Optimal configuration method of flywheel energy storage and medium voltage energy feedback for urban rail traction power supply

Abstract

The application provides a kind of urban rail transit traction power supply flywheel energy storage and medium voltage energy feedback collaborative optimization configuration method, flywheel energy storage unit FESE and inverter energy feedback unit IEFE are introduced in urban rail transit power supply system simultaneously, and urban rail transit AC-DC mixed power flow analysis model is constructed;With daily total cost and static recovery cycle as target, build comprehensive cost index;The number of FESE and the number of IEFE are taken as discrete variables, the charging voltage threshold of FESE, the discharging voltage threshold and the action voltage threshold of inverter energy feedback unit are taken as continuous variables, a hierarchical swarm intelligence optimization method is used, and the minimum comprehensive cost index is taken as the target, under the premise of meeting the operation constraints of power supply system, the collaborative optimization configuration scheme containing the number of FESE, the number of IEFE and of each traction substation is output.

Description

Technical Field

[0001] This invention relates to the field of urban rail transit traction power supply technology, specifically to a method for the coordinated optimization configuration of flywheel energy storage and medium-voltage energy feeder in urban rail transit traction power supply. Background Technology

[0002] Regenerative braking energy generated during frequent starts and stops of urban rail transit trains accounts for 30%-55% of the total traction energy consumption. Currently, regenerative braking energy recovery mainly relies on two technical routes: flywheel energy storage units (FESE) and medium-voltage energy feeder units (IEFE).

[0003] Flywheel energy storage units have efficient dynamic buffering capabilities, can store regenerative braking energy in real time and use it directly for train traction, resulting in significant energy savings, but the equipment cost is high. Medium-voltage energy feed-in equipment is technically mature and flexible in capacity expansion, and can quickly feed electrical energy back to the medium-voltage AC grid, but its energy utilization depends on the load demand on the grid side. If there is no vehicle traction demand or the power and lighting load cannot be fully absorbed by the adjacent traction substation during energy feedback, it may cause reverse power feeding or energy circulation to the urban grid, resulting in energy loss.

[0004] Existing research largely focuses on improving the performance and optimizing individual types of devices, lacking in-depth analysis of the collaborative mechanisms between flywheel energy storage and medium-voltage energy feedforward at the capacity matching and energy flow interaction levels. Especially in mixed-variable optimization problems involving discrete device configuration parameters such as the number of devices and continuous voltage threshold parameters, traditional algorithms suffer from bottlenecks such as the curse of dimensionality, variable coupling, and inefficient resource allocation, making it difficult to achieve a comprehensive optimization of system-level energy efficiency and economy. Therefore, there is an urgent need for a collaborative optimization configuration method that balances collaborative mechanisms and optimization efficiency between flywheel energy storage and medium-voltage energy feedforward. Summary of the Invention

[0005] This invention proposes a method for the coordinated optimization configuration of flywheel energy storage and medium-voltage energy feeder in urban rail transit traction power supply, in order to solve the technical problem that existing technologies only consider flywheel energy storage unit FESE and medium-voltage energy feeder equipment IEFE, which makes it difficult to achieve the comprehensive optimization of system-level energy efficiency and economy.

[0006] To address the aforementioned technical problems, this invention provides a method for the coordinated and optimized configuration of flywheel energy storage and medium-voltage energy feeder in urban rail transit traction power supply, comprising the following steps:

[0007] Step S1: Simultaneously introduce flywheel energy storage unit FESE and inverter energy feeder unit IEFE into the urban rail transit power supply system, and construct an AC / DC hybrid power flow analysis model for urban rail transit.

[0008] Step S2: Construct a comprehensive cost index with daily total cost and static payback period as targets;

[0009] Step S3: Treat the number of FESEs and the number of IEEFEs as discrete variables, and set the charging voltage threshold for FESEs. Discharge voltage threshold and the operating voltage threshold of the inverter energy feeder unit As a continuous variable, a hierarchical swarm intelligence optimization method is adopted, aiming to minimize the comprehensive cost index. Under the premise of satisfying the power supply system operation constraints, the output includes the number of FESEs and IEFEs of each traction substation. , , A collaborative optimization configuration scheme.

