Capacity optimization allocation method of photovoltaic-temperature difference-hydrogen-storage hybrid micro-energy system based on SAO algorithm

The capacity optimization allocation of hybrid micro-energy systems is an important link in micro-energy systems, which can effectively improve the reliability and economy of the power grid. In this paper, the capacity optimization allocation method of a photovoltaic-temperature difference-hydrogen-storage hybrid micro-energy system is studied. First, a capacity optimization configuration model based on photovoltaic, temperature difference, and hydrogen storage hybrid micro-energy system was established to maximize revenue. Then, the SAO algorithm is used to optimize its strategy. When the strategy does not change, it indicates that the income has been maximized. Finally, concerning the snow ablation mechanism, it can balance development and exploration well and can optimize the global optimal solution even in a more complex environment. The proposed method can ensure the reliability under premise of relatively low cost, and effectively improve the rationality of the power grid capacity allocation method in the photovoltaic thermal power generation scenario.


Introduction
With the arrival of the international energy transition trend, environment-friendly clean energy is developing rapidly.It was proposed to rapidly build new energy industries such as photovoltaic power generation, among which photovoltaic power generation is relatively mature, and new energy stations mainly photovoltaic are constantly built, and how to promote its consumption and efficiency utilization has become an urgent problem to be solved [1].
As an ideal clean energy, hydrogen production equipment has a long service life, simple raw material acquisition, hydrogen storage, and transportation are more convenient, providing a new optimal solution for the large-scale development, utilization, and consumption of photovoltaic.Hydrogen storage systems and supercapacitors are used for wind farms to reduce output fluctuations [2]; In [3], a battery was added to the wind-light-hydrogen coupling system to reduce the influence of fluctuating power.A battery-hydrogen storage system is used to realize the flexible grid [4].All the studies on capacity optimization allocation of multi-energy complementary microgrid systems in the above literature only consider single stakeholders of microgrids.However, with the investment and operation of a large number of "incremental distribution network" pilots, the interests of investors must be considered in capacity allocation and energy management of microgrids.As an effective way to solve such problems, game theory maximizes the interests of all game participants by finding Nash equilibrium points [5].
The traditional photovoltaic power generation system can only convert a small part of the irradiance into electricity, most of which is wasted in the form of heat, causing the temperature of the photovoltaic panel to rise, and the efficiency of the photovoltaic panel is further reduced.Integrating a photovoltaic (PV) system with a TEG system, a combined heat and power generation device, and a photovoltaic-TEG hybrid power generation system, is capable of using waste heat as a source of energy production, resulting in a higher output power density.Previous studies have considered multiagent and complex environments, and PSO is easy to fall into local optimal, so it is difficult to get the best power grid capacity allocation scheme in a short time.SAO algorithm, by referring to the snow ablation mechanism, can balance development and exploration well and can optimize the global optimal solution even in a more complex environment.

Model
The photovoltaic power probability model can be derived as: where: Γ (•) is gamma function; α and β are shape parameters of Beta distribution;  and  are PV power and maximum power respectively.
Considering the photovoltaic temperature difference in power generation, a thermal power generation module is added based on traditional photovoltaic power generation, and thermal power generation is related to the temperature of the hot end and the cold end.In the practical application of the new energy station, the power generation power of the photovoltaic thermal power generation system is increased according to many thermal power generation modules installed on photovoltaic panels.According to the investigation of the existing station, the output power can be increased by 8% by installing thermal power generation modules.
The model of the hydrogen production-storage-power generation system is constructed, and the system is used as an energy buffer unit to smooth the fluctuating power of the microgrid.The mathematical model of energy conversion and hydrogen storage is as follows: where: W EC,t is Hydrogen production capacity; P EC,t is EC power; η EC is efficiency.λ is the hydrogen capacity produced by electrolysis per degree.P FC,t is FC power; W FC,t is the hydrogen consumed by FC during the period of t; ηFC is the efficiency of FC; μ is hydrogen required to produce each kwh of electricity.E HST,t, and E HST,t+1 are the hydrogen storage capacity, respectively. and  dis are the efficiencies of hydrogen storage and discharge, respectively. , and − , dis are hydrogen storage and hydrogen release at time t, respectively.
The cost includes investment installation cost, operation and maintenance cost, power purchase cost, load interruption compensation cost, and light abandonment penalty cost.The income from photovoltaic thermal power generation is the income from electricity sales, the income from hydrogen energy storage in addition to the income of electricity sales, and the income from hydrogen and oxygen sales.In addition, both enjoy certain government subsidies according to the actual situation.
According to the above formula, combined with the actual market situation, the function of calculating the income of photovoltaic temperature difference and hydrogen energy storage in a certain scene is obtained.With particle swarm optimization, the smaller the population size, the faster the calculation speed, but the easier it is to fall into local optimization.A random initial solution space is set for the two populations of light and hydrogen, PV is used to identify photoelectric, and EC, HST, and FC are used to identify the process of hydrogen production and hydrogen storage.According to the function of the photovoltaic temperature difference and hydrogen energy storage, the income value of different solutions is calculated.When one party performs the calculation, the other party can arbitrarily take a reasonable value.
For each population, the three individuals with the highest fitness, the second highest fitness, and the third highest fitness were released into the elite pool, and the individuals with the top 50% fitness were regarded as a whole and their centroid was placed into the elite pool.The maximum and minimum values of PV, EC, HST, and FC are set based on site conditions.In SAO, a population is divided into two parts, exploration and development, and the two small populations use different location calculation formulas to achieve a balance between exploration and development.

