CN113346549A

CN113346549A - Multi-vector clean energy system under genetic algorithm and optimization method thereof

Info

Publication number: CN113346549A
Application number: CN202110732285.8A
Authority: CN
Inventors: 陈湘萍
Original assignee: Guizhou University
Current assignee: Guizhou University
Priority date: 2021-06-30
Filing date: 2021-06-30
Publication date: 2021-09-03

Abstract

The invention discloses a multi-vector clean energy system under a genetic algorithm, and correspondingly discloses an optimization method of the system, belonging to the technical field of renewable energy utilization and comprising a wind driven generator; the power output end of the wind driven generator is connected with the DC bus through the rectifier; the DC bus is connected into the electrolytic water tank and the power grid through the DC-DC converter and the first inverter respectively; the anode and cathode electrolytic reaction part of the electrolytic water tank is respectively communicated with the oxygen tank and the hydrogen tank through connecting pipes; the oxygen storage tank and the hydrogen storage tank are both connected with the fuel cell; the fuel cell is connected with the power grid through a second inverter; the power grid is connected with a household power utilization system; the rectifier, the DC-DC converter, the first inverter and the second inverter are all electrically connected with the central optimization controller, and the wind-hydrogen-fuel cell power system effectively solves the problems of low efficiency and high operation cost in the current wind-hydrogen-fuel cell power system.

Description

Multi-vector clean energy system under genetic algorithm and optimization method thereof

Technical Field

The invention relates to the technical field of renewable energy utilization, in particular to a multi-vector clean energy system under a genetic algorithm and an optimization method thereof.

Background

The technology of incorporating renewable energy into smart grid is being developed vigorously, renewable energy including wind energy, solar energy and biological energy has become the most promising resource in the world, and it is noted that wind energy accounts for 45% of the total amount of renewable energy as a main contributor to power generation by renewable energy. For rural areas with grid connections, power generated from renewable energy sources may be input to the main grid. For rural areas without grid connection, the generated electricity needs to be stored (e.g., batteries) to balance load and demand. Currently, the research on wind-hydrogen-fuel cell power systems is only in the laboratory operating phase, and the industrial application thereof is limited by the low efficiency and high cost of the system; to address these problems, an efficient and optimized operating scheme is needed to handle the operating transactions while managing the flow of electrical power between the various components. Therefore, in order to store a certain amount of electric energy during the off-peak period and output it during the peak period, energy storage is necessary.

Disclosure of Invention

The technical problem to be solved by the invention is as follows: a multi-vector clean energy system under a genetic algorithm is provided to solve the problems of low efficiency and high operation cost in the current wind-hydrogen-fuel cell power system.

In order to solve the problems, the invention provides the following technical scheme:

a multi-vector clean energy system under a genetic algorithm comprises a wind driven generator; the power output end of the wind driven generator is connected with the DC bus through the rectifier; the DC bus is connected into the electrolytic water tank and the power grid through the DC-DC converter and the first inverter respectively; the anode and cathode electrolytic reaction part of the electrolytic water tank is respectively communicated with the oxygen tank and the hydrogen tank through connecting pipes; the oxygen storage tank and the hydrogen storage tank are both connected with the fuel cell; the fuel cell is connected with the power grid through a second inverter; the power grid is connected with a household power utilization system; the rectifier, the DC-DC converter, the first inverter and the second inverter are all electrically connected with the central optimization controller.

The invention also provides an optimization method of the multi-vector clean energy system under the genetic algorithm, which comprises the following steps:

s1, calculating the output power of the wind driven generator according to the measured data and combining the formula 1 or the formula 2 and recording the output power as p_wind；

Wherein, P_r，v_{cut_in}，v_{cut_out}，v_rAnd v represents the rated power, input wind speed, output wind speed, rated wind speed and turbine blade wind speed of the wind generator (1), respectively; ρ, A and v represent local air density, equivalent area of the blade acting on the wind turbine and average wind speed, respectively.