[0010] Preferably, the AC / DC hybrid power flow analysis model for urban rail transit in step S1 includes: equating the 24-pulse rectifier unit in the traction power supply system to a piecewise linearized external characteristic model and dividing it into a constant voltage regulation zone and an overload droop zone; equating the train to a constant power load model; modeling the flywheel energy storage unit FESE and the inverter energy feed unit IEFE as multi-mode switching models based on DC grid voltage; and using an alternating AC / DC solution method to complete the energy flow calculation and convergence determination.

[0011] Preferably, the modeling expression for the 24-pulse rectifier unit in the AC / DC hybrid power flow analysis model for urban rail transit is as follows:

[0012] ;

[0013] In the formula, This indicates the DC side voltage of the rectifier unit; This indicates the no-load voltage of the rectifier unit; Indicates the drooping slope; Indicates the DC-side output current; This indicates the rated current of the rectifier unit; This indicates the rated DC output voltage.

[0014] Preferably, the modeling expression for trains in the AC / DC hybrid power flow analysis model for urban rail transit is:

[0015] ;

[0016] In the formula, The voltage of the traction network at the pantograph; Indicates the train's current; Indicates the train's mechanical power; This indicates the electromechanical conversion efficiency.

[0017] Preferably, in step S2, the total daily cost includes operating electricity costs, equipment configuration costs, and equipment operation and maintenance costs;

[0018] The operating electricity cost The expression is:

[0019] ;

[0020] The equipment configuration cost The expression is:

[0021] ;

[0022] The equipment operation and maintenance cost The expression is:

[0023] ;

[0024] In the formula, Indicates electricity price; This indicates the number of all main substations in the power supply system; This indicates the electrical energy consumed by each switch. This indicates the total number of traction substations within the system; and Let represent the nonlinear cost functions of FESE and IEFE, respectively; and They represent the first The number of FESEs and IEFEs configured in each substation; and These represent the full lifecycle of FESE and IEFE, respectively. This represents the equipment operation and maintenance cost coefficient.

[0025] Preferably, in step S2, the static recycling cycle The expression is:

[0026] ;

[0027] In the formula, This represents the total energy expenditure of the base value scheme.

[0028] Preferably, step S3 includes:

[0029] Step S31: Based on prior knowledge of the intensity of site load fluctuations, the particle swarm is divided into several subgroups by setting the number of groups, and a constraint on the total number of devices is set for each subgroup.

[0030] Step S32: Initialize the particle swarm according to the total number of devices in each subgroup, generate a discrete solution containing the number of FESEs and IEFEs of each traction substation, and mark the leading particle and the following particle in each group;

[0031] Step S33: Perform discrete decision-making layer optimization: fix continuous variables and perform iterative optimization of discrete variables based on the comprehensive cost index;

[0032] Step S34: Perform continuous decision-making layer optimization: Based on the equipment layout output by the discrete decision-making layer, fix the number of FESEs and IEFEs of the traction substation, and iteratively optimize the continuous variables based on the comprehensive cost index;

[0033] Step S35: When any layer satisfies the early stop condition or reaches the iteration limit, output the optimal result, and combine the output of the discrete decision layer with the output of the continuous decision layer to obtain the number of FESEs and IEEFEs for each traction substation. , , A collaborative optimization configuration scheme.

[0034] Preferably, in steps S33 and S34, the leading particle and the following particle are updated differently.

[0035] The guiding particle is updated using the following expression:

[0036] ;

[0037] The following particles are updated using the following expression:

[0038] ;

[0039] In the formula, Indicates the first Particles in the next iteration In dimensions The speed on; Indicates the first Inertia weights in the next iteration; and Indicates the first The learning factor at the next iteration; and Represents a random number within a unit interval; Represents particles In dimensions The optimal position of an individual on the [surface / structure]. Represented as dimension The global optimal position; Representing dimensions The optimal position within the group.

