Online secondary frequency modulation method considering regulation capacity polymerization of hydrogen fuel cell vehicle

Abstract

The invention discloses an online secondary frequency modulation method considering regulation capacity aggregation of a hydrogen fuel cell vehicle. The method comprises the following steps: 1) constructing a power system operation model of the hydrogen fuel cell vehicle HFCV based on a hydrogen fuel cell vehicle cluster; 2) constructing a Minkowski sum-based hydrogen fuel cell vehicle frequency modulation capacity aggregation model; 3) solving the frequency modulation capacity aggregation model of the hydrogen fuel cell vehicle to obtain a frequency modulation capacity feasible region of an HFCV cluster; 4) on the basis of the power system operation model, constructing an online secondary frequency modulation model of the hydrogen fuel cell vehicle; 5) based on the frequency modulation capacity feasible region, the system scheduling mechanism adopts a Lyapunov optimization method to solve the online secondary frequency modulation model of the hydrogen fuel cell vehicle to obtain an AGC instruction signal, and the AGC instruction signal is issued to an aggregator; and 6) the aggregator distributes the AGC instruction to each hydrogen fuel cell vehicle in real time to participate in the secondary frequency modulation of the power system. According to the method, AGC signal prediction information is not needed, and the frequency response and frequency modulation operation cost can be better balanced.

Description

Technical Field

[0001] The present invention relates to the technical field of new energy power systems, and in particular to an online secondary frequency modulation method considering the aggregation of hydrogen fuel cell vehicle regulation capacity. Background Art

[0002] Many traditional fossil fuel power plants are being replaced by renewable energy sources (RES), which are expected to become the primary source of electricity in the near future. However, the significant variation in renewable energy sources poses a significant challenge to maintaining frequency stability. Furthermore, hydrogen fuel cell vehicles (HFCVs), which no longer rely on fossil fuels, are expected to play a major role in future transportation systems, reducing carbon emissions from transportation.

[0003] HFCVs' energy supply consists of hydrogen fuel cells and batteries, which have different operating characteristics. Due to their rapid response, these fuel cells and batteries can control their charge and discharge power to mitigate frequency fluctuations. However, the individual charge or discharge power of a single HFCV is limited and insufficient to provide frequency services. Therefore, it is necessary to coordinate multiple HFCVs in the form of an aggregator, thereby establishing an aggregation model for grid-connected HFCVs. This aggregation model quantifies the aggregate frequency regulation capability of the HFCVs to ensure sufficient frequency reserves, which the aggregator can effectively manage. Due to the differences in operating characteristics between fuel cells and batteries, the feasible domain of the aggregation model's frequency regulation capability is equal to the Minkowski sum of the feasible domains of all fuel cells and batteries. However, the exact Minkowski sum between any arbitrary polyhedron is computationally intractable. In addition to providing aggregate frequency regulation capability to the system operator, the HFCV aggregator also distributes automatic gearing (AGC) tasks assigned by the higher-level system to each HFCV. However, the AGC signal is uncertain and difficult to accurately predict. To quickly and efficiently maintain frequency within a normal range, it is necessary to develop an online optimization method that determines the response strategy of each HFCV based on the current AGC signal. Summary of the Invention

[0004] The object of the present invention is to provide an online secondary frequency modulation method considering the aggregation of hydrogen fuel cell vehicle regulation capacity, comprising the following steps:

[0005] 1) Aggregators record the number of hydrogen fuel cell vehicle clusters connected to decentralized charging and discharging piles, as well as the time these hydrogen fuel cell vehicles reported leaving the vehicle;

[0006] 2) Based on the hydrogen fuel cell vehicle cluster, build a power system operation model for hydrogen fuel cell vehicles (HFCVs);

[0007] 3) Construct a frequency regulation capacity aggregation model for hydrogen fuel cell vehicles based on Minkowski sum;

[0008] 4) Solve the frequency regulation capacity aggregation model of hydrogen fuel cell vehicles and obtain the feasible region of frequency regulation capacity of HFCV clusters;

[0009] 5) Based on the power system operation model, build an online secondary frequency modulation model for hydrogen fuel cell vehicles;

[0010] 6) Based on the feasible region of frequency regulation capacity, the system dispatching agency uses the Lyapunov optimization method to solve the online secondary frequency regulation model of hydrogen fuel cell vehicles, obtains the AGC command signal, and sends it to the aggregator;

[0011] 7) The aggregator distributes AGC instructions to each hydrogen fuel cell vehicle in real time and participates in the secondary frequency regulation of the power system.

[0012] Furthermore, the power system operation model of the hydrogen fuel cell vehicle HFCV includes a hydrogen fuel cell operation model and an electrochemical energy storage operation model.

[0013] Furthermore, considering the discharge power limit and the hydrogen storage capacity limit of the hydrogen storage tank, the hydrogen fuel cell operation model is as follows:

[0014]

[0015] Where, is the maximum discharge power of the fuel cell of the g-th HFCV; is the remaining hydrogen storage capacity of the g-th HFCV at time t; τ is the time step of the AGC signal. and are the lower and upper limits of hydrogen storage capacity respectively; P t h ,max is the maximum discharge power of all fuel cells in the HFCV cluster at time t; is the start and end time;

[0016] Among them, the hydrogen consumption of hydrogen fuel cells during V2G is as follows:

[0017]

[0018] Where g is the set of hydrogen fuel cell vehicles The index of t is the time set The index of is the discharge power of the fuel cell of the g-th HFCV at time t; m g,t is the hydrogen consumption of the g-th HFCV at time t; η h is the energy conversion efficiency; HHV is the higher heating value coefficient of hydrogen.

[0019] Furthermore, considering the charging and discharging power limitations and energy limitations, the electrochemical energy storage operation model is as follows:

[0020]

[0021] Where, and are the charging power and discharging power of the energy storage of the g-th HFCV at time t, respectively. is the maximum discharge power of the energy storage of the g-th HFCV. is the remaining energy stored in the g-th HFCV at time t. and are the lower and upper limits of energy storage respectively; P t b,max is the maximum charge and discharge power of all energy storage in the HFCV cluster at time t; η c and η d It is the charge and discharge efficiency.

[0022] Furthermore, the hydrogen fuel cell vehicle frequency regulation capacity aggregation model based on the Minkowski sum includes the fuel cell aggregation model within the HFCV cluster and the energy storage aggregation model;

[0023] The fuel cell aggregation model includes discharge power constraints, frequency modulation capacity constraints, and hydrogen storage capacity constraints;

[0024] The energy storage aggregation model includes charging and discharging power constraints, frequency regulation capacity constraints and energy constraints.

[0025] Further, the fuel cell aggregation model is shown below:

[0026]

[0027] Where, P t h M is the discharge power of the fuel cell in the HFCV cluster at time t. t is the hydrogen consumption of the HFCV cluster at time t; The frequency modulation capacity that can be provided by fuel cells in the HFCV cluster; is the remaining hydrogen storage capacity of the HFCV cluster at time t;

[0028] Among them, the change in hydrogen storage energy caused by the change in the grid connection state of the HFCV cluster at time t The state variable X that represents the vehicle grid connection state g,t As shown below:

[0029]

[0030] Where, and are the hydrogen storage capacity of HFCV when it is connected to the grid and when it is off the grid; X g,t=1 means the g-th vehicle is connected to the grid at time t;

[0031] The energy storage aggregation model is as follows:

[0032]

[0033] Where, P t b,c and P t b,d are the charging and discharging powers of the energy storage in the HFCV cluster at time t, respectively; The frequency regulation capacity that energy storage can provide within the HFCV cluster. is the remaining energy stored in the HFCV cluster at time t;

[0034] Among them, the energy storage energy change caused by the change of the grid connection state of the HFCV cluster at time t As shown below:

[0035]

[0036] Where, and They are the energy storage capacity of HFCV when it is connected to the grid and when it is off the grid.

