An Extended Timeline Particle Swarm Optimization Algorithm Applied to Energy Routing Management

Abstract

The present invention discloses an extended timeline particle swarm optimization algorithm applied to energy routing management, which is applied to the field of energy management of an energy routing system. It fully considers the state of charge constraints of the battery at future time points and finds an optimal scheduling curve based on time. Even when the distributed power capacity is less than the maximum power demand of the system at a certain future moment, the algorithm can still plan ahead to make the battery store sufficient electrical energy so that the scheduling power and the state of charge at any time point are within the constraints. The extended timeline particle swarm optimization algorithm proposed by the present invention overcomes the limitation that the standard particle swarm timeline can only flow unidirectionally. Compared with the dynamic programming algorithm, it does not require a specific algorithm structure and there is no curse of dimensionality problem, enabling it to search for the optimal curve in multiple space-time ranges and effectively solving the problem of mutual coupling of the optimization results at the front and back time points.

Description

Technical Field

[0001] The present invention relates to the field of smart grid dispatching, and in particular to an extended time-axis particle swarm algorithm applied to energy routing management. Background Art

[0002] Today, with the increasingly warming global climate, reducing carbon emissions has become an urgent pursuit in many industries. How to replace fossil energy with renewable energy and green hydrogen energy has become a research hotspot. However, renewable energy power generation has strong randomness, and its wide access increases the power fluctuation of the power grid and affects the stability of the power grid. The energy routing system can suppress the power fluctuation of new energy, participate in power grid dispatching, assist in peak shaving and valley filling, and earn certain economic benefits. Therefore, taking it as the entry point for new energy grid connection and conducting economic and efficient energy management is of great significance for improving the friendliness of new energy grid connection and the penetration rate of new energy.

[0003] However, with the gradual increase in the requirements for the energy routing system, as well as the access of devices such as photovoltaic, wind power, energy storage batteries, hydrogen electrolyzers, and hydrogen fuel cells, the energy fluctuation of the energy routing system becomes more complex, significantly increasing the difficulty of energy management. It is necessary to select appropriate control strategies for each device according to different energy management objectives and device characteristics, design an energy management system, conduct automatic control and intelligent dispatching decisions, and under the conditions of multiple constraints, with the highest new energy utilization rate, the lowest power generation cost, and the maximum economic benefit as the goals, adjust the output of controllable distributed power sources and loads to achieve the functions of peak shaving and valley filling and economic dispatching, providing guarantee for the safe, stable, economic and efficient operation of the system.

[0004] Since energy management is a multi-objective, multi-constraint, multi-stage, multi-variable and complex non-linear optimal dispatching problem, it is necessary to introduce an optimization algorithm to calculate the optimal dispatching arrangement, accelerate the dispatching decision speed, and improve the economic level of dispatching. However, the currently common algorithms can only perform single-point optimization for multiple time points unidirectionally within continuous time periods. Especially when the maximum power demand of the system is greater than the capacity of the distributed power source, they cannot take into account the impact of the current moment's dispatching on the SOC of the battery at future moments. Therefore, it is of great significance to design a new algorithm to solve the above problems. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide an extended time-axis particle swarm algorithm applied to energy routing management, which fully considers the SOC of the battery at future time points and other constraint conditions, and finds the optimal curve based on time. Even when the capacity of the distributed power source is less than the maximum power demand of the system, it can plan ahead to ensure that the dispatching power and SOC at any time point are within the constraint range.