[0040] Preferably, in steps S33 and S34, the inertia weight and learning factor are updated using the following expressions:

[0041] ;

[0042] ;

[0043] ;

[0044] In the formula, Indicates the first The inertia weight at the next iteration; T represents the maximum number of iterations for the corresponding layer; and These represent the maximum and minimum weights of the inertia weight, respectively. Represents learning factor The maximum value; and Represents learning factor The maximum and minimum values.

[0045] Preferably, during discrete decision-making layer optimization, after every Q iterations, the number of particles in each group is recalculated and redistributed according to the following formula:

[0046] ;

[0047] ;

[0048] ;

[0049] ;

[0050] ;

[0051] In the formula, Let the weights be the assigned weights for the g-th group; This represents the optimized baseline scene fitness value when FESE and IEFE are not configured. This represents the fitness value corresponding to the current optimal solution in the g-th group; This represents the initial number of basic particles allocated to the g-th group; This represents the total number of particles in the particle swarm. Indicates the number of groups; This represents the remaining particles after the initial allocation; This represents a vector of group indexes sorted in descending order of weight. The function returns the index values ​​in ascending order. Represents the weight sequence vectors of each group; This represents the index of the group with the highest weight after sorting. This represents the final number of particles allocated to the optimal group.

[0052] The beneficial effects of this invention include at least the following: the synergistic optimization mechanism of flywheel energy storage and inverter energy feedback enables the two to complement each other's functions, avoiding the limitations of single equipment in energy recovery and economy. It improves the utilization efficiency of regenerative braking energy through a composite energy path of energy storage buffering and grid feedback, while balancing equipment investment and operational benefits. This innovative synergistic optimization framework and algorithm provides a feasible path for system-level optimization of regenerative braking energy recovery in urban rail transit power supply systems, taking into account both energy utilization efficiency and economy, and contributing to the efficient and low-carbon operation of the power supply system. Attached Figure Description

[0053] Figure 1 This is a schematic diagram of the method flow according to an embodiment of the present invention;

[0054] Figure 2 This is a schematic diagram of an urban rail transit power supply system according to an embodiment of the present invention;

[0055] Figure 3 This is a schematic diagram of the power flow calculation simulation model according to an embodiment of the present invention;

[0056] Figure 4 This is a schematic diagram of the economic model of flywheel energy storage and inverter energy feed-in according to an embodiment of the present invention;

[0057] Figure 5 This is a schematic diagram of the solution process in an embodiment of the present invention. Detailed Implementation

[0058] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of the present invention.

[0059] like Figure 1 As shown in the figure, this invention provides a method for the coordinated optimization configuration of flywheel energy storage and medium-voltage energy feeder in an urban rail transit power supply system, including the following steps:

[0060] Step S1: Simultaneously introduce flywheel energy storage unit FESE and inverter energy feeder unit IEFE into the urban rail transit power supply system, and construct an AC / DC hybrid power flow analysis model for urban rail transit.

[0061] Specifically, such as Figure 2As shown, the urban rail transit power supply system exhibits typical AC / DC hybrid power supply characteristics, mainly composed of an AC power supply network and a DC traction network. The AC power supply network adopts a 35kV dual-power zoned power supply structure, consisting of a main substation and a traction step-down hybrid substation. The main substation is responsible for drawing medium-voltage AC power from the urban power grid and distributing it to each traction step-down hybrid substation. Each station's traction step-down hybrid substation is a standardized integrated unit, internally containing a 24-pulse rectifier unit, an inverter energy feeder unit, a flywheel energy storage unit, and station power and lighting loads. The rectifier unit converts 35kV AC power into DC traction network voltage through a rectifier transformer, supplying power to the DC traction network; the step-down transformer reduces the medium-voltage AC power to low-voltage AC, meeting the power and lighting load requirements of station lighting, elevators, air conditioning, etc. The DC traction network, as the core of energy transmission, connects each traction step-down hybrid substation to the train, undertaking the transmission function of train traction and regenerative braking energy. Therefore, the urban rail transit power supply system exhibits significant distributed and AC / DC coordinated characteristics.