[0037] Furthermore, the feasible domain of the frequency modulation capacity of the HFCV cluster is as follows:

[0038]

[0039] Where, It represents the feasible domain of fuel cell frequency regulation capacity within the HFCV cluster. A represents the controllable variable of the fuel cell in the HFCV cluster. h and B h is the coefficient matrix; It represents the feasible domain of energy storage frequency regulation capacity within the HFCV cluster. A represents the controllable variable of energy storage in the HFCV cluster. b and B b is the coefficient matrix; represents the Minkowski sum. U represents a point on the boundary of the feasible region of the set.

[0040] Furthermore, the steps for solving the frequency modulation capacity aggregation model of hydrogen fuel cell vehicles include:

[0041] 3.1) Integrate HFCV cluster grid connection information They represent grid connection time, off-grid time, grid-connected energy storage, grid-connected hydrogen storage, and off-grid hydrogen storage respectively;

[0042] 3.2) Based on the hydrogen fuel cell vehicle frequency modulation capacity aggregation model, the feasible domain of the frequency modulation capacity of fuel cells and batteries in the HFCV cluster is calculated separately;

[0043] 3.3) According to the set feasible region boundary formula U=U b +U h , determine the Minkowskisum boundary vertex O of the feasible region of HFCV cluster frequency modulation capacity j ;

[0044] 3.4) Construct the sorting point set Ω={O1,O2,...,O J}; Among them, the minimum point O1 in the sorted point set is selected as the reference point;

[0045] 3.5) If The boundary vertex O j Write the Minkowski sum vertex set of the frequency modulation capacity feasible domain middle;

[0046] 3.6) Determine whether j>J holds. If not, set j=j+1 and return to step 3.5). If so, proceed to step 3.7);

[0047] 3.7) Minkowski sum vertex set of the feasible region of the connection frequency modulation capacity Each point determines the exact feasible region of frequency regulation capability

[0048] Furthermore, the steps of constructing an online secondary frequency modulation model for hydrogen fuel cell vehicles include:

[0049] 5.1) Convert the time coupling constraints (2) and (7) into time average constraints, that is:

[0050]

[0051] Where E is the expectation; parameter

[0052] 5.2) To reduce the life cycle cost of HFCVs and allocate SFR power to each HFCV, an online secondary frequency regulation model for hydrogen fuel cell vehicles is constructed, namely:

[0053]

[0054] st(1),(4),(5)-(6),(25)-(26)(29)

[0055]

[0056] Where, constraints (30) and (31) prohibit HFCV g from being charged and discharged simultaneously. Constraints (32) and (33) determine the charge and discharge state of HFCV g.

[0057] Among them, the degradation cost of fuel cells is as follows:

[0058]

[0059] Where, is the degradation cost of the fuel cell at time t; is the fuel cell cost coefficient;

[0060] Fuel cell degradation coefficient d caused by V2G h As shown below:

[0061]

[0062] Where d1 represents the degradation rate of the fuel cell under maximum power conditions;

[0063] In the V2G process, the hydrogen consumption cost of the fuel cell is as follows:

[0064]

[0065] Where, is the hydrogen consumption cost of HFCV g at time t; is the hydrogen cost coefficient;

[0066] In the SFR process, the energy storage charging and discharging degradation cost is as follows:

[0067]

[0068] Where, represents the energy storage degradation cost of HFCV g at time t;

[0069] Battery degradation cost coefficient As shown below:

[0070]

[0071] Where λ g N represents the battery investment cost of HFCV g. b Energy storage under energy constraints the number of cycles that can be performed;

[0072] Frequency Regulation Mismatch Penalty Cost As shown below:

[0073]

[0074] Where r t is the normalized AGC signal, r t ∈[-1,1]. R t is the total frequency regulation capacity determined by the HFCV aggregator. p is the penalty cost coefficient.

[0075] Furthermore, the steps of solving the online secondary frequency modulation model of hydrogen fuel cell vehicles using the Lyapunov optimization method include:

[0076] 6.1) Construct a virtual queue and convert the time average constraint into a queue stability constraint. The steps include:

[0077] 6.1.1) Construct a virtual hydrogen energy queue for fuel cells and an expression for the dynamic characteristics of the virtual hydrogen energy queue, namely:

[0078]

[0079] Where, is the disturbance parameter;

[0080] 6.1.2) Constructing a virtual hydrogen fleet H for HFCV g,t The average rate stability constraint is as follows:

[0081]

[0082] 6.1.3) Construct a virtual energy queue for energy storage and an expression for the dynamic characteristics of the virtual energy queue, namely:

[0083]

[0084] Where, is the disturbance parameter;

[0085] 6.1.4) Constructing Virtual Energy Queue B g,t The average rate stability constraint is:

[0086]

[0087] 6.1.5) Combine steps 6.1.1)-6.1.4) to construct a concatenated vector Θ of the virtual queue t =(H t ,B t ),Right now

[0088]

[0089] 6.2) Define the Lyapunov function to obtain Lyapunov drift and drift plus penalty;

[0090] Among them, in the size of the virtual hydrogen and battery energy queues, the Lyapunov function L(Θ t ) is as follows:

[0091]

[0092] Lyapunov drift Δ(Θ t ) is as follows:

[0093]

[0094] Where, Θ t is a random variable.

[0095] The drift penalty is as follows:

[0096]

[0097] Where V is the weight coefficient, which is used to balance the stability and operating cost of the virtual queue; F t For operating costs;

[0098] 6.3) Construct the Lyapunov drift expression within the AGC signal gap, namely:

[0099]

[0100] 6.4) Substituting formulas (55)-(56) into formula (53), minimizing the upper limit of drift plus penalty, we obtain

[0101]

[0102] Where, is the maximum hydrogen consumption of the g-th HFCV;

[0103] 6.6) Based on formula (57), the online SFR problem of each AGC signal time slot is constructed, that is:

[0104]

[0105] 6.7) Before the start of each AGC signal time slot, update the virtual queue using equations (44) and (47);

[0106] 6.8) Based on the given current virtual queue, total frequency regulation capacity and AGC signal, solve the online SFR problem of each AGC signal time slot and determine the AGC command signal for each HFCV.

[0107] The technical effect of the present invention is unquestionable. The present invention is applicable to power grids with a high proportion of new energy. The present invention fully considers the operating characteristics of hydrogen fuel cell vehicles and proposes an online secondary frequency regulation method that considers the aggregation of hydrogen fuel cell vehicle regulation capacity. This method does not require AGC signal prediction information and can better balance frequency response and frequency regulation operating costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0108] Figure 1 It is a flow chart of an online secondary frequency regulation algorithm considering the aggregation of hydrogen fuel cell vehicle regulation capacity;

[0109] Figure 2 It is the feasible domain diagram of aggregate frequency regulation capacity of hydrogen fuel cell vehicles;

[0110] Figure 3 This is a comparison of the total FM power output of the HFCV cluster under different methods;

[0111] Figure 4 It is the energy change of HFCV in response to AGC signal;

[0112] Figure 5 is the gap between different AGC signal sequences. DETAILED DESCRIPTION

[0113] The present invention will be further described below with reference to the following examples, but it should not be understood that the scope of the present invention is limited to the following examples. Without departing from the above technical ideas of the present invention, various substitutions and modifications can be made according to common technical knowledge and customary means in the art, and all should be included in the scope of protection of the present invention.