[0006] To solve the above technical problems, the technical solution adopted by the present invention is as follows: An extended time-axis particle swarm algorithm applied to energy routing management, comprising the following steps:

[0007] Step S1: Use the income from selling hydrogen and oxygen produced by the hydrogen electrolyzer throughout the day, the income from assisting the power grid in peak shaving, and the cost of purchasing electricity from the power grid to form the system income / cost objective function;

[0008] Step S2: Based on the parameters of the hydrogen electrolyzer, hydrogen fuel cell, battery capacity, and state of charge, set the constraint conditions at each moment;

[0009] Step S3: Initialize the parameters including the position X of all particles, the moving speed V, the number of iterations n, the learning factors c 1 and c 2 , and the inertia coefficient ω, and generate particles with the time t dimension;

[0010] Step S4: Calculate the fitness of each particle according to the system income / cost objective function in Step S1 to obtain the individual optimal particle and the group optimal particle;

[0011] Step S5: Update the velocity V i,d,t and position X i,d,t of all particles at each moment according to the positions of the individual optimal particle and the group optimal particle, and constrain the range of the updated particle velocity and position;

[0012] Step S6: Determine whether the current number of iterations has reached the maximum number of iterations. If so, stop the iteration and output the optimal solution; otherwise, return to Step S4 to continue the iterative calculation.

[0013] A further improvement of the technical solution of the present invention is that: the system income / cost objective function in Step S1 is:

[0014]

[0015] In the formula, is the total amount of electricity taken from the power grid by the energy routing system throughout the day; is the total amount of electricity fed into the power grid by the energy routing system throughout the day; is the total amount of hydrogen produced by the hydrogen electrolyzer throughout the day; is the total amount of hydrogen used by the hydrogen fuel cell throughout the day; C b is the price of purchasing electricity from the power grid; C s is the price of feeding electricity into the power grid; C H2 is the price of buying and selling hydrogen; C O2 is the price of selling oxygen;

[0016] The expression for the total amount of electricity taken from the power grid by the energy routing system throughout the day is:

[0017]

[0018] Wherein, is the electric power purchased by the energy routing system from the power grid at time t;

[0019] The expression for the total amount of electricity fed by the energy routing system to the power grid throughout the day is:

[0020]

[0021] Wherein, is the electric power sold by the energy routing system to the power grid at time t;

[0022] The expression for the total amount of hydrogen produced by the hydrogen electrolyzer throughout the day is:

[0023]

[0024] Wherein, is the hydrogen production rate of the hydrogen electrolyzer in the energy routing system at time t.

[0025] The expression for the total amount of hydrogen consumed by the hydrogen fuel cell throughout the day is:

[0026]

[0027] Wherein, is the hydrogen consumption rate of the hydrogen fuel cell in the energy routing system at time t.

[0028] A further improvement of the technical solution of the present invention lies in that: the power P of the hydrogen electrolyzer at time t in step S2 el (t) has the following constraint conditions:

[0029]

[0030] Wherein, P el,max and P el,min are respectively the maximum boundary and the minimum boundary of the operating power of the hydrogen electrolyzer; is a switching quantity, when the hydrogen electrolyzer operates, otherwise it stops;

[0031] The power P of the hydrogen fuel cell at time t fc (t) has the following constraint conditions:

[0032]

[0033] Wherein, P fc,max and P fc,min are respectively the maximum boundary and the minimum boundary of the operating power of the hydrogen fuel cell; is a switching quantity, when the hydrogen fuel cell operates, otherwise it stops;

[0034] Charging power of the battery at time t and the discharging power have the following constraint conditions:

[0035]

[0036]

[0037] In the formula, and are the maximum charging power boundary and the maximum discharging power boundary of the battery respectively; σ bat is a switching quantity. When σ bat = 1, the battery is discharging. When σ bat = 0, the battery is charging;

[0038] The constraint conditions of the state of charge (SOC) of the battery are as follows:

[0039] SOC min ≤ SOC(t) ≤ SOC max ,

[0040] In the formula, SOC max and SOC min are the maximum boundary and the minimum boundary of the state of charge of the battery respectively;

[0041] The equality constraint conditions of the battery are as follows:

[0042]

[0043] In the formula, P pv (t) is the output power of the photovoltaic at time t, P load (t) is the demand power of the load at time t, η bat is the charge-discharge efficiency of the battery, η pv is the photovoltaic power generation efficiency, η fc is the power generation efficiency of the hydrogen fuel cell, η el is the hydrogen production efficiency of the hydrogen electrolyzer, η grid is the power interaction efficiency between the energy routing system and the power grid;