[0062] like Figure 3 As shown, the rectifier unit can be equivalent to a voltage source with unidirectional current flow, wherein... This is the no-load voltage of the rectifier unit. For output current, The equivalent resistance of the rectifier unit; The operating threshold voltage of the medium-voltage energy feed system. The equivalent resistance of a medium-voltage energy feed system; flywheel energy storage is equivalent to two unidirectional voltage sources. This is the charging threshold voltage. For charging resistors, This is the discharge threshold voltage. The resistor is the discharge resistor; the train can be considered an equivalent current source. For train current, The braking resistor's operating threshold. To assist in energy consumption equivalent resistance; uplink / downlink variable resistor and These are used to characterize the dynamic resistance value between the train and the station due to the distance difference, and their magnitude depends on the distance between the up / down train and the substation.

[0063] First, the 24-pulse rectifier unit in the traction power supply system is modeled, and its expression is:

[0064] ;

[0065] In the formula, This indicates the DC side voltage of the rectifier unit; This indicates the no-load voltage of the rectifier unit; The droop slope represents the regulation characteristic of the load current on the output voltage. Indicates the DC-side output current; This indicates the rated current of the rectifier unit; This indicates the rated DC output voltage.

[0066] Secondly, the constant power load of the train is modeled, and its expression is:

[0067] ;

[0068] In the formula, The voltage of the traction network at the pantograph; Indicates the train current. >0 indicates traction flow extraction. <0 indicates regenerative energy supply; Indicates the train's mechanical power. >0 indicates traction power. <0 indicates braking power; This indicates the electromechanical conversion efficiency.

[0069] Furthermore, the flywheel energy storage unit model includes five operating modes: constant power flywheel energy storage charging and discharging mode, constant voltage flywheel energy storage charging and discharging mode, and flywheel energy storage shutdown mode, whose operating status is dynamically switched according to the DC grid voltage.

[0070] Finally, the inverter power supply device model includes three operating modes: inverter power supply shutdown mode, constant voltage inverter mode, and constant power inverter mode, and its operating status is dynamically switched according to the DC grid voltage.

[0071] After verification and analysis, a power flow calculation model for the entire DC traction power supply system was formed based on the modular modeling method. The magnitude of the up / down variable resistance and rail resistance depends on the distance between the up / down train and the substation.

[0072] In summary, the 24-pulse rectifier unit is equivalent to a piecewise linearized black box model, divided into a constant voltage regulation zone and an overload droop zone; the train is equivalent to a constant power source model, and the load characteristics of traction and braking conditions are dynamically represented by piecewise functions; the flywheel energy storage system is equivalent to a five-stage operation model, including constant power charging and discharging, constant voltage charging and discharging, and shutdown mode; the inverter energy feeder is equivalent to a multi-mode switching model, dynamically switching between shutdown, constant voltage inverter, and constant power inverter modes according to the DC grid voltage; the AC power supply network is modeled using the improved Newton-Raphson method to form a dual-power zoned power supply structure, and the coordinated optimization of the global energy flow of the system is achieved through alternating AC and DC solutions.

[0073] Step S2: Construct a comprehensive cost index with daily total cost and static payback period as targets.

[0074] Specifically, such as Figure 4As shown, a comprehensive economic model for flywheel energy storage and inverter energy feedback is established. The comprehensive cost index uses daily total cost and static payback period as core parameters. Through normalization and weight allocation, the objective function is obtained. Its expression is:

[0075] ;

[0076] In the formula, Represents the variables in the mixed continuous-discrete search space; This is the daily total cost function, which includes operating electricity costs. Equipment configuration costs and maintenance costs ; This is a static payback period function; Represents the normalization function; , These are the weighting coefficients of total cost and static payback period in the objective function, used to balance economic efficiency and the speed of return on investment.

[0077] Operating electricity costs The formula, calculated using a hybrid AC / DC power flow iterative algorithm, is as follows:

[0078] ;

[0079] In the formula, For electricity price; This indicates the number of all main substations in the power supply system; This indicates the electrical energy consumed by each switch.