[0114] Example 1:

[0115] See also Figures 1 to 3 , an online secondary frequency regulation method considering the aggregation of hydrogen fuel cell vehicle regulation capacity, comprising the following steps:

[0116] 1) Aggregators record the number of hydrogen fuel cell vehicle clusters connected to decentralized charging and discharging piles, as well as the time these hydrogen fuel cell vehicles reported leaving the vehicle;

[0117] 2) Based on the hydrogen fuel cell vehicle cluster, build a power system operation model for hydrogen fuel cell vehicles (HFCVs);

[0118] 3) Construct a frequency regulation capacity aggregation model for hydrogen fuel cell vehicles based on Minkowski sum;

[0119] 4) Solve the frequency regulation capacity aggregation model of hydrogen fuel cell vehicles and obtain the feasible region of frequency regulation capacity of HFCV clusters;

[0120] 5) Based on the power system operation model, build an online secondary frequency modulation model for hydrogen fuel cell vehicles;

[0121] 6) Based on the feasible region of frequency regulation capacity, the system dispatching agency uses the Lyapunov optimization method to solve the online secondary frequency regulation model of hydrogen fuel cell vehicles, obtains the AGC command signal, and sends it to the aggregator;

[0122] 7) The aggregator distributes AGC instructions to each hydrogen fuel cell vehicle in real time and participates in the secondary frequency regulation of the power system.

[0123] Example 2:

[0124] An online secondary frequency modulation method considering the aggregation of hydrogen fuel cell vehicle regulation capacity, the technical content is the same as that of Example 1, further, the power system operation model of the hydrogen fuel cell vehicle HFCV includes a hydrogen fuel cell operation model and an electrochemical energy storage operation model.

[0125] Example 3:

[0126] An online secondary frequency modulation method considering the aggregation of hydrogen fuel cell vehicle regulation capacity, the technical content is the same as any one of Examples 1-2, further considering the discharge power limit and the hydrogen storage capacity limit of the hydrogen storage tank, the hydrogen fuel cell operation model is as follows:

[0127]

[0128] Where, is the maximum discharge power of the fuel cell of the g-th HFCV; is the remaining hydrogen storage capacity of the g-th HFCV at time t; τ is the time step of the AGC signal. and are the lower and upper limits of hydrogen storage capacity respectively; P t h ,max is the maximum discharge power of all fuel cells in the HFCV cluster at time t; is the start and end time;

[0129] Among them, the hydrogen consumption of hydrogen fuel cells during V2G is as follows:

[0130]

[0131] Where g is the set of hydrogen fuel cell vehicles The index of t is the time set The index of is the discharge power of the fuel cell of the g-th HFCV at time t; m g,t is the hydrogen consumption of the g-th HFCV at time t; η his the energy conversion efficiency; HHV is the higher heating value coefficient of hydrogen.

[0132] Example 4:

[0133] An online secondary frequency modulation method considering the capacity aggregation of hydrogen fuel cell vehicles is provided. The technical content is the same as any one of Examples 1-3. Furthermore, considering the charge and discharge power limit and energy limit, the electrochemical energy storage operation model is as follows:

[0134]

[0135] Where, and are the charging power and discharging power of the energy storage of the g-th HFCV at time t, respectively. is the maximum discharge power of the energy storage of the g-th HFCV. is the remaining energy stored in the g-th HFCV at time t. and are the lower and upper limits of energy storage respectively; P t b,max is the maximum charge and discharge power of all energy storage in the HFCV cluster at time t; η c and η d It is the charge and discharge efficiency.

[0136] Example 5:

[0137] An online secondary frequency modulation method considering hydrogen fuel cell vehicle regulation capacity aggregation, the technical content is the same as any one of Examples 1-4, further, the hydrogen fuel cell vehicle frequency modulation capacity aggregation model based on Minkowski sum includes a fuel cell aggregation model within the HFCV cluster and an energy storage aggregation model;

[0138] The fuel cell aggregation model includes discharge power constraints, frequency modulation capacity constraints, and hydrogen storage capacity constraints;

[0139] The energy storage aggregation model includes charging and discharging power constraints, frequency regulation capacity constraints and energy constraints.

[0140] Example 6:

[0141] An online secondary frequency modulation method considering the aggregation of hydrogen fuel cell vehicle regulation capacity, the technical content is the same as any one of Examples 1-5, further, the fuel cell aggregation model is as follows:

[0142]

[0143]

[0144] Where, P t hM is the discharge power of the fuel cell in the HFCV cluster at time t. t is the hydrogen consumption of the HFCV cluster at time t; The frequency modulation capacity that can be provided by fuel cells in the HFCV cluster; is the remaining hydrogen storage capacity of the HFCV cluster at time t;

[0145] Among them, the change in hydrogen storage energy caused by the change in the grid connection state of the HFCV cluster at time t The state variable X that represents the vehicle grid connection state g,t As shown below:

[0146]

[0147] Where, and are the hydrogen storage capacity of HFCV when it is connected to the grid and when it is off the grid; X g,t =1 means the g-th vehicle is connected to the grid at time t;

[0148] The energy storage aggregation model is as follows:

[0149]

[0150] Where, P t b,c and P t b,d are the charging and discharging powers of the energy storage in the HFCV cluster at time t, respectively; The frequency regulation capacity that energy storage can provide within the HFCV cluster. is the remaining energy stored in the HFCV cluster at time t;

[0151] Among them, the energy storage energy change caused by the change of the grid connection state of the HFCV cluster at time t As shown below:

[0152]

[0153] Where, and They are the energy storage capacity of HFCV when it is connected to the grid and when it is off the grid.

[0154] Example 7:

[0155] An online secondary frequency regulation method considering the aggregation of hydrogen fuel cell vehicle regulation capacity, the technical content is the same as any one of Examples 1-6, further, the feasible domain of the frequency regulation capacity of the HFCV cluster is as follows:

[0156]

[0157] Where, It represents the feasible domain of fuel cell frequency regulation capacity within the HFCV cluster. A represents the controllable variable of the fuel cell in the HFCV cluster. h and B h is the coefficient matrix; It represents the feasible domain of energy storage frequency regulation capacity within the HFCV cluster. A represents the controllable variable of energy storage in the HFCV cluster. b and B b is the coefficient matrix; represents the Minkowski sum. U represents a point on the boundary of the feasible region of the set.