[0044] The interaction constraint conditions of the active power between the energy routing system and the power grid at time t are as follows:

[0045]

[0046]

[0047] In the formula, and are the maximum power purchase boundary from the power grid and the maximum power selling boundary to the power grid of the energy routing system respectively; σ gridis a switching quantity, σ grid When σ grid = 1, only buy electricity from the power grid, and when σ

[0048] A further improvement of the technical solution of the present invention lies in that: the particle expression with the time t dimension generated in the initialization in step S3 is:

[0049] X i,d,t = (X max - X min ) × rand + X min ,

[0050] V i,d,t = (V max - V min ) × rand + V min ,

[0051] In the formula, i represents the particle ordinal number; d represents the particle position dimension; t represents the moment when the particle is located;

[0052] The learning factors c 1 and c 2 do not change with the number of iterations, and the inertia coefficient ω is updated as the number of iterations n changes:

[0053]

[0054] In the formula, N is the total number of iterations, ω max is the maximum boundary of the change of the inertia coefficient, and ω min is the minimum boundary of the change of the inertia coefficient.

[0055] A further improvement of the technical solution of the present invention lies in that: the expressions for updating the velocities V i,d,t and positions X i,d,t of all particles at each moment in step S5 are:

[0056]

[0057]

[0058] In the formula, represents the individual optimal position of the i-th particle at the t-th time point in the d-th dimension; represents the global optimal position of the i-th particle at the t-th time point in the d-th dimension; n represents the number of iterations;

[0059] For the updated particle velocity and position , the range constraint conditions are as follows:

[0060]

[0061]

[0062] In the formula, represents the maximum and minimum values of the velocity of the particle at t moments in the d dimension; represents the maximum and minimum boundaries of the position of the particle at the t moment in the d dimension.

[0063] Due to the adoption of the above technical solution, the technical progress achieved by the present invention is as follows:

[0064] 1. The extended time-axis particle swarm optimization algorithm proposed by the present invention is applied to the energy management field of the energy routing system, which can fully consider the state of charge constraint conditions of the battery at future time points and find the optimal scheduling curve based on time. Even when the distributed power capacity is less than the maximum power demand of the system at a certain future moment, the algorithm can still plan ahead to make the battery store sufficient electric energy so that the scheduling power and the state of charge at any time point are within the constraints;

[0065] 2. The extended time-axis particle swarm optimization algorithm proposed by the present invention extends the time axis on the standard particle swarm and endows each optimizing particle with a time dimension. Using the idea of solving problems in this dimension by increasing the dimension, the two-dimensional particle matrix with only the number of particles and the spatial dimension is increased by a third dimension "time", so that the particle swarm can search for the optimal scheduling from multiple spaces simultaneously, and then can take into account each state variable in each space-time at the same time and make it within the constraints;

[0066] 3. The extended time-axis particle swarm optimization algorithm proposed by the present invention overcomes the limitation that the time line of the standard particle swarm can only flow unidirectionally. Compared with the dynamic programming algorithm, it does not require a specific algorithm structure and there is no problem of dimensionality explosion, so that it can search for the optimal curve in multiple space-time ranges and effectively solve the problem that the optimization results of the front and back time points are coupled with each other. By adding constraints to all particles in the whole time period and taking the sum of the optimal fitness of the objective function at each moment in the whole time period as the global optimum on the basis of ensuring that the optimization results at each moment meet the constraints, the particles in each space-time are connected together, so that the particle behaviors at all previous and subsequent time points can affect each other, and finally the optimal curve of the whole time period is solved in the multi-space-time constraint range. The proposed extended time-axis particle swarm optimization algorithm broadens the applicable range of the standard particle swarm, and the problems that the algorithm proposed by the present invention may apply to also include: path planning, warehouse management, loading, resource allocation, capacity determination of microgrid equipment, etc. Description of the Drawings

[0067] Figure 1 is the flow chart of the extended time-axis particle swarm optimization algorithm;

[0068] Figure 2 is the optimization schematic diagram of the extended time-axis particle swarm optimization algorithm;

