Optimal Dispatching of Distributed Generation Cluster Systems Considering Multiply Optimization Objectives

In the trend of increasing energy shortages and environmental pollution, integrating clean micro sources as cluster systems has become the mainstream trend in the DG cluster system research field. In an effort to ensure the economy and environmental friendliness of Distributed Generation (DG) cluster systems, a multi-objective optimization scheduling model is established to minimize the operating and environmental costs of DG cluster systems. Based on a multi-objective particle swarm optimization algorithm, this paper conducts a study on the capacity optimal configuration of wind turbines, photovoltaic generators, and energy storage devices. The model proposed in this paper is applied to analyse a DG cluster system in a rural area dominated by agriculture. The results show that the proposed model has a significant effect in reducing the operating and environmental costs of the DG cluster system.


Introduction
The dependence on fossil fuels and environmental pollution have led traditional power systems to face increasingly severe conflicts between resources and the environment.Therefore, the research on flexible and efficient Distributed Generation (DG) cluster systems containing distributed energy is receiving widespread attention from the power industry.Nowadays, a lot of attention has been focused on the study of economic, environmental, and reliability indicators of DG cluster systems [1][2][3].Especially, the optimization and scheduling of DG cluster systems that balance economy and environment is currently the mainstream research direction [4][5][6].
The uncertainty of DG cluster systems has greatly increased because of the fluctuating and intermittent characteristics of distributed wind and photovoltaic energy generation [7].These shortcomings will be detrimental to the safety and reliability of power grid operation [8][9].Power grids are usually equipped with a certain capacity of energy storage devices to improve their reliability and promote the consumption of distributed energy [10].How to reasonably configure the capacity of wind power, photovoltaic and energy storage devices in DG cluster systems is an important subject worth studying now.
To solve the problems mentioned above, this paper sets operation cost and environmental cost of DG cluster systems as the optimization goals, taking the characteristics of load and distributed energy generation into account, to achieve the optimal configuration and scheduling of photovoltaic, wind power and energy storage in DG cluster systems.

DG cluster system model
The grid-connected DG cluster system model in this paper is composed of a diesel generator set, an energy storage system, a wind driven generator set and a photovoltaic generator set. Figure 1 shows the topology structure of the DG cluster system.DG cluster system can purchase electricity from the main grid to maintain power balance, or sell electricity to generate revenue.

Mathematical model of energy storage
The state of charge value of the energy storage system can be expressed as, where S OC (t) and S OC (t 0 ) represent the state of charge value of the energy storage system at the time of t and t 0 , P BESS represents the charge and discharge power of the energy storage device, which is positive when charging and negative when discharging, η denotes the charging and discharging efficiency and Δt refers to the sampling time period.

Mathematical model of wind-driven generator
The mathematical model of a wind generator can be expressed as, 0 where v in , v out and v r are respectively the cut-in, cut-out and rated wind speeds of the wind turbine which can only generate electricity normally when the wind speed is between the cut-in and cut-out wind speeds, and P r represents the rated output power.

Mathematical model of photovoltaic power generation device
The mathematical model of a photovoltaic power generation device is as,

Objective function
Considering the economic and environmental factors of DG cluster system operation, a multiobjective optimal scheduling model has been built to minimize the operating and environmental cost of the DG cluster system.

Operating cost.
The expression of the operating cost of the DG cluster system is as, where H 1 represents the cost caused by purchasing electricity from the main grid, H 2 refers to the fuel cost of power generation, mainly from the diesel generator set, H 3 represents the cost of operation and maintenance, H 4 denotes the depreciation cost of the energy storage equipment, H 5 represents the cost of equipment purchase, and H 6 refers to the cost of constraint deviation penalty.

Environmental cost.
Pollutant emission penalties for diesel generator sets and main grids are mainly considered.The expression of environmental cost for the DG cluster system can be expressed as, where λ DE is the pollutant penalty coefficient of the diesel generator set, Q DE is the power generated by the diesel generator set, λ grid is the penalty coefficient for main network pollutants and Q grid is the power output from the main network.

Power balance constraint.
The expression of the power balance constraint of the DG cluster system is shown as, where P load denotes total load power, P is the power of the diesel generator set, N represents the number of units, P WT , P PV and P ES are the power generation of the wind turbine, photovoltaic system, and energy storage system respectively, and P grid represents the interaction power of the main network.