[0158] Example 8:

[0159] An online secondary frequency modulation method considering hydrogen fuel cell vehicle regulation capacity aggregation, the technical content of which is the same as any one of Examples 1-7, further comprising the step of solving a hydrogen fuel cell vehicle frequency modulation capacity aggregation model comprising:

[0160] 3.1) Integrate HFCV cluster grid connection information They represent grid connection time, off-grid time, grid-connected energy storage, grid-connected hydrogen storage, and off-grid hydrogen storage respectively;

[0161] 3.2) Based on the hydrogen fuel cell vehicle frequency modulation capacity aggregation model, the feasible domain of the frequency modulation capacity of fuel cells and batteries in the HFCV cluster is calculated separately;

[0162] 3.3) According to the set feasible region boundary formula U=U b +U h , determine the Minkowskisum boundary vertex O of the feasible region of HFCV cluster frequency modulation capacity j ;

[0163] 3.4) Construct the sorting point set Ω={O1,O2,...,O J}; Among them, the minimum point O1 in the sorted point set is selected as the reference point;

[0164] 3.5) If The boundary vertex O j Write the Minkowski sum vertex set of the frequency modulation capacity feasible domain middle;

[0165] 3.6) Determine whether j>J holds. If not, set j=j+1 and return to step 3.5). If so, proceed to step 3.7);

[0166] 3.7) Minkowski sum vertex set of the feasible region of the connection frequency modulation capacity Each point determines the exact feasible region of frequency regulation capability Example 9:

[0167] An online secondary frequency modulation method considering the aggregation of hydrogen fuel cell vehicle regulation capacity, the technical content of which is the same as any one of Examples 1-8, further comprising the steps of constructing an online secondary frequency modulation model for a hydrogen fuel cell vehicle:

[0168] 5.1) Convert the time coupling constraints (2) and (7) into time average constraints, that is:

[0169]

[0170]

[0171] Where E is the expectation; parameter

[0172] 5.2) To reduce the life cycle cost of HFCVs and allocate SFR power to each HFCV, an online secondary frequency regulation model for hydrogen fuel cell vehicles is constructed, namely:

[0173]

[0174] st(1),(4),(5)-(6),(25)-(26)(29)

[0175]

[0176] Where, constraints (30) and (31) prohibit HFCV g from being charged and discharged simultaneously. Constraints (32) and (33) determine the charge and discharge state of HFCV g.

[0177] Among them, the degradation cost of fuel cells is as follows:

[0178]

[0179] Where, is the degradation cost of the fuel cell at time t; is the fuel cell cost coefficient;

[0180] Fuel cell degradation coefficient d caused by V2G h As shown below:

[0181]

[0182] Where d1 represents the degradation rate of the fuel cell under maximum power conditions;

[0183] In the V2G process, the hydrogen consumption cost of the fuel cell is as follows:

[0184]

[0185] Where, is the hydrogen consumption cost of HFCV g at time t; is the hydrogen cost coefficient;

[0186] In the SFR process, the energy storage charging and discharging degradation cost is as follows:

[0187]

[0188]

[0189] Where, represents the energy storage degradation cost of HFCV g at time t;

[0190] Battery degradation cost coefficient As shown below:

[0191]

[0192] Where λ g N represents the battery investment cost of HFCV g. b Energy storage under energy constraints the number of cycles that can be performed;

[0193] Frequency Regulation Mismatch Penalty Cost As shown below:

[0194]

[0195] Where r t is the normalized AGC signal, r t ∈[-1,1]. R t is the total frequency regulation capacity determined by the HFCV aggregator. p is the penalty cost coefficient.

[0196] Example 10:

[0197] An online secondary frequency modulation method considering the aggregation of hydrogen fuel cell vehicle regulation capacity, the technical content of which is the same as any one of Examples 1-9, further comprising the steps of using the Lyapunov optimization method to solve the online secondary frequency modulation model of the hydrogen fuel cell vehicle, comprising:

[0198] 6.1) Construct a virtual queue and convert the time average constraint into a queue stability constraint. The steps include:

[0199] 6.1.1) Construct a virtual hydrogen energy queue for fuel cells and an expression for the dynamic characteristics of the virtual hydrogen energy queue, namely:

[0200]

[0201] Where, is the disturbance parameter;

[0202] 6.1.2) Constructing a virtual hydrogen fleet H for HFCV g,t The average rate stability constraint is as follows:

[0203]

[0204] 6.1.3) Construct a virtual energy queue for energy storage and an expression for the dynamic characteristics of the virtual energy queue, namely:

[0205]

[0206] Where, is the disturbance parameter;

[0207] 6.1.4) Constructing Virtual Energy Queue B g,t The average rate stability constraint is:

[0208]

[0209] 6.1.5) Combine steps 6.1.1)-6.1.4) to construct a concatenated vector Θ of the virtual queue t =(H t ,B t ),Right now

[0210]

[0211] 6.2) Define the Lyapunov function to obtain Lyapunov drift and drift plus penalty;

[0212] Among them, in the size of the virtual hydrogen and battery energy queues, the Lyapunov function L(Θ t ) is as follows:

[0213]

[0214] Lyapunov drift Δ(Θ t ) is as follows:

[0215]

[0216] Where, Θ t is a random variable.

[0217] The drift penalty is as follows:

[0218]

[0219] Where V is the weight coefficient, which is used to balance the stability and operating cost of the virtual queue; F t For operating costs;

[0220] 6.3) Construct the Lyapunov drift expression within the AGC signal gap, namely:

[0221]

[0222] 6.4) Substituting formulas (55)-(56) into formula (53), minimizing the upper limit of drift plus penalty, we obtain

[0223]

[0224] Where, is the maximum hydrogen consumption of the g-th HFCV;

[0225] 6.6) Based on formula (57), the online SFR problem of each AGC signal time slot is constructed, that is:

[0226]

[0227] 6.7) Before the start of each AGC signal time slot, update the virtual queue using equations (44) and (47);

[0228] 6.8) Based on the given current virtual queue, total frequency regulation capacity and AGC signal, solve the online SFR problem of each AGC signal time slot and determine the AGC command signal for each HFCV.

[0229] Example 11:

[0230] An online secondary frequency modulation method considering the aggregation of hydrogen fuel cell vehicle regulation capacity, the steps are as follows:

[0231] 1) The main steps of the secondary frequency modulation method based on the operational flexibility of hydrogen fuel cell vehicles are as follows:

[0232] Hydrogen fuel cell vehicles (HFCVs) provide online secondary frequency regulation (SFR) services to the power system. When a vehicle owner arrives at work or returns home, they connect their HFCV to the nearest decentralized charging and discharging station and report their departure time to the aggregator. HFCV aggregators collect information about vehicles providing vehicle-to-grid (V2G) services. Based on the Minkowski sum, aggregators generate frequency regulation capacity for each time period. This frequency regulation capacity, as reported by the aggregator, participates in market bidding for frequency regulation ancillary services. Based on the market bidding results, the system dispatcher sends an automatic generation control (AGC) signal to the aggregator. Based on the Lyapunov optimization method, the aggregator distributes AGC instructions to each HFCV in real time, participating in the secondary frequency regulation of the power system.

[0233] The HFCV's power system consists of a hydrogen fuel cell and electrochemical energy storage. The fuel cell can store and release energy and transfer energy between different times. The hydrogen consumption of the fuel cell during V2G can be expressed as:

[0234]

[0235] Where g is the set of hydrogen fuel cell vehicles The index of t is the time set The index of is the discharge power of the fuel cell of the g-th HFCV at time t. m g,t is the hydrogen consumption of the g-th HFCV at time t. h is the energy conversion efficiency. HHV is the higher heating value of hydrogen.

[0236] Considering the discharge power limit and the hydrogen storage capacity limit of the hydrogen storage tank, the fuel cell operation model is as follows:

[0237]

[0238] Where, is the maximum discharge power of the fuel cell of the g-th HFCV. is the remaining hydrogen storage capacity of the g-th HFCV at time t. τ is the time step of the AGC signal. and They are the lower and upper limits of hydrogen storage capacity respectively.