[0069] Figure 3 It is an optimization scheduling scenario composed of photovoltaic power, grid-connected power, and load;

[0070] Figure 4 It is a line graph of the optimization result of the standard particle swarm optimization;

[0071] Figure 5 It is a line graph of the optimization result of the extended time-axis particle swarm algorithm. Specific implementation manner

[0072] The present invention will be further described in detail below in conjunction with embodiments:

[0073] As Figure 1 shown, an extended time-axis particle swarm algorithm applied to energy routing management includes the following steps:

[0074] Step S1: Use the income from selling hydrogen / oxygen produced by the hydrogen electrolyzer throughout the day, the income from assisting the power grid in peak shaving, and the cost of purchasing electricity from the power grid to form the system income / cost objective function;

[0075] The system income / cost objective function is:

[0076]

[0077] In the formula, is the total amount of electricity taken from the power grid by the energy routing system throughout the day; is the total amount of electricity fed into the power grid by the energy routing system throughout the day; is the total amount of hydrogen produced by the hydrogen electrolyzer throughout the day; is the total amount of hydrogen used by the hydrogen fuel cell throughout the day; C b is the price of taking electricity from the power grid; C s is the price of feeding electricity into the power grid; C H2 is the price of buying and selling hydrogen; C O2 is the price of selling oxygen;

[0078] The expression for the total amount of electricity taken from the power grid by the energy routing system throughout the day is:

[0079]

[0080] In the formula, is the power of the energy routing system buying electricity from the power grid at time t;

[0081] The expression for the total amount of electricity fed into the power grid by the energy routing system throughout the day is:

[0082]

[0083] In the formula, is the power of the energy routing system selling electricity to the power grid at time t;

[0084] The expression for the total hydrogen production of the hydrogen electrolyzer throughout the day is:

[0085]

[0086] Wherein, is the hydrogen production rate of the hydrogen electrolyzer in the energy routing system at time t.

[0087] The expression for the total hydrogen consumption of the hydrogen fuel cell throughout the day is:

[0088]

[0089] Wherein, is the hydrogen consumption rate of the hydrogen fuel cell in the energy routing system at time t.

[0090] Step S2: Set the constraint conditions at each moment based on the hydrogen electrolyzer, hydrogen fuel cell, battery capacity, and state of charge parameters;

[0091] The power P of the hydrogen electrolyzer at time t el (t) has the following constraint conditions:

[0092]

[0093] Wherein, P el,max and P el,min are respectively the maximum and minimum boundaries of the operating power of the hydrogen electrolyzer. The maximum boundary value is 100 kW, and the minimum boundary is 10 kW; is a switch quantity, when the hydrogen electrolyzer operates, otherwise it shuts down;

[0094] The power P of the hydrogen fuel cell at time t fc (t) has the following constraint conditions:

[0095]

[0096] Wherein, P fc,max and P fc,min are respectively the maximum and minimum boundaries of the operating power of the hydrogen fuel cell. The maximum boundary value is 100 kW, and the minimum boundary is 10 kW; is a switch quantity, when the hydrogen fuel cell operates, otherwise it shuts down;

[0097] The charging power of the battery at time t and the discharging power have the following constraint conditions:

[0098]

[0099]

[0100] In the formula, and are the maximum power charging boundary and the maximum power discharging boundary of the battery respectively. The values of the maximum power charging boundary and the maximum power discharging boundary are both 100 kW; σ bat is a switching quantity, σ bat = 1 means the battery is discharging, and σ bat = 0 means the battery is charging;

[0101] The constraint conditions for the state of charge (SOC) of the battery are as follows:

[0102] SOC min ≤ SOC(t) ≤ SOC max ,

[0103] In the formula, SOC max and SOC min are the maximum boundary and the minimum boundary of the state of charge of the battery respectively. The value of the maximum boundary is 0.9, and the value of the minimum boundary is 0.1;

[0104] To ensure power balance in the energy routing system by the battery, the battery should also satisfy the equality constraint. The equality constraint conditions for the battery are as follows:

[0105]

[0106] In the formula, P pv (t) is the output power of the photovoltaic at time t, P load (t) is the power demand of the load at time t, η bat is the charge and discharge efficiency of the battery, η pv is the photovoltaic power generation efficiency, η fc is the hydrogen fuel cell power generation efficiency, η el is the hydrogen production efficiency of the hydrogen electrolyzer, η grid is the power interaction efficiency between the energy routing system and the power grid;

[0107] The interaction constraint conditions for the active power between the energy routing system and the power grid at time t are as follows:

[0108]

[0109]

[0110] In the formula, and are the maximum power buying boundary from the power grid and the maximum power selling boundary to the power grid of the energy routing system respectively; σ grid is a switching quantity, σ grid = 1 means only buying power from the power grid, and σ grid = 0 means only selling power to the power grid.

[0111] Step S3: Initialize parameters including the positions X, moving speeds V, number of iterations n, learning factors c 1 and c 2 , and the inertia coefficient ω, and generate particles with a time t dimension, the principle of which is as Figure 2 shown;

[0112] The expression of the generated particles with a time t dimension after initialization is:

[0113] X i,d,t =(X max -X min )×rand+X min ,

[0114] V i,d,t =(V max -V min )×rand+V min ,

[0115] In the formula, i represents the particle ordinal number; d represents the particle position dimension; t represents the moment when the particle is located;

[0116] The learning factors c 1 and c 2 do not change with the number of iterations, and the inertia coefficient ω is updated as the number of iterations n changes:

[0117]

[0118] In the formula, N is the total number of iterations, ω max is the maximum boundary of the change of the inertia coefficient, and ω min is the minimum boundary of the change of the inertia coefficient.

[0119] Step S4: Calculate the fitness of each particle according to the system revenue / cost objective function in Step S1, and obtain the individual optimal particle and the global optimal particle;

[0120] Step S5: Update the speeds V i,d,t and positions X i,d,t of all particles at each moment according to the positions of the individual optimal particle and the global optimal particle, and constrain the ranges of the updated particle speeds and positions;

[0121] The expressions for updating the speeds V i,d,t and positions X i,d,t of all particles at each moment are:

[0122]

[0123]

[0124] In the formula, represents the individual optimal position of the i-th particle at the t-th time point in the d-th dimension; represents the global optimal position of the i-th particle at the t-th time point in the d-th dimension; n represents the number of iterations;

[0125] For the updated particle velocity and position the range constraint conditions are as follows:

[0126]

[0127]

[0128] In the formula, represents the maximum and minimum values of the velocity of the particle at the t-th moment in the d-th dimension; represents the maximum and minimum boundaries of the position of the particle at the t-th moment in the d-th dimension.

[0129] Step S6: Determine whether the current number of iterations has reached the maximum number of iterations. If so, stop the iteration and output the optimal solution; otherwise, return to Step S4 to continue the iterative calculation.

[0130] The objective of the present invention is to maximize the objective function, i.e., the system revenue result. There are several parameters of the extended time-axis particle swarm algorithm that need to set initial values. Let: the number of particles be 3000, the learning factor c 1 be 2, c 2 be 10, the inertia coefficient ω linearly decreases from 0.1 to 1.1, and the maximum number of iterations is 200.