[0080] Equipment configuration cost Considering the life-cycle amortization of flywheel energy storage FESE and inverter energy feeder IEFE, the formula is:

[0081] ;

[0082] In the formula, This indicates the total number of traction substations within the system; and Let represent the nonlinear cost functions of FESE and IEFE, respectively; and They represent the first The number of FESEs and IEFEs configured in each substation; and These represent the full lifecycle of FESE and IEFE, respectively.

[0083] Operation and maintenance costs Based on a proportional discount of equipment configuration costs, the formula is as follows:

[0084]

[0085] In the formula, The equipment operation and maintenance cost coefficient is selected as 2%-5% in this embodiment based on experience.

[0086] Static recycling cycle The ratio of initial total investment to annual net cash flow is given by the following formula:

[0087] ;

[0088] In the formula, The base value represents the total electricity cost, the numerator represents the total equipment investment, and the denominator represents the average daily net cost savings.

[0089] In one embodiment of the present invention, the constraint setting includes:

[0090] The number of devices configured is a non-negative integer;

[0091] Voltage threshold meets > > The collaborative logic;

[0092] The flywheel SOC is kept within a safe range of 20%-100% to ensure the physical feasibility of the index calculation.

[0093] Step S3: Treat the number of FESEs and the number of IEEFEs as discrete variables, and set the charging voltage threshold for FESEs. Discharge voltage threshold and the operating voltage threshold of the inverter energy feeder unit As a continuous variable, a hierarchical swarm intelligence optimization method is adopted, aiming to minimize the comprehensive cost index. Under the premise of satisfying the power supply system operation constraints, the output includes the number of FESEs and IEFEs of each traction substation. , , A collaborative optimization configuration scheme.

[0094] In one embodiment of the present invention, as follows: Figure 5The layered swarm intelligence optimization method shown above seeks the optimal solution. First, it initializes the particle swarm based on prior knowledge, setting constraints on the total number of devices and particles in each group. Decision variables are initialized, and the first particle in each group is designated as the leader particle, with the rest as followers. Then, it enters the discrete decision layer iteration. In each iteration, inertia weights and learning factors are calculated, continuous variables are fixed, particle fitness is evaluated, and individual and global optima are updated. The velocity and position of discrete variables are updated according to differentiated strategies for leader and follower particles. Mutation operations are performed, and constraints on discrete variables are addressed. Every Q iterations, the fitness of each group is reordered, and the leader particle and particle number allocation are updated. If the change in global optima fitness is less than a threshold, the iteration at this layer ends early. Next, it enters the continuous decision layer iteration. After fixing discrete variables, parameters are calculated, fitness is evaluated, and the optima are updated. The velocity and position of continuous variables are updated according to roles. Bounded perturbations are performed, and constraints on continuous variables are addressed. If the change in global optima fitness is less than a threshold, the iteration at this layer ends early. Finally, the global optimal solution is returned.

[0095] Specifically, the swarm intelligence optimization method in this embodiment includes the following steps.

[0096] 1) Implementation steps of the layered decoupling mechanism

[0097] 1.1) Layered framework initialization

[0098] Define the type of optimization variables: the discrete variables are the number of flywheel energy storage FESE and inverter energy feeder IEFE in each traction substation CS. and The continuous variable is the charge / discharge voltage threshold of FESE. and IEFE operating voltage threshold .

[0099] The hierarchical objectives are set as follows: the discrete decision layer focuses on equipment layout optimization, the continuous decision layer focuses on voltage parameter calibration, and the two are connected by outputting discrete configuration quantity parameters to the continuous voltage threshold input.

[0100] 1.2) Discrete Decision-Making Layer Optimization Steps

[0101] Fixed continuous variable: The voltage threshold is temporarily set to an engineering experience value, such as... =1750V, eliminating its interference with equipment configuration.