[0239] Considering the charging and discharging power limit and energy limit, the energy storage operation model is as follows:

[0240]

[0241]

[0242] Where, and are the charging power and discharging power of the energy storage of the g-th HFCV at time t, respectively. is the maximum discharge power of the energy storage of the g-th HFCV. is the remaining energy stored in the g-th HFCV at time t. and They are the lower and upper limits of energy storage capacity respectively.

[0243] 2) The main steps of the hydrogen fuel cell vehicle frequency modulation capacity aggregation model based on Minkowski sum are as follows:

[0244] Due to the variability of hydrogen fuel cell vehicle grid connection time, it is necessary to standardize the dispatch to a unified time so that the aggregator can accurately determine the spare capacity of the frequency regulation business. g,t To characterize the vehicle grid connection status, as shown below

[0245]

[0246] Where, X g,t =1 means that the g-th vehicle is connected to the grid at time t.

[0247] The charging and discharging power limit and energy limit of the HFCV cluster can be expressed as

[0248]

[0249] Where, P t h,max is the maximum discharge power of all fuel cells in the HFCV cluster at time t. t b,max is the maximum charge and discharge power of all energy storage in the HFCV cluster at time t. and are the lower and upper limits of all hydrogen storage capacities within the HFCV cluster, respectively. and They are the lower and upper limits of all energy storage energies in the HFCV cluster, respectively.

[0250] In addition to the coupling relationship between power and energy, HFCV clusters can also provide SFR services for the power system. The frequency regulation capacity of HFCV represents the upward and downward regulation space based on the active power trajectory. The aggregation model of fuel cells in the HFCV cluster includes discharge power constraints, frequency regulation capacity constraints, and hydrogen storage capacity constraints, which can be expressed as

[0251]

[0252] Where, P t h M is the discharge power of the fuel cell in the HFCV cluster at time t. t is the hydrogen consumption of the HFCV cluster at time t. Fuel cells in HFCV clusters can provide frequency modulation capabilities. is the remaining hydrogen storage capacity of the HFCV cluster at time t. is the change in hydrogen storage energy caused by the change in the grid-connected state of the HFCV cluster at time t, which can be expressed as

[0253]

[0254] Where, e h,ag and e h,dg They are the hydrogen storage capacity of HFCV when it is connected to the grid and when it is off the grid, respectively.

[0255] The feasible domain of the frequency modulation capacity of fuel cells in an HFCV cluster is essentially a polyhedron constrained by a set of linear inequalities. According to Equations (11)-(16), the equation can be rearranged into the following compact form.

[0256]

[0257] Where, It represents the feasible domain of fuel cell frequency regulation capacity within the HFCV cluster. A represents the controllable variable of the fuel cell in the HFCV cluster. h and B h is the coefficient matrix.

[0258] The aggregation model of energy storage in the HFCV cluster includes charging and discharging power constraints, frequency regulation capacity constraints, and energy constraints, which can be expressed as

[0259]

[0260] Where, P t b,c and P t b,d are the charging and discharging powers of the energy storage in the HFCV cluster at time t, respectively. The frequency regulation capacity that energy storage can provide within the HFCV cluster. is the remaining energy stored in the HFCV cluster at time t. is the energy storage change caused by the change of the grid-connected state of the HFCV cluster at time t, which can be expressed as

[0261]

[0262] Where, e b,ag and e b,dg They are the energy storage capacity of HFCV when it is connected to the grid and when it is off the grid.

[0263] Similarly, according to equations (18)-(22), the feasible region of frequency regulation capacity of energy storage in the HFCV cluster can be expressed in the following compact form:

[0264]

[0265] Where, It represents the feasible domain of energy storage frequency regulation capacity within the HFCV cluster. A represents the controllable variable of energy storage in the HFCV cluster. b and B b is the coefficient matrix.

[0266] The feasible region of the frequency modulation capacity of the HFCV cluster is composed of fuel cells and batteries. In essence, it is the Minkowski sum of the feasible regions of the frequency modulation capacity of fuel cells and batteries, which can be expressed as:

[0267]

[0268] Where, represents the Minkowski sum. U represents a point on the boundary of the feasible region of the set.

[0269] To reduce computational complexity, a method for accurately calculating the Minkowskisum of the frequency modulation capacity of an HFCV cluster is proposed. See Algorithm 1 for details. The core idea of ​​this method is to calculate the sum of the vertices of the feasible region for the frequency modulation capacity of the fuel cell and battery to obtain the set of Minkowskisum boundaries for the HFCV cluster. The convex hull of this set is then constructed to determine the exact feasible region for the frequency modulation capacity.

[0270]

[0271] 3) The main steps of the online secondary frequency modulation algorithm for hydrogen fuel cell vehicles are as follows:

[0272] Lyapunov optimization can solve online optimization problems with time-average constraints. Therefore, in order to respond to real-time AGC signals, the time-coupling constraints (3) and (7) must be converted into time-average constraints. The following is the derivation process of the time-average constraint. First, the sum of (3) and (7) is as follows:

[0273]

[0274] Calculate the expected values ​​of (26) and (27) and then divide them by T as follows:

[0275]

[0276]

[0277] When T tends to infinity,

[0278]

[0279] because and is a finite value, the right side of equations (30) and (31) is zero. Therefore, the above equation can be expressed as

[0280]

[0281] Constraints (32) and (33) are time-averaged constraints, which are relaxations of constraints (3)-(4) and (7)-(8), respectively.

[0282] The goal of the HFCV aggregator is to reduce the life cycle cost of HFCVs while efficiently allocating SFR power to each HFCV. The objective function includes the HFCV fuel cell lifetime degradation cost and battery lifetime degradation cost, as well as the penalty cost for frequency regulation mismatch.

[0283] Fuel cell operating cost: During the V2G response period, the fuel cell operating cost includes the hydrogen consumption cost and the fuel cell degradation cost. Without considering the fuel cell degradation caused by traffic-related factors, the fuel cell degradation of SFR can be expressed as

[0284]

[0285] Where, is the degradation cost of the fuel cell at time t. h is the fuel cell degradation coefficient caused by V2G, which can be expressed as

[0286]

[0287] Where d1 represents the degradation rate of the fuel cell under maximum power conditions.

[0288] In the V2G process, the hydrogen consumption cost of the fuel cell can be expressed as

[0289]

[0290] Where, is the hydrogen consumption cost of HFCV g at time t. is the hydrogen cost coefficient.

[0291] Energy storage degradation cost: According to the energy storage degradation mechanism, the degree of energy storage life degradation is affected by the depth of discharge cycle (DoD). To prevent overcharge and overdischarge of energy storage, when the battery operating range is limited to a specific DoD range, the energy storage life degradation cost increases monotonically with the increase of cycle depth. The cycle DoD shows a linear relationship with the number of battery charge and discharge times, as described below.

[0292]

[0293] In the SFR process, the energy storage charge and discharge degradation cost is proportional to the cycle DoD and can be expressed as

[0294]

[0295] Where, represents the energy storage degradation cost of HFCV g at time t. is the battery degradation cost coefficient, which can be expressed as

[0296]

[0297] Where λ g N represents the battery investment cost of HFCV g. b Energy storage under energy constraints The number of cycles that can be performed.

[0298] Frequency regulation mismatch penalty cost: During the frequency regulation response, the aggregator is penalized for the regulation mismatch due to the absolute error between the AGC signal and the actual response power. The frequency regulation mismatch penalty cost for the HFCV aggregator can be expressed as

[0299]

[0300] Where r t is the normalized AGC signal, r t ∈[-1,1]. R t is the total frequency regulation capacity determined by the HFCV aggregator. p is the penalty cost coefficient.