S2, calculating the electric energy consumption power of the electrolytic water tank according to the formula 3, and recording the electric energy consumption power as p_e；

Wherein;

η_we，U_nthe hydrogen flow rate (liter/hour), the oxygen flow rate (liter/hour), the conversion efficiency of the electrolytic water tank and the voltage are respectively expressed.

S3, calculating the actual output power of the fuel cell and recording as p_fc，

S4 introduction of ESOC_STo evaluate the hydrogen and oxygen remaining in the oxygen tank and the hydrogen tankBy a level of

Equation 4 calculates the ESOC_S；

Wherein, V_HAnd V_oEquivalent volume states, ESOC, of the hydrogen and oxygen tanks, respectively_HAnd ESOC_OThe levels of the hydrogen tank and the oxygen tank, respectively, can be determined by equations 5 and 6:

P_h，

P_oand

the current pressure of the hydrogen tank, the full load pressure of the hydrogen tank, the current pressure of the oxygen tank and the full load pressure of the oxygen tank are respectively set;

s5, establishing a GA objective function by equation 7, and defining constraint conditions of the objective function, equation 7a, equation 7b, equation 7c, and equation 7 d:

p_dir(k)+p_fc(k)＝p_load(k) (7b)

0.2≤ESOC_S(k)≤0.9 (7c)

0≤p_grid(k)≤p_wind(k) (7d)

wherein p is_grid(k) Is electric power output to the power grid; p is a radical of_load(k) For the load power of the domestic electric system, p_dir(k) The electric quantity generated by wind power generation is used for meeting the power demand of a household power system;

s6 for finding p_grid(k) Further establishing an SSM fitness function through an equation 8, and defining constraint conditions of the fitness function, namely an equation 9, an equation 10 and an equation 11;

x₁(k)＝x₁(k-1)+ξ₁·x₂(k)+φ₁ (8)

x₂(k)＝y(k)＝p_dir(k) (9)

p_dir(k)+p_fc(k)＝p_loc (10)

wherein x is₁(k) And x₂(k) For a defined state variable, x₁(k)＝ESOC_SAnd y (k) is a defined output variable; xi₁Is an empirical coefficient, phi₁A relaxation variable when establishing a state space;

defining a state variable x₁(k) And x₂(k) (ii) a And defining an output variable y (k); k is 1-24; let k equal to 1, and pair x₁(0) Assign value to x₁(0) Is equal to any number in the interval of 0.2 to 0.9; and for the time-dependent parameters obtained in steps S1 through S4: p is a radical of_wind、p_eAnd ESOC_SPerforming population initialization processing to obtain a candidate solution;

s7, performing adaptability evaluation on the obtained candidate solution through the SSM fitness function, judging whether the candidate solution meets the stopping standard of the SSM function, and if the candidate solution meets the stopping standard, entering the step S8; if not, processing the candidate solution by using a cross and mutation method, performing iterative computation by using an SSM fitness function to generate a new generation of candidate solution, and repeating the steps;

s8, if the candidate solution meets the stopping standard, the candidate solution is taken as a final solution; obtaining p in GA target function according to optimal solution_grid(k) And obtaining the corresponding p_wind、p_eAnd ESOC_SIndex parameters are then optimized by the central optimization controller according to p_wind、p_eAnd ESOC_SSending control instructions to the wind driven generator, the electrolytic water tank, the oxygen storage tank and the hydrogen storage tank to adjust the working states of all the devices;

preferably, the fuel cell is a PEM fuel cell; its actual output power p is calculated in step S3_fcThe calculation can be carried out according to formula 13;

p_fc＝i·E^cell·m (13)

where i is the current density, E^cellFor the actual output voltage, m is the effective mass of the PEM fuel cell;

E^cellcan be obtained by calculation of equation 14:

E^cell＝E⁰-η^act-η^ohmic-η^con (14)

wherein E is⁰Representing the open circuit potential or thermodynamic equilibrium potential, η^act，η^ohmicAnd η^conRespectively representing activation loss, resistivity loss and concentration loss;

E⁰、η^act、η^ohmic、η^conthe calculation can be obtained by

equations

15, 16, 17, and 18, respectively:

η^ohm＝i·(R_els+R_ion) (17)

wherein R represents a gas universal constant, T represents a temperature, alpha represents a charge transfer coefficient, n represents the number of electrons entering the reaction, represents a Faraday constant, i represents a current density, i represents⁰Represents the equivalent current density; r_eleAnd R_ionElectronic resistance and ionic resistance respectively; i.e. i^LA limit value representing the current density; Δ H is the reaction enthalpy; Δ S is the entropy of reaction. .

Preferably, ξ in step S2₁The calculation can be made by:

ESOC_S(k)＝ESOC_S(k-1)+ξ₁ (18)

wherein k is 1-24 and represents different time periods in a day.

Preferably, the criterion for determining whether the candidate solution meets the stopping criterion of the SSM fitness function in step S7 may be performed in two ways, the first way is by determining whether the fitness function converges under the candidate solution, and if so, determining that the candidate solution meets the stopping criterion; the second is to judge whether the actual iteration number reaches the iteration number standard set by the system, and if so, the actual iteration number can be judged to meet the stop standard.

The invention has the beneficial effects that:

the invention mainly takes wind energy as the only energy source of the whole system, and combines the GA genetic algorithm to control the wind power generation to mainly meet the power demand of a target family; the residual wind energy is stored by electrolyzing water to generate hydrogen and oxygen; the redundant energy generated by the wind driven generator is transmitted to the smart grid; if the demand exceeds the power provided by wind energy, the fuel cell will start to generate electricity and act as a backup power source, and the beneficial effects are shown in the following points:

(1) in the present invention, energy requirements are balanced by using an energy storage system for the hydrogen fuel cell.

(2) During system operation, the load is balanced by the energy storage system, consuming wind energy locally.

(3) After the optimization processing is carried out through the GA optimization algorithm, the conversion efficiency of converting wind energy into a chemical energy (hydrogen) or electric energy form is high.

(4) The advanced technology provides a new method for grid-connected operation of a power grid, and intermittent renewable energy can be effectively utilized.

Drawings

FIG. 1 is a schematic diagram of the system configuration in the present embodiment;

FIG. 2 is a system block diagram between step S5 and step S9 in the present embodiment;

3(a) - (d) are schematic diagrams respectively illustrating energy consumption of a survey family in four seasons of spring, summer, autumn and winter in the embodiment;

fig. 4(a) - (d) are graphs of the average single-day time-load demand variation of the survey family in four seasons of spring, summer, fall and winter in the present embodiment, respectively;

FIG. 5 is a graph of wind speed for the present example surveying home local yearly;

FIGS. 6(a) - (d) are schematic diagrams respectively illustrating the average single-day time-wind speed variation of the survey family in four seasons of spring, summer, autumn and winter in the present embodiment;

FIGS. 7(a) - (d) are respectively the average single-day time-ESOC of the survey family in four seasons of spring, summer, autumn and winter in this example_SChange four intentions of (1);

FIGS. 8(a) - (d) are schematic diagrams of the power-time variation of the average single-day system input grid of the survey family in four seasons of spring, summer, autumn and winter in the present embodiment, respectively;

description of reference numerals: 1. the system comprises a wind driven generator, 2, a rectifier, 3, a DC bus, 4, a central optimization controller, 5, a DC-DC converter, 6, a first inverter, 7, a power grid, 8, an electrolytic water tank, 9, an oxygen storage tank, 10, a hydrogen storage tank, 11, a household power system, 12 and a fuel cell.