[0102] Population initialization: Particles are generated according to grouping constraints. Each particle contains information on the number of devices at each site, and the first particle in each group is marked as the leader particle. The grouping constraints are constructed using empirical values. For example, group A particles are configured with a total of 1 FW and 1 INV across the entire line, but their locations are unknown and require optimization. Group B particles are configured with a total of 2 FWs and 1 INV across the entire line, but their locations are also unknown and require optimization. The advantage of this approach is that more particles can be allocated to more likely particle populations, thus allocating more computational resources.

[0103] Then, iterative optimization is performed, which includes parameter updates, fitness evaluation, particle updates, mutation operations, and constraint verification.

[0104] Parameter update: Calculate the inertia weight and learning factor according to the following formula to achieve time-varying adaptation.

[0105] ;

[0106] ;

[0107] ;

[0108] In the formula, Indicates the first The inertia weight at the next iteration; T represents the maximum number of iterations for the corresponding layer; and These represent the maximum and minimum weights of the inertia weight, respectively. Represents learning factor The maximum value; and Represents learning factor The maximum and minimum values.

[0109] Fitness assessment: Based on the AC / DC hybrid power flow model, the daily total cost and static payback period are calculated and integrated into a comprehensive cost index J.

[0110] Particle Update: Guides particles to update the position of the global optimal information through fusion, and follows the particles to update based on the optimal information within the group, thereby adjusting the discrete variable, i.e., the number of devices.

[0111] Mutation operations: Perform device location swaps and crossovers within groups with probability, while maintaining the total number of devices.

[0112] Constraint verification: If the total number of devices does not meet the grouping constraint, randomly add or delete the number of devices at the site.

[0113] Resource reallocation: Every Q iterations, the number of particles in each group is recalculated according to the following formula, and more resources are allocated to the groups with better fitness:

[0114] ;

[0115] ;

[0116] ;

[0117] ;

[0118] ;

[0119] In the formula, Let the weights be the assigned weights for the g-th group; This represents the optimized baseline scene fitness value when FESE and IEFE are not configured. This represents the fitness value corresponding to the current optimal solution in the g-th group; This represents the initial number of basic particles allocated to the g-th group; This represents the total number of particles in the particle swarm. Indicates the number of groups; This represents the remaining particles after the initial allocation; This represents a vector of group indexes sorted in descending order of weight. The function returns the index values ​​in ascending order. Represents the weight sequence vectors of each group; This represents the index of the group with the highest weight after sorting. This represents the final number of particles allocated to the optimal group.

[0120] Calculate the weights for each group , with the best fitness within the group Inversely proportional; calculate the initial number of particles in each group. It is linked to the group weight ratio; calculate the remaining particles. Allocate them to the group with the highest weight.

[0121] Finally, a convergence check is performed: if the change in the global optimal fitness is less than the threshold, the discrete layer iteration is exited, and the optimal device configuration scheme is output.

[0122] 1.3) Continuous decision-making level optimization steps

[0123] Fixed discrete variables: The equipment layout output by the discrete decision layer is used as the basis, and the number and location of the equipment are not changed.

[0124] Iterative optimization of the continuous decision layer includes parameter updates, fitness evaluation, particle updates, mutation operations, and constraint verification.

[0125] Parameter update: Updated using the same formula as the discrete decision layer.

[0126] Fitness assessment: With fixed equipment configuration, calculate the comprehensive cost index J, and update the individual and global optimum.

[0127] Particle Update: Adjust continuous variables according to the leader-follower mechanism.

[0128] Mutation operation: Bounded perturbation is applied to the voltage threshold to enhance exploratory behavior. In this embodiment, Gaussian perturbation is used, as shown in the following equation:

[0129] ;

[0130] In the formula, = Define the specific perturbation scale for each dimension: This represents the bounded perturbation constraints for each dimension; , ] represents the variable boundary, that is, the upper and lower boundaries of the d-th variable.

[0131] Constraint verification: An elastic boundary reflection strategy is adopted to bounce out-of-bounds voltage values ​​back to the safe range.

[0132] Finally, a convergence check is performed: if the change in the global optimal fitness is less than the threshold, the iteration of the continuous layer is terminated, and the optimal voltage threshold scheme is output.