[0301] Based on the time average constraint and objective function, the time average optimization model of HFCV participating in SFR is expressed as:

[0302]

[0303] st(1)-(2),(5)-(6),(33)-(34) (44)

[0304]

[0305] Where, constraints (45) and (46) prohibit HFCV g from charging and discharging simultaneously. Constraints (47) and (48) determine the charge and discharge states of HFCV g. Constraints (49) and (50) specify that the frequency response power of the aggregator must not exceed its total frequency regulation capability. Based on the above time-averaged optimization model, we use the Lyapunov optimization method to solve the online SFR problem.

[0306] Based on the Lyapunov optimization method, the key steps for the online response of HFCVs to AGC signals are outlined as follows: 1) Construct a virtual queue and transform the time-averaged constraint into a queue stability constraint. 2) Define the Lyapunov function to obtain the Lyapunov drift and the drift-plus-penalty. 3) Minimize the upper bound of the drift-plus-penalty to obtain the solution to the online SFR problem.

[0307] The virtual hydrogen fleet of fuel cells is defined as follows:

[0308]

[0309] Where, is a perturbation parameter that ensures that the hydrogen stored in the HFCV remains within the limits of constraint (4).

[0310] The dynamics of the virtual hydrogen fleet can be expressed as:

[0311]

[0312] According to formula (32), the virtual hydrogen energy queue H of HFCV can be derived: g,t With average rate stability, as shown below:

[0313]

[0314] Similar to the fuel cell virtual queue, the virtual energy queue for energy storage is defined as:

[0315]

[0316] Where, is a perturbation parameter that ensures that the energy stored in the HFCV is maintained within the bounds of constraint (8).

[0317] The dynamic characteristics of the virtual energy queue can be expressed as

[0318]

[0319] According to formula (33), the virtual energy queue B of HFCV g,t With average rate stability, as shown below:

[0320]

[0321] Equations (53) and (56) represent the average rate stability conditions for the above virtual hydrogen and battery energy queues, indicating that the virtual queues will not grow linearly over time.

[0322] The concatenated vector Θ of the virtual queue t =(H t ,B t ) is defined as

[0323]

[0324] In the sizes of virtual hydrogen and battery energy queues, the Lyapunov function is defined as follows

[0325]

[0326] In order to alleviate the congestion of virtual hydrogen and energy storage queues, it is crucial to maintain a small Lyapunov function value. The change of the Lyapunov function in two consecutive cycles should not be too large. Therefore, the Lyapunov drift Δ(Θ t ) is defined as:

[0327]

[0328] Where, Θ t is a random variable.

[0329] Lyapunov drift is a metric used to quantify the magnitude of the queue Θ given its current state. t The expected growth of the size. By minimizing the Lyapunov drift, the stability of the virtual queue can be maintained. However, simply minimizing the Lyapunov drift often leads to higher frequency adjustment costs. Therefore, we adopt a drift plus penalty strategy to balance the Lyapunov drift Δ(Θ t ) and operating costs F t The formula for drift plus penalty is as follows:

[0330]

[0331] Where V is the weight coefficient, which is used to balance the stability and operating cost of the virtual queue.

[0332] Due to the definition of Lyapunov drift, the drift plus penalty term is still closely related to time. In order to achieve online frequency regulation, the control decision of HFCV is obtained by minimizing the upper limit of Lyapunov drift plus penalty function. The Lyapunov drift in the AGC signal gap can be expressed as

[0333]

[0334] After substituting equations (63) and (64) into equation (61), the drift plus penalty can be reformulated as

[0335]

[0336] Where, is the maximum hydrogen consumption of the g-th HFCV.

[0337] By minimizing the upper bound of the sum of drift and penalty, the online SFR problem for each AGC signal time slot can be formulated as:

[0338]

[0339] Before the start of each AGC signal time slot, the parameters are updated according to equations (52) and (55). Then, given the current virtual queue, total frequency regulation capacity, and AGC signal, the online SFR problem P2 is optimized to determine the frequency regulation output of each HFCV without predicting the future AGC signal. Since A is a constant in the optimization problem, it is omitted from the optimization model.

[0340] By selecting the weight coefficient V, the perturbation parameters can be determined to ensure that the hydrogen and battery energies remain within the constraints (4) and (8), as described in the following proposition.

[0341] Proposition 1: For any V∈[0,V max ], when the disturbance parameter is equal to

[0342]

[0343] The online SFR problem P2 can obtain the sequence of hydrogen and battery energy that satisfies constraints (4) and (8), where

[0344]

[0345] To analyze the gap between offline and online algorithms, we first determine the offline SFR algorithm as follows:

[0346]

[0347] st( 1 )-( 8 ),( 45 )-( 5 0)(70)

[0348] The suboptimality of the online SFR algorithm refers to the gap between the optimal solution of the online problem P2 and the offline problem P3, as stated in Proposition 2.

[0349] Proposition 2: The time average cost of the online problem P2 and the offline problem P3 are expressed as and The gap between the optimal solutions of the online optimization problem and the offline optimization problem is

[0350]

[0351] From the above proposition, we can see that the gap between the optimal solutions of the online problem P2 and the offline problem P3 is inversely proportional to the weight coefficient V. An increase in the value of V will lead to a decrease in the optimality gap.

[0352] Example 12:

[0353] The verification of an online secondary frequency regulation method considering the aggregation of hydrogen fuel cell vehicle regulation capacity is as follows:

[0354] To verify the correctness of the proposed online secondary frequency regulation algorithm that considers the aggregation of hydrogen fuel cell vehicle regulation capacity, a simulation analysis was conducted using 1,000 hydrogen fuel cell vehicles participating in the secondary frequency regulation of the power system. The relevant parameters of the simulation model are shown in Table 1.

[0355] TABLE 1 Simulation parameters

[0356]

[0357] For the convenience of calculation, this paper assumes that hydrogen vehicles only participate in V2G at home. Taking 18:00 as an example, the feasible domain of the aggregate frequency regulation capacity of all hydrogen fuel cell vehicles is as follows: Figure 2 shown.

[0358] Based on Algorithm 1, the feasible domain of the frequency regulation capacity of the energy storage and fuel cell of hydrogen fuel cell vehicles is first determined. The feasible domain of the energy storage frequency regulation capacity for all hydrogen fuel cell vehicles connected to the grid at 18:00 is determined by the combined effects of the energy storage charge and discharge power constraints, the frequency regulation capacity constraints, and the energy storage SOC constraints. As the energy storage charge and discharge power increases, the maximum frequency regulation capacity provided by the energy storage gradually decreases. The feasible domain of the fuel cell frequency regulation capacity for all hydrogen fuel cell vehicles connected to the grid at 18:00 is determined by the combined effects of the fuel cell discharge power constraints, the frequency regulation capacity constraints, and the hydrogen tank hydrogen storage capacity constraints. Unlike energy storage, fuel cells participate in frequency regulation only through discharge. As the fuel cell discharge power increases, the frequency regulation capacity provided by the fuel cell gradually increases.