Detailed Description

The invention will be further described with reference to the following drawings and specific embodiments:

example (b):

referring to fig. 1, the present embodiment provides a multi-vector clean energy system under genetic algorithm, which includes a wind power generator 1; the power output end of the wind driven generator 1 is connected with a DC bus 3 through a rectifier 2; the DC bus 3 is respectively connected into an electrolytic water tank 8 and a power grid 7 through a DC-DC converter 5 and a first inverter 6; the anode and cathode electrolytic reaction part of the electrolytic water tank 8 is respectively communicated with an oxygen tank 9 and a hydrogen tank 10 through connecting pipes; the oxygen tank 9 and the hydrogen tank 10 are both connected with a fuel cell 12; the fuel cell 12 is connected to the grid 7 via a second inverter 7; the power grid 7 is connected with a household power utilization system 11; the rectifier 2, the DC-DC converter 5, the first inverter 6 and the second inverter 12 are all electrically connected to a central optimization controller 4.

s1, calculating the output power of the wind driven generator 1 according to the measured data and combining the formula 1 or the formula 2 and recording the output power as p_wind；

Wherein, P_r，v_{cut_in}，v_{cut_out}，v_rAnd v represents the rated power, input wind speed, output wind speed, rated wind speed and turbine blade wind speed of the wind turbine 1, respectively; ρ, A and v represent local air density, equivalent area of the blade acting on the wind turbine 1 and average wind speed, respectively.

S2, calculating the electric power consumption of the electrolytic water tank 8 according to the formula 3, and recording the electric power consumption as p_e；

Wherein;

η_we，U_nthe hydrogen flow rate (liter/hour), the oxygen flow rate (liter/hour), the conversion efficiency of the electrolytic water tank 8 and the voltage are respectively expressed.

S3, calculating the actual output power of the fuel cell 12 and recording as p_fc，

S4 introduction of ESOC_STo estimate the levels of hydrogen and oxygen remaining in the oxygen tank 9 and the hydrogen tank 10, the ESOC is calculated by equation 4_S；

P_h，

P_oand

p_dir(k)+p_fc(k)＝p_load(k) (7b)

0.2≤ESOC_S(k)≤0.9 (7c)

0≤p_grid(k)≤p_wind(k) (7d)

wherein, P_grid(k) Is electric power output to the power grid; p is a radical of_load(k) For the load power, p, of the domestic electric system 11_dir(k) The electric quantity generated by the wind power generation is used for meeting the required power of the household power system 11;

x₁(k)＝x₁(k-1)+ξ₁·x₂(k)+φ₁ (8)

x₂(k)＝y(k)＝p_dir(k) (9)

p_dir(k)+p_fc(k)＝p_loc (10)

defining a state variable x₁(k) And x₂(k) (ii) a And defining an output variable y (k); k is 1-24; let k equal to 1, and pair x₁(0) Assign value to x₁(0) Equal to over the interval 0.2 to 0.9Any one number; and for the time-dependent parameters obtained in steps S1 through S4: p is a radical of_wind、p_eAnd ESOC_SPerforming population initialization processing to obtain a candidate solution;

s8, if the candidate solution meets the stopping standard, the candidate solution is taken as a final solution; obtaining P in GA target function according to optimal solution_grid(k) And obtaining the corresponding p_wind、p_eAnd ESOC_SIndex parameters are then optimized by the central optimization controller 4 according to p_wind、p_eAnd ESOC_SThe index parameters of the wind power generator send control instructions to the wind power generator 1, the electrolytic water tank 8, the oxygen storage tank 9 and the hydrogen storage tank 10 so as to adjust the working state of each device;

the fuel cell 12 is a PEM fuel cell; the actual output power pfc calculated in step S3 can be obtained by calculation using equation 13;

p_fc＝i·E^cell·m (13)

E^cellcan be obtained by calculation of equation 14:

E^cell＝E⁰-η^act-η^ohmic-η^con (14)

E⁰、η^act、η^ohmic、η^conthe calculation can be obtained by

equations

15, 16, 17, and 18, respectively:

η^ohm＝i·(R_ele+R_ion) (17)

wherein R represents a gas universal constant, T represents a temperature, alpha represents a charge transfer coefficient, n represents the number of electrons entering the reaction, represents a Faraday constant, i represents a current density, i represents⁰Represents the equivalent current density; r_eleAnd R_ionElectronic resistance and ionic resistance respectively; i.e. i^LA limit value representing the current density; Δ H is the reaction enthalpy; Δ S is the entropy of reaction.