[0133] 2) Implementation steps of the group evolution mechanism

[0134] 2.1) Initial group settings

[0135] Grouping Criteria: Based on prior knowledge of site load fluctuation intensity, the particle swarm optimization is divided into G subgroups, with each subgroup constrained by a fixed total number of equipment configurations. For example, the total number of devices in group 1 is set to 1, and the total number of devices in group 2 is set to 2.

[0136] Particle allocation: Prioritize allocating more than 60% of particles to target groups with large load fluctuations and high optimization potential.

[0137] 2.2) Roles within the group and co-evolution

[0138] Role division: The first particle in each group is initially the leader particle, and the rest are follower particles; after every Q iterations, the particle with the best fitness in the group is re-marked as the leader particle.

[0139] Differential Update: In this embodiment, the leading particle and the following particle are updated differently as follows:

[0140] Leading the speed of particle fusion in updating individual and global optimal information, driving global exploration:

[0141] ;

[0142] By following the update speed of individual and group-wide optimal information from particle fusion, local development is enhanced.

[0143] ;

[0144] In the formula, Indicates the first Particles in the next iteration In dimensions The speed on; Indicates the first Inertia weights in the next iteration; and Indicates the first The learning factor at the next iteration; and Represents a random number within a unit interval; Represents particles In dimensions The optimal position of an individual on the [surface / structure]. Represented as dimension The global optimal position; Representing dimensions The optimal position within the group.

[0145] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described; only preferred embodiments of the present invention are illustrated. The descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of the present invention. As long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0146] It should be noted that those skilled in the art can make various modifications and improvements without departing from the inventive concept, and these all fall within the scope of protection of this invention. Therefore, the scope of protection of this invention should be determined by the appended claims.

Claims

1. A method for optimal configuration of flywheel energy storage and medium voltage energy feedback for urban rail traction power supply, characterized in that: Includes the following steps: Step S1: Simultaneously introduce flywheel energy storage unit FESE and inverter energy feeder unit IEFE into the urban rail transit power supply system, and construct an AC / DC hybrid power flow analysis model for urban rail transit. Step S2: Construct a comprehensive cost index with daily total cost and static payback period as targets; Step S3: Treat the number of FESEs and the number of IEEFEs as discrete variables, and set the charging voltage threshold for FESEs. Discharge voltage threshold and the operating voltage threshold of the inverter energy feeder unit As a continuous variable, a hierarchical swarm intelligence optimization method is adopted, aiming to minimize the comprehensive cost index. Under the premise of satisfying the power supply system operation constraints, the output includes the number of FESEs and IEFEs of each traction substation. , , A collaborative optimization configuration scheme; Step S3 includes: Step S31: Based on prior knowledge of the intensity of site load fluctuations, the particle swarm is divided into several subgroups by setting the number of groups, and a constraint on the total number of devices is set for each subgroup. Step S32: Initialize the particle swarm according to the total number of devices in each subgroup, generate a discrete solution containing the number of FESEs and IEFEs of each traction substation, and mark the leading particle and the following particle in each group; Step S33: Perform discrete decision-making layer optimization: fix continuous variables and perform iterative optimization of discrete variables based on the comprehensive cost index; Step S34: Perform continuous decision-making layer optimization: Based on the equipment layout output by the discrete decision-making layer, fix the number of FESEs and IEFEs of the traction substation, and iteratively optimize the continuous variables based on the comprehensive cost index; Step S35: When any layer satisfies the early stop condition or reaches the iteration limit, output the optimal result, and combine the output of the discrete decision layer with the output of the continuous decision layer to obtain the number of FESEs and IEEFEs for each traction substation. , , A collaborative optimization configuration scheme.

2. The method according to claim 1, characterized in that: The AC / DC hybrid power flow analysis model for urban rail transit in step S1 includes: equating the 24-pulse rectifier unit in the traction power supply system to a piecewise linearized external characteristic model and dividing it into a constant voltage regulation zone and an overload droop zone; equating the train to a constant power load model; modeling the flywheel energy storage unit FESE and the inverter energy feed unit IEFE as multi-mode switching models based on DC grid voltage; and using an alternating AC / DC solution method to complete the energy flow calculation and convergence determination.