[0359] Then, the Minkowski sum of the hydrogen fuel cell vehicle's frequency modulation capacity was accurately calculated to determine the vertices of the feasible region for the aggregate frequency modulation capacity. The vertices of the feasible region for the frequency modulation capacity were connected to determine the precise feasible region for the aggregate frequency modulation capacity of the hydrogen fuel cell vehicle. As the discharge power of the hydrogen fuel cell vehicle increases, the maximum frequency modulation capacity it can provide increases first and then decreases. At a discharge power of 2625.98 kW, the frequency modulation capacity provided by the hydrogen fuel cell vehicle reaches its maximum, reaching 19005.98 kW. Within the range of [-1452.45, 0], as the charging power of the hydrogen fuel cell vehicle increases, the maximum frequency modulation capacity it can provide gradually decreases.

[0360] In order to highlight the advantages of the convex hull-based frequency modulation capacity aggregation method proposed in the present invention, Table 2 compares the aggregation results of different aggregation methods. Compared with the zonotope method and the box method, the convex hull-based aggregation method proposed in the present invention can give full play to the flexibility of hydrogen fuel cell vehicles and determine the maximum feasible domain of the aggregated frequency modulation capacity of hydrogen fuel cell vehicles. The box-based aggregation method cannot construct a coupling model between the frequency modulation capacity and the charge and discharge power of hydrogen fuel cell vehicles. Therefore, this method sacrifices the most flexibility and has the smallest feasible domain of aggregated frequency modulation capacity. The zonotope-based aggregation method can only construct a symmetric feasible domain, sacrificing the asymmetric feasible domain. The feasible domain of aggregated frequency modulation capacity is smaller than that of the method proposed in the present invention, and some flexibility of frequency modulation of hydrogen fuel cell vehicles is lost.

[0361] TABLE 2 Aggregation results of different aggregation methods

[0362]

[0363] Four scenarios are used to compare and analyze the online frequency modulation power allocation method. The four scenarios are as follows:

[0364] Scenario 1: Online optimization of secondary frequency modulation power allocation using a greedy algorithm;

[0365] Scenario 2: Optimizing secondary frequency modulation power allocation using an offline method;

[0366] Scenario 3: Adopting the online optimization of secondary frequency modulation power allocation proposed by the present invention.

[0367] Taking the frequency regulation capacity from 18:00 to 18:15 as a benchmark, Table 3 shows the accumulated operating costs over time under different methods. Greedy algorithm S1 has the worst performance and the highest cost. Offline optimization algorithm S2 has the lowest cost and the best performance. Since the AGC signal is difficult to predict, this algorithm is not feasible in practice. Algorithm S3 proposed in the present invention has the best performance, lower operating cost than greedy algorithm S1, and does not require AGC signal prediction. Compared with algorithms S1 and S2, the algorithm proposed in the present invention significantly reduces the total cost of secondary frequency regulation from $33,690.8 to $24,451.36, a large decrease of 27.42%.

[0368] TABLE 3 Comparison of secondary frequency regulation operating costs in different scenarios ($)

[0369]

[0370] Figure 3 The total frequency modulation power output of an HFCV cluster using different methods was compared. For greedy algorithm S1, the HFCV cluster fully responded to the AGC signal through charging and discharging within 0-210 seconds. For offline algorithm S2, the HFCV cluster fully responded to the AGC signal, ensuring minimal operating costs throughout the entire AGC signal response period. However, due to the difficulty in predicting future AGC signals, the offline algorithm is not practical. Compared to greedy algorithm S1, online algorithm S3 achieved a longer response time. Figure 4 The energy variation of the HFCV in response to the AGC signal is shown. For the online algorithm S3, the battery energy and hydrogen level are always kept within a limited range, satisfying the constraints (4) and (8), verifying the correctness of Proposition 1.

[0371] To evaluate the effectiveness of the proposed online algorithm, 30 different AGC signal sequences are randomly selected to calculate the gap between the online and offline algorithms. The gaps between different AGC signal sequences are shown in Figure 2. Figure 5 As shown in Figure 2, the maximum difference between the two algorithms reaches 24.81% of the best result achieved by the offline algorithm, while the minimum difference is exactly 0. In particular, Figure 5 All gaps in are strictly smaller than the maximum gap 77.42$ shown in Equation (71), verifying the correctness of Proposition 2.

Claims

1. An online secondary frequency modulation method considering the aggregation of hydrogen fuel cell vehicle regulation capacity, characterized in that: The following steps are involved: 1) Aggregators record the number of hydrogen fuel cell vehicle clusters connected to decentralized charging and discharging piles, as well as the time these hydrogen fuel cell vehicles reported leaving the vehicle; 2) Based on the hydrogen fuel cell vehicle cluster, build a power system operation model for hydrogen fuel cell vehicles (HFCVs); 3) Construct a frequency regulation capacity aggregation model for hydrogen fuel cell vehicles based on Minkowski sum; 4) Solve the frequency regulation capacity aggregation model of hydrogen fuel cell vehicles and obtain the feasible region of frequency regulation capacity of HFCV clusters; 5) Based on the power system operation model, build an online secondary frequency modulation model for hydrogen fuel cell vehicles; 6) Based on the feasible region of frequency regulation capacity, the system dispatching agency uses the Lyapunov optimization method to solve the online secondary frequency regulation model of hydrogen fuel cell vehicles, obtains the AGC command signal, and sends it to the aggregator; 7) The aggregator distributes AGC instructions to each hydrogen fuel cell vehicle in real time and participates in the secondary frequency regulation of the power system.

2. The online secondary frequency modulation algorithm considering the aggregation of hydrogen fuel cell vehicle regulation capacity according to claim 1 is characterized in that: The power system operation model of the hydrogen fuel cell vehicle HFCV includes the hydrogen fuel cell operation model and the electrochemical energy storage operation model.

3. The online secondary frequency modulation algorithm considering the aggregation of hydrogen fuel cell vehicle regulation capacity according to claim 2 is characterized in that: Considering the discharge power limit and the hydrogen storage capacity limit of the hydrogen storage tank, the hydrogen fuel cell operation model is as follows: Where, is the maximum discharge power of the fuel cell of the g-th HFCV; is the remaining hydrogen storage capacity of the g-th HFCV at time t; τ is the time step of the AGC signal. and are the lower and upper limits of hydrogen storage capacity, respectively; P t h,max is the maximum discharge power of all fuel cells in the HFCV cluster at time t; is the start and end time; Among them, the hydrogen consumption of hydrogen fuel cells during V2G is as follows: Where g is the set of hydrogen fuel cell vehicles The index of t is the time set The index of is the discharge power of the fuel cell of the g-th HFCV at time t; m g,t is the hydrogen consumption of the g-th HFCV at time t; η h is the energy conversion efficiency; HHV is the higher heating value coefficient of hydrogen.

4. The online secondary frequency modulation algorithm considering the aggregation of hydrogen fuel cell vehicle regulation capacity according to claim 2 is characterized in that: Considering the charging and discharging power limitations and energy limitations, the electrochemical energy storage operation model is as follows: Where, and are the charging power and discharging power of the energy storage of the g-th HFCV at time t, respectively. is the maximum discharge power of the energy storage of the g-th HFCV. is the remaining energy stored in the g-th HFCV at time t. and are the lower and upper limits of energy storage capacity, respectively; P t b,max is the maximum charge and discharge power of all energy storage in the HFCV cluster at time t; η c and η d It is the charge and discharge efficiency.