ξ in step S2₁The calculation can be made by:

ESOC_S(k)＝ESOC_S(k-1)+ξ₁ (18)

wherein k is 1-24 and represents different time periods in a day.

The basis for determining whether the candidate solution meets the stopping criterion of the SSM fitness function in step S7 may be performed in two ways, the first way is by determining whether the fitness function converges under the candidate solution, and if so, determining that it meets the stopping criterion; the second is to judge whether the actual iteration number reaches the iteration number standard set by the system, and if so, the actual iteration number can be judged to meet the stop standard.

In particular, this embodiment also performs tracking adjustment on a family to which the system is applied, and the recording time of the survey is one year, as shown in fig. 3(a) - (d), each chart shows the scenes of working days and weekends in each season.

Similar activity patterns are found from the four charts. These weekday activities can generally be divided into three modes, including morning, noon, and evening. For example, on a weekday in spring, they start at 7:00, 13:00, and 16:30, respectively, as shown in fig. 3 (a). From point 7 onwards, the household performs different activities. The summer (fig. 3(b)) working day starts from 5:30, which is earlier than the other seasons. From the irregular polygon image in the morning, the tenants are clearly distinguished because the family members have different calendars on weekdays. The pattern in autumn (fig. 3(c)) is very similar to spring, except that every part of autumn begins earlier than spring. For example, the morning in autumn starts at 6.30 and is half an hour earlier than the spring. The midday in winter (fig. 3(d)) starts slightly earlier than the other seasons. Because of the cold weather in winter, a tenant may return to home earlier in winter. The working days in winter occupy the most (11:00 to 15: 00). The weekend can be divided into day and night. The weekend activities of the four seasons were almost the same except that the night of the spring ended earlier than the other seasons. These differences result in variations in energy consumption.

In the present invention, the proposed method is verified by an investigation of four different daily electricity usage patterns for four seasons. Key features of power generation and demand for a rural user can be summarized in table 1, and the waveform in fig. 4 illustrates the demand curve for these days.

As shown in Table 1, the load varied from 0 to 5.8kW in the spring. The demand for electricity rises to 9.7kW in autumn and the power consumed in autumn exceeds the next three days to reach maximum (16.1 kWh). A large amount of electric power is also consumed in winter because of poor lighting and cold weather, which increases the need for lighting and heating. From the view of seasons and low demand, the four seasons show similar trends, and the power consumption in winter and autumn is generally larger than that in spring and summer. In fig. 8, the same trend shows that the peak time of four typical days is less than 1 hour even in autumn and winter. These three areas, named early, noon and late, are shown in fig. 4. In spring for example, the power consumption increases from 6:30 in the morning to a peak at 8: 15. Subsequently, the power consumption fluctuates before dropping below 1 kW. Starting at 13:00, power consumption again increased until 14:00 pm. In the noon region, the energy consumption is also fluctuating, because different appliances are used during this time. The power consumption peak occurs at 17:00 and the next peak occurs at around 19: 00. These three areas become less distinctive on selected winter days, as compared to weekdays, with a different activity schedule for the residents on weekends.