[0131] The present invention conducts economic optimal scheduling for an energy routing system containing a photovoltaic cell, a hydrogen electrolyzer, a hydrogen fuel cell, a storage battery, a grid-connected converter, and a load, proposes an extended time-axis particle swarm algorithm to solve the objective function, obtains the optimal solution of the objective function, that is, obtains the hydrogen electrolyzer power, hydrogen production rate, hydrogen fuel cell power, hydrogen consumption rate, storage battery power, and state of charge SOC of the energy routing system at each moment during the scheduling period, thereby improving the economy of the energy routing system operation. Using the optimization scheduling scenario composed of photovoltaic power, grid-connected power, and load as shown in Figure 3 After simulating and comparing the standard particle swarm algorithm and the algorithm of the present invention, it is proved that the extended time-axis particle swarm algorithm can obtain the maximum revenue scheduling for the whole period, while the standard particle swarm algorithm can only perform single-point optimization one by one during the long-term optimization process and cannot consider whether the SOC of the storage battery at future moments can be within the constraints. As shown in Figure 4As shown, within 1 - 2 hours after startup, the hydrogen electrolytic cell will operate at maximum power in the optimization result of the standard particle swarm algorithm to produce as much hydrogen as possible. As a result, the electrical energy in the battery will be quickly depleted, causing the battery SOC to reach the lower limit. And due to the goal of maximizing economic benefits, the energy management system will dispatch the hydrogen fuel cell to generate electricity as little as possible because this means a reduction in economic benefits as hydrogen is consumed. Therefore, the battery SOC will remain slightly above the lower limit for a long time. Although the SOC has not exceeded the lower limit at this time, the ability of the energy routing system to handle sudden increases in load has become very low.

[0132] However, at 21 - 22 hours, due to the load P load and the power fed into the power grid P grid the sum of which exceeds the rated power of the hydrogen fuel cell, 100 kW. But at this time, the SOC is still near the lower boundary of 0.1, and the battery does not have enough electrical energy to handle the suddenly increased load, resulting in the battery SOC exceeding the lower limit of 0.1. This problem is caused by the fact that the standard PSO algorithm can only optimize single points one by one in a long time period and cannot take into account the scheduling constraints at future time points, which will endanger the reliable and stable operation of the system.

[0133] The present invention utilizes the fast search ability of the particle swarm algorithm, extends the time axis on the standard particle swarm, endows each optimized particle with a time dimension, and proposes an improved extended time - axis particle swarm algorithm. Similarly, in the Figure 3 optimization scheduling scenario shown, the scheduling result of the algorithm of the present invention is as Figure 5 shown. This extension overcomes the limitation that the time line of the standard particle swarm can only flow unidirectionally, enables the particles to search for the optimal curve in multiple space - time ranges, and effectively solves the problem that the scheduling at any time point has a two - way coupling with the scheduling at all other moments along the time axis.

Claims

1. An extended time-axis particle swarm optimization algorithm applied to energy routing management, characterized in that: It includes the following steps: Step S1: Use the daily 24-hour revenue from selling hydrogen / oxygen produced by the hydrogen electrolyzer, the revenue from assisting the power grid in peak shaving, and the cost of purchasing electricity from the power grid to form the system revenue / cost objective function; Step S2: Based on the hydrogen electrolyzer, hydrogen fuel cell, battery capacity, and state of charge parameters, set the constraint conditions at each moment; Step S3: Initialize various parameters including the positions X, movement speeds V, number of iterations n, learning factors c 1 and c 2 , and the inertia coefficient ω, and generate particles with the time t dimension; Step S4: Calculate the fitness of each particle according to the system revenue / cost objective function in Step S1 to obtain the individual optimal particle and the global optimal particle; Step S5: Update the velocity V of all particles at each moment according to the positions of the individual optimal particle and the swarm optimal particle i,d,t and the position X i,d,t , and constrain the updated particle velocity and position within a range Step S6: Determine whether the current iteration number reaches the maximum iteration number. If it reaches, stop the iteration and output the optimal solution; otherwise, return to Step S4 to continue the iterative calculation.