Distributed power output constraint.
The distributed power output constraint of the DG cluster system can be expressed as, min max k k k P P P   ( 7 ) where P k represents the output of the k-th distributed power source, and P kmin and P kmax represent the lower and upper output limit of the k-th distributed power generation.

Diesel generator climbing constraint.
The expression of the diesel generator climbing constraint of the DG cluster system is as, where SOC represents the state of charge of the energy storage device at any time, and SOC min and SOC max denote the minimum and maximum state of charge limit for the energy storage device.

Simulation Parameters
The multi-objective particle swarm optimization (MOPSO) was adopted in this paper to solve the proposed model by taking economic and environmental benefits into account and the simulation is realized by MATLAB.In the MOPSO algorithm, the size of the particle swarm is 40, and the maximum iteration times and crossover rate are 100 and 0.9 respectively.The research object of the simulation is a DG cluster system of a rural area dominated by agriculture.Typical daily load characteristics are shown in Figure 2. A load of rural areas dominated by agriculture mainly consists of agricultural load and residential load.The latter has a significant peak in electricity consumption at night, while the former relies mainly on daytime power.System model parameters are set as follows: the environmental protection coefficient of the wind turbine and the main network are 0.98 and 1.83 respectively, in CNY/kWh.Equipment operation and maintenance coefficient of photovoltaic generator set, wind turbine and diesel generator set are 0.0296, 0.0096 and 0.128 respectively, in CNY/kWh.The energy storage depreciation coefficient is 0.026, in CNY/kWh.Table 1 shows the initial configuration of each output device.

Optimization results
Figure 3 shows the Pareto frontier of the MOPSO algorithm, which can be seen as a compromise to achieve a reduction in the operation and environmental cost of the DG cluster system.Figure 4 shows the optimization scheduling results.Figure 5 and Figure 6 illustrate the predicted wind and solar output.As shown in Figure 4, in the afternoon and evening peak hours of the load (14:00-16:00 and 18:00-20:00), the power is mainly provided by energy storage systems and diesel units to meet the load demand.In the morning peak hours (06:00-08:00), the power is mainly provided by diesel units and the main grid.During the flat period (10:00-13:00), the output of wind and photovoltaic power increases.In an effort to prevent the waste of wind and solar resources, the energy storage system charges at this time and supplies power to the DG cluster system during the evening peak hours of the load, reducing the amount of electricity purchased from the main grid.During the valley period (01:00-04:00), the wind turbine and main grid have a high output, so the energy storage system will collect electricity at this time.
It can be analysed from Figure 5 and Figure 6 that 1 wind turbine, 35 photovoltaic panels, and 17 energy storage devices need to be configured.The wind turbine capacity is 10.9581 kW, the photovoltaic capacity is 11.2841k W, and the scheduled battery capacity is 20.0496 kW.The output of the wind turbine, photovoltaic, and energy storage are 73.05%,45.14%, and 66.83% of their installed capacity respectively.The capacity ratio of wind power, photovoltaic, and energy storage is 0.971:1:1.777.

Conclusion
This paper conducts a study on the optimal configuration and scheduling of photovoltaic, wind power and energy storage in DG cluster systems.Characteristics of rural load, distributed energy output and other factors are comprehensively considered in the research process.Based on the MOPSO algorithm, the optimization process is implemented to realize the minimization of operating and environmental costs.The optimization configuration results indicate that energy storage devices can reduce the waste of wind energy and photovoltaic energy, and have a certain effect on peak shaving and valley filling.The optimization scheduling results indicate that the multi-objective DG cluster system optimization scheduling model proposed in this article can fully consider the characteristics of photovoltaic power and wind energy, leverage energy storage advantages, and ultimately generate reasonable scheduling strategies.The optimization model proposed above can provide a reasonable reference for the research on the scheduling of DG cluster systems.

Figure 2 .
Figure 2. Load characteristics in a rural area dominated by agriculture.
DE,t and P DE,t-1 denote the climbing rate of the unit at the time of t and t-1, and ΔP DE,t represents the maximum climbing speed of the unit at time t.