TABLE 1. Power demand of a common household within four days

The average wind speed measured and recorded at the test site ranged between 0 and 21m/s and the power produced ranged between 0 and 4kW, as shown in figure 5. There is a fast wind speed for each season, but there is no fixed wind speed pattern for any one season. The four graphs in fig. 6 illustrate the variation of the wind power per day. Spring wind power as shown in fig (6a), gradually drops from 1.4kW peak at 1:00 to 11:00 lowest point and maintains low state to 16: 00. Subsequently, the power was slowly increased to around 0.5kW at around 21 points. As shown in fig (6b), the output power during most of the summer is less than 1kW, except for a peak power of 20: 00. As shown in fig (6c), the wind power variation in fall was relatively small, reaching a maximum of 1.2kW at 12: 00. As shown in fig (6d), the wind power in winter ranged from 0.2 to 0.9kW with some variation. The waveform may show the wind power. By comparison, autumn and winter produce more energy than spring and summer. The energy provided in the summer is minimal during the selected time.

In this study, demand and wind energy within the home were intermittent, but their changes exhibited completely different characteristics. Therefore, it is desirable to employ hydrogen-based energy storage systems to make up for the gap, and GA methods to maintain optimal operation of the energy storage system.

These features can be summarized as follows:

1. the peak demand exceeds 4kW, but only lasts for a few minutes. The summer demand is 4kW, and the winter demand is less than 1 kW;

2. there are 3 peaks of energy demand on selected days, and wind energy is random on these days.

3. Peak demand and peak supply do not occur simultaneously on any selected date.

4. Typically, during any selected day, the wind energy produced during the night is higher than the energy demand.

ESOC_SThe results of the test over the selected four days are shown in figure 7. If it is ESOC in spring_SIncrease from 0:00 to 7: 30; during this time, energy demand is low and wind energy maintains a high output. Most of the wind energy is used in the energy storage system (hydrogen is generated by electrolysis). Then, wind energy and ESOC_SThe energy demand is reduced to meet both wind power and fuel cell as shown in figure (7 a).

As shown in FIG. 7b, the ESOC_SIncreasing continuously from 0:00 to 5:00 until reaching its upper limit (0.9), starting at 6:30, and dropping sharply until 14: 00. After some fluctuation, the ESOC_SThe upper limit is reached at 20: 00. In summer, the wind power becomes stronger, the wind power demand is reduced, and the wind power is continuously and rapidly increased.

Some fluctuation also occurred in autumn as shown in fig. 7 c. ESOC_SThe increase was maintained during the night until 6: 30. The energy demand at that time was higher than wind energy. Starting from 9:30, ESOC_SGrowth is restored within 12:00 to 17:00 before peak demand is met in the afternoon. Subsequently, the ESOC_SAnd decreased until the lower limit (0.2) was reached.

As shown in FIG. 7d, ESOC in winter_SMaintained at 0.5 or above. The main descent processes occur at 9:00 and 17:00 at the start of peak demand in the morning and evening. After the peak period, the ESOC_SBegins to rise as more power is input into the storage system from the wind turbine.

The four graphs in fig. 8 represent the power input to the grid. As shown in fig. 8(a), the grid is hardly supplied with any wind energy in spring, and the wind energy is fully used to supply the demand of the household or to charge the energy storage system. As shown in FIG. 8d, the wind energy input to the grid is the most in winter because the generated wind energy is far from the gridFar exceeding local requirements. As shown in fig (8b), there are two periods of wind energy input into the grid in summer, one period being 4:00 to 6:00 before peak demand in the morning. The other one starts with a peak value of 20: 00. As shown in FIG. 8c, the majority of the time of wind energy input in autumn occurs between 12:30 and 16:30, at which time the ESOC occurs_SIs sufficiently charged. The additional power of the wind generator is between 12:00 and 17:00 input to the grid, and the ESOC_SThe upper limit is maintained and no charging or discharging of the storage system is required.

The important indices of the system architecture are summarized in table 2.

