j0.1021  14 + j10  60  61  0.5075 + j0.2585  1244 + j888 
26  27  0.1732 + j0.0572  14 + j10  61  62  0.0974 + j0.0496  32 + j23 
3  28  0.0044 + j0.0108  26 + j18.6  62  63  0.145 + j0.0738  0 + j0 
28  29  0.064 + j0.1565  26 + j18.6  63  64  0.7105 + j0.3619  227 + j162 
29  30  0.3978 + j0.1315  0 + j0  64  65  1.041 + j0.5302  59 + j42 
30  31  0.0702 + j0.0232  0 + j0  11  66  0.2012 + j0.0611  18 + j13 
31  32  0.351 + j0.116  0 + j0  66  67  0.0047 + j0.0014  18 + j13 
32  33  0.839 + j0.2816  14 + j10  12  68  0.7394 + j0.2444  28 + j20 
33  34  1.708 + j0.5646  19.5 + j14  68  69  0.0047 + j0.0016  28 + j20 
34  35  1.474 + j0.4873  6 + j4  –  –  –  – 
Base voltage/kV  12.66 
Baseline capacity/MVA  10 
Baseline impedance/kΩ  16.03 
PV unit capacity cost/(yuan·kW^{−1})  8000 
WPG unit capacity cost/(yuan·kW^{−1})  7500 
FC unit capacity cost/(yuan·kW^{−1})  7000 
MGT unit capacity cost/(yuan·kW^{−1})  7000 
Discount rate  5.5% 
The MOABC control parameters in the table above are most consistent with the research topic of this paper.
PV#1  PV#2  WPG#1  WPG#2  FC  MGT  

Upper limit/kW  700  750  450  500  300  500 
Lower limit/kW  30  25  20  10  10  50 
Algorithm  PV  WPG  FC  MGT  

#1 
#1 
#2 
#2 
#1 
#1 
#2 
#2 
Capa 
Install 
Capa 
Install 

MOABC  31  3  30  31  300  18  34  16  44  33  210  13 
MOGOA  531  28  56  32  425  23  157  8  20  12  300  15 
MOPSO  215  13  741  32  127  16  86  2  20  7  164  18 
MODA  295  27  310  32  162  2  214  10  97  12  324  31 
MOABC  MODA  MOGOA  MOPSO  No DG is configured  

LCC/yuan  –  
Total voltage deviation/p.u  38.42  41.22  45.83  66.20  
Voltage fluctuation/p.u  1.021  1.024  1.030  1.097  
System loss/kW  2107.10  2524.03  2789.88  4061.87  
Pollution discharge coefficient/kg/h  151.29  257.371  231.294  –  
Tieline power deviation/kW  1139.12  1147.56  1152.38  1209.19  
Weight of meteorological index  0.4854  0.5108  0.5075  – 
Algorithm  PV  WPG  FC  MGT  

#1 
#1 
#2 
#2 
#1 
#1 
#2 
#2 
Capa 
Install 
Capa 
Install 

MOABC  593  68  600  61  30  51  132  46  30  26  217  64 
MOGOA  427  56  579  61  144  46  173  9  98  59  68  65 
MOPSO  653  2  28  61  176  15  300  10  20  7  149  65 
MODA  311  16  440  61  166  50  117  18  53  67  132  68 
MOABC  MODA  MOGOA  MOPSO  No DG is configured  

LCC/yuan  –  
Total voltage deviation/p.u  38.94  39.78  36.50  71.39  
Voltage fluctuation/p.u  0.773  0.761  0.760  1.183  
System loss/kW  1758.04  1430.87  1420.43  4449.99  
Pollution discharge coefficient/kg/h  144.514  115.41  98.344  –  
Tieline power deviation/kW  834.994  825.935  821.28  1241.85  
Weight of meteorological index  0.7730  0.6008  0.5683  – 
Considering the operation stability of DN, the investment economy of DG, the environmental friendliness and the full utilization of scenic resources, this paper carries out the capacity allocation location of four typical DG based on MOABC. Based on the improved grey target decision, the optimal compromise solution is obtained. The main conclusions are as follows:
1) This paper establishes a multiobjective programming model based on DG LCC, pollution emission index, meteorological index weight and DN voltage fluctuation, voltage distribution, network loss and tieline power deviation. It can achieve the maximum improvement of DN stability and economy on the premise of ensuring the interests of DG investors.
2) In this paper, an improved grey target decision scheme based on Euclidean distance and Mahalanobis distance is adopted, which can effectively avoid the influence of subjective decision on the optimal compromise solution.
3) In this paper, a simulation experiment based on IEEE 33 standard node test system is designed to verify the validity of the proposed configuration method. The experimental results show that the total voltage deviation, voltage fluctuation and system network loss of DN decreased by 49.67%, 7.47% and 48.12% respectively after the DG configured by MOABC. The stability of DN operation is effectively improved.
4) In this paper, a simulation experiment based on IEEE 69 standard node test system is designed to verify the validity of the proposed configuration method. The experimental results show that the total voltage deviation, voltage fluctuation and system network loss of DN decreased by 54.98%, 35.93% and 75.17% respectively after the DG configured by MOABC. The stability of DN operation is effectively improved.
In this paper, based on IEEE 33 and IEEE 69 distribution network structure, MOABC is used to locate and determine the capacity of four types of DG. After the optimized configuration of DG is connected to DN, the total voltage deviation and voltage fluctuation of DN are significantly reduced, which is conducive to the more stable and economical operation of DN. It shows that the distribution of distributed generation is crucial to the success of grid planning [
Thank you very much for the support provided by State Grid Shaanxi Electric Power Company and Central Southern China Electric Power Design for this paper.
The authors received no specific funding for this study.
Qiangfei Cao: Conceptualization, Writingreviewing, and editing; Huilai Wang: Writingoriginal draft preparation, Investigation; Zijia Hui: Visualization and contribution to the discussion of the topic; Lingyun Chen: Validation.
All data are from a region in Southwest China, and the data used cannot be disclosed according to data confidentiality requirements.
The authors declare that they have no conflicts of interest to report regarding the present study.