3. The method of claim 2, wherein the method is characterized by: The modeling expression for the 24-pulse rectifier unit in the AC / DC hybrid power flow analysis model of urban rail transit is as follows: ； In the formula, This indicates the DC side voltage of the rectifier unit; This indicates the no-load voltage of the rectifier unit; Indicates the drooping slope; Indicates the DC-side output current; This indicates the rated current of the rectifier unit; This indicates the rated DC output voltage.

4. The method of claim 2, wherein the method is characterized by: The modeling expression for trains in the AC / DC hybrid power flow analysis model of urban rail transit is as follows: ； wherein is the catenary voltage at the pantograph; denotes the train current; denotes the train mechanical power; denotes the electromechanical conversion efficiency.

5. The method for coordinated optimization configuration of flywheel energy storage and medium-voltage energy feeder for urban rail transit traction power supply according to claim 1, characterized in that: In step S2, the total daily cost includes operating electricity costs, equipment configuration costs, and equipment operation and maintenance costs; The operating electricity cost The expression is: ； The device configuration cost The expression is: ； The device operation and maintenance cost The expression is: ； In the formula, Indicates electricity price; This indicates the number of all main substations in the power supply system; This indicates the electrical energy consumed by each switch. This indicates the total number of traction substations within the system; and Let represent the nonlinear cost functions of FESE and IEFE, respectively; and They represent the first The number of FESEs and IEFEs configured in each substation; and These represent the full lifecycle of FESE and IEFE, respectively. This represents the equipment operation and maintenance cost coefficient.

6. The method for coordinated optimization configuration of flywheel energy storage and medium-voltage energy feeder for urban rail transit traction power supply according to claim 5, characterized in that: In step S2, the static recovery period The expression is: ； wherein represents the base value scheme total electrical energy expenditure.

7. The method of claim 1, wherein the method is a method of optimal configuration of flywheel energy storage and medium voltage energy feedback for urban rail traction power supply, characterized in that: In steps S33 and S34, the leading particle and the following particle are updated differentially. The guiding particle is updated using the following expression: ； The following particles are updated using the following expression: ； In the formula, Indicates the first In the next iteration, the particle In dimensions The speed on; Indicates the first Inertia weights in the next iteration; and Indicates the first The learning factor at the next iteration; and Represents a random number within a unit interval; Represents particles In dimensions The optimal position of an individual on the [unclear]; Represented as dimension The global optimal position; Representing dimensions The optimal position within the group.

8. The method of claim 7, wherein the method is characterized by: In steps S33 and S34, the inertia weight and learning factor are updated using the following expressions: ； ； ； In the formula, Indicates the first The inertia weight at the next iteration; T represents the maximum number of iterations for the corresponding layer; and These represent the maximum and minimum weights of the inertia weight, respectively. Represents learning factor The maximum value; and Represents learning factor The maximum and minimum values.

9. The method of claim 1, wherein the method is a method of optimal configuration of flywheel energy storage and medium voltage energy feedback for urban rail traction power supply, characterized in that: When performing discrete decision-level optimization, after every Q iterations, the number of particles in each group is recalculated and redistributed according to the following formula: ； ； ； ； ； In the formula, Let the weights be the assigned weights for the g-th group; This represents the optimized baseline scene fitness value when FESE and IEFE are not configured. This represents the fitness value corresponding to the current optimal solution in the g-th group; This represents the initial number of basic particles allocated to the g-th group; This represents the total number of particles in the particle swarm. Indicates the number of groups; This represents the remaining particles after the initial allocation; This represents a vector of group indexes sorted in descending order of weight. The function returns the index values ​​in ascending order. Represents the weight sequence vectors of each group; This represents the index of the group with the highest weight after sorting. This represents the final number of particles allocated to the optimal group.

Images