Energy index:

the demand of the households was 9.86kWh, 10.40kWh, 14.65kWh and 11.75kWh, respectively, in the selected four days, with the power demand reaching a maximum in the late autumn. In this study, wind power and fuel cells meet the demand. In spring for example, the wind energy is 9.99 kWh. The energy of the fuel cell was 6.25kWh, and the energy of electrolysis was 6.36 kWh. The total energy required by the load is 9.99+ 6.25-6.36-9.88 kWh. However, the load actually consumed 9.86kWh, the remainder being the power loss during operation (0.02 kWh). The losses for the remaining time were 0.99kWh, 0.94kWh and 0.04kWh, respectively.

It can be seen that autumn and winter consume more energy than spring and summer. The power of the fuel cell is used to compensate for these selected date differences. In addition to powering the load, a portion of the wind energy is used to generate hydrogen in order to store the energy for later use. When wind power in winter exceeds local demand, additional wind power is input into the grid.

The power index is as follows:

the daily peak demand power in the four seasons was 5.81kW, 6.10kW, 9.63kW, and 8.91kW, respectively. The fuel cell modules provide an energy gain, providing maximum power of 5.71kW, 5.82kW, 8.68kW and 8.29kW for the corresponding time. However, the peak power of wind energy is only 4 kW. The power fed back into the grid reaches a maximum value (3.75kW) in summer. In this study, the peak demand and peak power of wind energy are characterized differently, and the fuel cell can balance this difference. Fuel cells reduce the need for any frequent and significant power exchange between the load and the grid. Thus, the system is designed to improve the recovery of the grid while maximizing the use of local renewable energy.

The time index is as follows:

only for a limited time can the load power reach 4kW in a selected four days. The peak required power lasts for 0.09 hour, 0.03 hour, 0.05 hour and 0.25 hour in spring, summer, autumn and winter, respectively. The demand is low for the remaining time. The fuel cell modules provide electrical power to the load for 14.6 hours, 10.3 hours, 9.2 hours, and 7.3 hours, respectively. The wind power generator continuously transmits power to the system, and the power transmission time is respectively 9.4 hours, 12.5 hours, 13.6 hours and 16.7 hours. The fuel cell and hydrogen plant may run for up to 24 hours per day. The energy storage system makes up the difference between power generation and power supply, thereby improving the dynamic response of the whole system

Other indexes are as follows:

ESOC_Smost of the time ranges between 0.2 and 0.9. However when the ESOC is_SBelow 20%, a potential risk may occur. This is the case because wind power is low and in a high demand state for a long time. By using the proposed GA optimization method, wind energy is mainly consumed locally except for the wind energy power supply network in winter. The best and worst case are 100% and 51.7% wind energy consumption in spring and winter, respectively. The storage system is kept in a safe range by adopting the GA optimized operation scheme. Therefore, the safety of the storage system is ensured while the renewable energy is maximally utilized.

Claims

1. A multi-vector clean energy system under genetic algorithm comprises a wind driven generator (1); the method is characterized in that: the power output end of the wind driven generator (1) is connected with the DC bus (3) through the rectifier (2); the DC bus (3) is respectively connected into an electrolytic water tank (8) and a power grid (7) through a DC-DC converter (5) and a first inverter (6); the anode and cathode electrolytic reaction part of the electrolytic water tank (8) is respectively communicated with the oxygen storage tank (9) and the hydrogen storage tank (10) through connecting pipes; the oxygen tank (9) and the hydrogen tank (10) are both connected with the fuel cell (12); the fuel cell (12) is connected with the power grid (7) through a second inverter (7); the power grid (7) is connected with a household power utilization system (11); the rectifier (2), the DC-DC converter (5), the first inverter (6) and the second inverter (12) are all electrically connected with the central optimization controller (4).

2. The method for optimizing multi-vector clean energy system under genetic algorithm according to claim 1, wherein: the method comprises the following steps:

s1, calculating the output power of the wind driven generator (1) according to the measured data and the combination of the formula 1 or the formula 2 and recording the output power as p_wind；

Wherein, P_r，v_{cut_in}，v_{cut_out}，v_rAnd v represents the rated power, input wind speed, output wind speed, rated wind speed and turbine blade wind speed of the wind generator (1), respectively; ρ, A and v represent local air density, equivalent area of the blade acting on the wind turbine (1) and average wind speed, respectively.