2. An extended time-axis particle swarm optimization algorithm applied to energy routing management according to Claim 1, characterized in that: The system revenue / cost objective function in Step S1 is: Wherein, is the total amount of electricity taken from the power grid by the energy routing system throughout the day; is the total amount of electricity fed into the power grid by the energy routing system throughout the day; is the total amount of hydrogen produced by the hydrogen electrolyzer throughout the day; is the total amount of hydrogen consumed by the hydrogen fuel cell throughout the day; C b is the electricity purchase price from the power grid; C s is the electricity feed-in price to the power grid; C H2 is the hydrogen buying and selling price; C O2 is the oxygen selling price; The expression for the total amount of electricity taken from the power grid by the energy routing system throughout the day is: Wherein, is the electric power purchased by the energy routing system from the power grid at time t; The expression for the total amount of electricity fed back to the power grid by the energy routing system throughout the day is: In the formula, is the electric power sold by the energy routing system to the power grid at time t; The expression for the total amount of hydrogen produced by the hydrogen electrolyzer throughout the day is: In the formula, is the hydrogen production rate of the hydrogen electrolyzer in the energy routing system at time t; The expression for the total amount of hydrogen consumed by the hydrogen fuel cell throughout the day is: Wherein, is the hydrogen consumption rate of the hydrogen fuel cell in the energy routing system at time t.

3. An extended time-axis particle swarm optimization algorithm applied to energy routing management according to Claim 2, characterized in that: The constraint condition of the power P el (t) of the hydrogen electrolyzer at time t is as follows: Wherein, P el,max and P el,min are respectively the maximum boundary and the minimum boundary of the operating power of the hydrogen electrolyzer; is a digital quantity, when the hydrogen electrolyzer operates, otherwise it shuts down; Power P of the hydrogen fuel cell at time t fc (t) has the following constraint conditions: Wherein, P fc,max and P fc,min are respectively the maximum boundary and the minimum boundary of the operating power of the hydrogen fuel cell; is a switching quantity, when the hydrogen fuel cell operates, otherwise it shuts down; Battery charging power at time t and discharge power are subject to the following constraints: In the formula, and are the maximum power charging boundary and the maximum power discharging boundary of the storage battery respectively; σ bat is a switching value. When σ bat = 1, the storage battery discharges. When σ bat = 0, the storage battery charges. The constraint conditions for the state of charge SOC of the battery are as follows: SOC min ≤SOC(t)≤SOC max , where SOC max and SOC min are the maximum and minimum boundaries of the state of charge of the battery, respectively; The equality constraint conditions for the battery are as follows: Where, P pv (t) is the output power of the photovoltaic at time t, P load (t) is the power demand of the load at time t, η bat is the charge and discharge efficiency of the battery, η pv is the photovoltaic power generation efficiency, η fc is the power generation efficiency of the hydrogen fuel cell, η el is the hydrogen production efficiency of the hydrogen electrolyzer, η grid is the electrical energy interaction efficiency between the energy routing system and the power grid; The interactive constraint conditions for the active power between the energy routing system and the power grid at time t are as follows: In the formula, and are the maximum power boundaries for the energy routing system to buy electricity from the power grid and sell electricity to the power grid respectively; σ grid is a switch quantity. When σ grid = 1, only buy electricity from the power grid. When σ grid = 0, only sell electricity to the power grid.

4. An extended time-axis particle swarm optimization algorithm applied to energy routing management according to Claim 3, characterized in that: The particle expression with the time t dimension generated in the initialization in Step S3 is: X i,d,t = (X max - X min ) × rand + X min , V i,d,t = (V max - V min ) × rand + V min , In the formula, i represents the particle ordinal number; d represents the particle position dimension; t represents the moment when the particle is located; Learning factor c 1 and c 2 do not change with the number of iterations. The inertia coefficient ω is updated as the number of iterations n changes: where N is the total number of iterations, ω max is the maximum boundary of the change in the inertia coefficient, ω min is the minimum boundary of the change in the inertia coefficient.

5. An extended time-axis particle swarm optimization algorithm applied to energy routing management according to Claim 4, characterized in that: In step S5, update the velocity V of all particles at each moment i,d,t and the position X i,d,t The expression is as follows: In the formula, represents the individual optimal position of the i-th particle at the t-th time point in the d-th dimension; represents the global optimal position of the i-th particle at the t-th time point in the d-th dimension; n represents the number of iterations; For the updated particle velocity and position the range constraint conditions are as follows: Wherein, represents the maximum and minimum velocities of the particle at t moments in the d dimension; represents the maximum and minimum boundaries of the position of the particle at the t-th moment in the d dimension.

Images