5. The online secondary frequency modulation algorithm considering the aggregation of hydrogen fuel cell vehicle regulation capacity according to claim 1 is characterized in that: The Minkowski sum-based hydrogen fuel cell vehicle frequency regulation capacity aggregation model includes the fuel cell aggregation model within the HFCV cluster and the energy storage aggregation model; The fuel cell aggregation model includes discharge power constraints, frequency modulation capacity constraints, and hydrogen storage capacity constraints; The energy storage aggregation model includes charging and discharging power constraints, frequency regulation capacity constraints and energy constraints.

6. The online secondary frequency modulation algorithm considering the aggregation of hydrogen fuel cell vehicle regulation capacity according to claim 1 is characterized in that: The fuel cell aggregation model is shown below: Where, P t h M is the discharge power of the fuel cell in the HFCV cluster at time t. t is the hydrogen consumption of the HFCV cluster at time t; The frequency modulation capacity that can be provided by fuel cells in the HFCV cluster; is the remaining hydrogen storage capacity of the HFCV cluster at time t; Among them, the change in hydrogen storage energy caused by the change in the grid connection state of the HFCV cluster at time t The state variable X that represents the vehicle grid connection state g,t As shown below: Where, and are the hydrogen storage capacity of HFCV when it is connected to the grid and when it is off the grid; X g,t =1 means the g-th vehicle is connected to the grid at time t; The energy storage aggregation model is as follows: Where, P t b,c and P t b,d are the charging and discharging powers of the energy storage in the HFCV cluster at time t, respectively; The frequency regulation capacity that energy storage can provide within the HFCV cluster. is the remaining energy stored in the HFCV cluster at time t; Among them, the energy storage energy change caused by the change of the grid connection state of the HFCV cluster at time t As shown below: Where, and They are the energy storage capacity of HFCV when it is connected to the grid and when it is off the grid.

7. The online secondary frequency modulation algorithm considering the aggregation of hydrogen fuel cell vehicle regulation capacity according to claim 1 is characterized in that: The feasible domain of the frequency modulation capacity of the HFCV cluster is as follows: Where, It represents the feasible domain of fuel cell frequency regulation capacity within the HFCV cluster. A represents the controllable variable of the fuel cell in the HFCV cluster. h and B h is the coefficient matrix; It represents the feasible domain of energy storage frequency regulation capacity within the HFCV cluster. A represents the controllable variable of energy storage in the HFCV cluster. b and B b is the coefficient matrix; represents the Minkowski sum. U represents a point on the boundary of the feasible region of the set.

8. The online secondary frequency modulation algorithm considering the aggregation of hydrogen fuel cell vehicle regulation capacity according to claim 1 is characterized in that: The steps to solve the frequency regulation capacity aggregation model of hydrogen fuel cell vehicles include: 3.1) Integrate HFCV cluster grid connection information They represent grid connection time, off-grid time, grid-connected energy storage, grid-connected hydrogen storage, and off-grid hydrogen storage respectively; 3.2) Based on the hydrogen fuel cell vehicle frequency modulation capacity aggregation model, the feasible domain of the frequency modulation capacity of fuel cells and batteries in the HFCV cluster is calculated separately; 3.3) According to the set feasible region boundary formula U=U b +U h , determine the Minkowski sum boundary vertex O of the feasible region of HFCV cluster frequency modulation capacity j ; 3.4) Construct the sorting point set Ω={O1,O2,...,O J }; Among them, the minimum point O1 in the sorted point set is selected as the reference point; 3.5) If The boundary vertex O j Write the Minkowski sum vertex set of the frequency modulation capacity feasible domain middle; 3.6) Determine whether j>J holds. If not, set j=j+1 and return to step 3.5). If so, proceed to step 3.7); 3.7) Minkowski sum vertex set of the feasible region of the connection frequency modulation capacity Each point determines the exact feasible region of frequency regulation capability 9. The online secondary frequency modulation algorithm considering the aggregation of hydrogen fuel cell vehicle regulation capacity according to claim 1 is characterized in that: The steps to build an online secondary frequency modulation model for hydrogen fuel cell vehicles include: 5.1) Convert the time coupling constraints (2) and (7) into time average constraints, that is: Where E is the expectation; parameter 5.2) To reduce the life cycle cost of HFCVs and allocate SFR power to each HFCV, an online secondary frequency regulation model for hydrogen fuel cell vehicles is constructed, namely: st(1),(4),(5)-(6),(25)-(26)(29) Where, constraints (30) and (31) prohibit HFCV g from being charged and discharged simultaneously. Constraints (32) and (33) determine the charge and discharge state of HFCV g; r t is the normalized AGC signal; Among them, the degradation cost of fuel cells is as follows: Where, is the degradation cost of the fuel cell at time t; is the fuel cell cost coefficient; Fuel cell degradation coefficient d caused by V2G h As shown below: Where d1 represents the degradation rate of the fuel cell under maximum power conditions; In the V2G process, the hydrogen consumption cost of the fuel cell is as follows: Where, is the hydrogen consumption cost of HFCV g at time t; is the hydrogen cost coefficient; In the SFR process, the energy storage charging and discharging degradation cost is as follows: Where, represents the energy storage degradation cost of HFCV g at time t; Battery degradation cost coefficient As shown below: Where λ g N represents the battery investment cost of HFCV g. b Energy storage under energy constraints the number of cycles that can be performed; Frequency Regulation Mismatch Penalty Cost As shown below: Where r t is the normalized AGC signal, r t ∈[-1,1]. R t is the total frequency regulation capacity determined by the HFCV aggregator. p is the penalty cost coefficient.

10. The online secondary frequency modulation algorithm considering the aggregation of hydrogen fuel cell vehicle regulation capacity according to claim 1 is characterized in that: The steps of solving the online secondary frequency modulation model of hydrogen fuel cell vehicles using the Lyapunov optimization method include: 6.1) Construct a virtual queue and convert the time average constraint into a queue stability constraint. The steps include: 6.1.1) Construct a virtual hydrogen energy queue for fuel cells and an expression for the dynamic characteristics of the virtual hydrogen energy queue, namely: Where, is the disturbance parameter; 6.1.2) Constructing a virtual hydrogen fleet H for HFCV g,t The average rate stability constraint is as follows: 6.1.3) Construct a virtual energy queue for energy storage and an expression for the dynamic characteristics of the virtual energy queue, namely: Where, is the disturbance parameter; 6.1.4) Constructing Virtual Energy Queue B g,t The average rate stability constraint is: 6.1.5) Combine steps 6.1.1)-6.1.4) to construct a concatenated vector Θ of the virtual queue t =(H t ,B t ),Right now 6.2) Define the Lyapunov function to obtain Lyapunov drift and drift plus penalty; in, In the size of the virtual hydrogen and battery energy queues, the Lyapunov function L(Θ t ) is as follows: Lyapunov drift Δ(Θ t ) is as follows: Where, Θ t is a random variable. The drift penalty is as follows: Where V is the weight coefficient, which is used to balance the stability and operating cost of the virtual queue; F t For operating costs; 6.3) Construct the Lyapunov drift expression within the AGC signal gap, namely: 6.4) Substituting formulas (55)-(56) into formula (53), minimizing the upper limit of drift plus penalty, we obtain Where, is the maximum hydrogen consumption of the g-th HFCV; 6.5) Based on formula (57), the online SFR problem of each AGC signal time slot is constructed, that is: 6.6) Before the start of each AGC signal time slot, update the virtual queue using equations (44) and (47); 6.7) Based on the given current virtual queue, total frequency regulation capacity and AGC signal, solve the online SFR problem of each AGC signal time slot and determine the AGC command signal for each HFCV.

Images