S2, calculating the electric energy consumption power of the electrolytic water tank (8) according to the formula 3, and recording the electric energy consumption power as p_e；

Wherein;

η_we，U_nthe hydrogen flow rate (liter/hour), the oxygen flow rate (liter/hour), the conversion efficiency of the electrolytic water tank (8) and the voltage are respectively expressed.

S3, calculating the actual output power of the fuel cell (12) and recording the actual output power as p_fc，

S4 introduction of ESOC_STo evaluate the levels of hydrogen and oxygen remaining in the oxygen tank (9) and the hydrogen tank (10), the ESOC is calculated by equation 4_S；

P_h，

P_oand

p_dir(k)+p_fc(k)＝p_load(k) (7b)

0.2≤ESOC_S(k)≤0.9 (7c)

0≤p_grid(k)≤p_wind(k) (7d)

wherein p is_grid(k) Is electric power output to the power grid; p is a radical of_load(k) For the load power, p, of a domestic electric system (11)_dir(k) The electric quantity generated by the wind power generation is used for meeting the required power of a household power utilization system (11);

x₁(k)＝x₁(k-1)+ξ₁·x₂(k)+φ₁ (8)

x₂(k)＝y(k)＝p_dir(k) (9)

p_dir(k)+p_fc(k)＝p_loc (10)

wherein x is₁(k) And x₂(k) For a defined state variable, x₁(k)＝ESOC_SAnd y (k) is a defined output variable;

ξ₁is an empirical coefficient, phi₁A relaxation variable when establishing a state space;

defining a state variable x₁(k) And x₂(k) (ii) a And defining an output variable y (k); k is 1-24; let k equal to 1, and pair x₁(0) Assign value to x₁(0) Is equal to any number in the interval of 0.2 to 0.9; and for the time-dependent parameters obtained in steps S1 through S4: p is a radical of_wind、p_eAnd ESOC_SPerforming population initialization processing toObtaining a candidate solution;

s8, if the candidate solution meets the stopping standard, the candidate solution is taken as a final solution; obtaining p in GA target function according to optimal solution_grid(k) And obtaining the corresponding p_wind、p_eAnd ESOC_SIndex parameters are then optimized by the central optimization controller (4) according to p_wind、p_eAnd ESOC_SThe index parameters of the device send control commands to the wind driven generator (1), the electrolytic water tank (8), the oxygen storage tank (9) and the hydrogen storage tank (10) to adjust the working state of each device.

3. The method for optimizing multi-vector clean energy system under genetic algorithm according to claim 2, wherein: the fuel cell (12) is a PEM fuel cell; its actual output power p is calculated in step S3_fcThe calculation can be carried out according to formula 13;

p_fc＝i·E^cell·m (12)

E^cellcan be obtained by calculation of equation 14:

E^cell＝E⁰-η^act-η^ohmic-η^con (13)

E⁰、η^act、η^ohmic、η^conthe calculation can be obtained by equations 15, 16, 17, and 18, respectively:

η^okm＝i·(R_ele+R_ion) (16)

4. The method for optimizing multi-vector clean energy system under genetic algorithm according to claim 2, wherein: ξ in step S2₁The calculation can be made by:

ESOC_S(k)＝ESOC_S(k-1)+ξ₁ (18)

wherein k is 1-24 and represents different time periods in a day.

5. The method for optimizing multi-vector clean energy system under genetic algorithm according to claim 2, wherein: the basis for determining whether the candidate solution meets the stopping criterion of the SSM fitness function in step S7 may be performed in two ways, the first way is by determining whether the fitness function converges under the candidate solution, and if so, determining that it meets the stopping criterion; the second is to judge whether the actual iteration number reaches the iteration number standard set by the system, and if so, the actual iteration number can be judged to meet the stop standard.