Algorithm flow
The algorithm flow chart is as follows:

Calculation of snowmelt rate
SAO uses the degree-day method to reflect the snowmelt process.The general form of the method is as follows: ( ) where M stands for snowmelt rate and  stands for daily average temperature.1 refers to the base temperature, which is usually set to 0.  represents the degree day factor, ranging from 0.35 to 0.6.In each iteration, the mathematical expression for the value to be updated, , is shown as:  indicates the termination standard.In SAO, the snow melt rate is calculated using the following formula: ( ) ( ) Calculate the average of all the solutions in the population.

New solutions in the computational exploration part
The location during exploration is calculated as follows: ( ) where () represents the position of the i th iteration, () represents a random number vector based on Brownian motion with Gaussian distribution, ⊗ represents multiplication by entries, and 1 represents a number randomly selected from [0,1].In addition,  () refers to the current best solution, () from a group of randomly selected individuals in some of the elite, and  () is the location of the entire group of the center of mass.

New solutions for the calculation development part
Corresponding to the snowmelt process in the physical process, the snow melts rate and position update equation obtained are as follows: is the snowmelt rate, and 2 represents the random number selected from [−1,1].This parameter shows the differences between individuals.At this stage, the use of cross terms -2×(()−()) and (1− 2 ) × (()−()), based on the current optimal search agent and the barycentric position of the group of information, makes the individual is more likely to reach the area of hope.
As iterations increase, the possibility of individuals making irregular movements with highly dispersed characteristics increases.During the iteration, the development part gradually decreases and the exploration part correspondingly increases.
According to the set boundary value, the solution that crosses the boundary in the solution space is set to the boundary value.After calculating the new solution for each population, the optimal value of the elite pool and individual of the population is updated.First, the new fitness is calculated.When one population is calculated, the optimal value of other populations is taken into the calculation.Secondly, after the fitness calculation is completed, the elite pool is updated according to the method of the calculation development part.
After obtaining the optimal solution for each population, compare whether the solution in this iteration and the last iteration is the same.If the solution is the same, it indicates that the optimal has been reached and convergence has been achieved; if not, it indicates that the optimal has not been reached and the iteration continues.

Analysis of numerical examples
In the experiment, it is believed that the maximum hydrogen storage capacity of a hydrogen storage tank under standard atmospheric pressure is 1000 m 3 , the rated power of the fuel cell is 2 kW, the rated hydrogen output per KWh is 0.19 m 3 , the cost of fuel cell is 5000-yuan, hydrogen storage tank is 7000-yuan, electrolytic cell is 4000 yuan, and rated power of fuel cell is 2 kW.The rated power of the electrolytic cell is 13.62 kW.Light intensity and microgrid load used in the experiment are shown in Figure 2.  Other parameters such as photovoltaic subsidies, oxygen prices, electricity prices, and so on refer to the actual research.Finally, in the experimental environment, 5000 photovoltaic panels, 21 electrolytic cells, 1 hydrogen storage tank, and 4 fuel cells are planned.The planning conforms to the actual situation, in the case of stable profit of photovoltaic temperature difference system, the more the profit is higher, and hydrogen energy storage needs to be built with the power generation of photovoltaic temperature difference power generation.In this plan, the income from selling electricity is 39167 yuan, the income from selling oxygen is 4160 yuan, the carbon penalty is 10717 yuan, the load interruption penalty is 56,228 yuan, the government subsidy for photovoltaic thermal power generation is 25,070 yuan, the subsidy for hydrogen energy storage is 11678 yuan, and the monthly profit is 9723 yuan after removing construction costs and maintenance costs.Table 1 shows optimization results between the SAO algorithm and PSO.

Conclusion
The proposed method considers the photovoltaic thermal power generation mode with higher energy efficiency, and adopts the efficient and renewable clean energy hydrogen as the energy storage mode, both of which are the most cutting-edge new energy modes.Compared with the general particle swarm optimization, the SAO method has better performance, the simulation results are more accurate, and the convergence can be reached faster.Hydrogen energy, light abandonment, and load interruption penalty are evaluated comprehensively.Hydrogen energy is a new method of energy storage.Although the domestic technology is not mature at present, it has great development prospects.Moreover, the paper not only considers the benefits of photovoltaic thermal power generation and hydrogen energy storage but also takes into account the grid and environmental factors.

Figure 1
Figure 1 Algorithm flow chart.3.2Calculation of snowmelt rateSAO uses the degree-day method to reflect the snowmelt process.The general form of the method is as follows:

Figure 3 Figure 2 .
is the iterative diagram of the objective function.The light intensity and microgrid load used in the experiment.

Table 1
Optimization results between the SAO algorithm and PSO.