Planning and optimal configuration method for active traction network energy system

Abstract

The invention relates to a planning and optimal configuration method for an active traction network energy system, and the method achieves the construction of a universal planning and optimal configuration method through three typical scenes, and carries out the planning and configuration of a new energy and energy storage system in a traction network in consideration of the operation cost and a full-life-cycle net present value. Therefore, it is guaranteed that the railway system has economic feasibility, carbon reduction of the railway system is guaranteed, and meanwhile energy self-consistency is achieved to the maximum extent. According to the method, a double-layer optimization model is adopted for capacity configuration optimization, an upper-layer optimization model adopts an intelligent algorithm to perform capacity planning on a renewable energy system and an energy storage system so as to achieve the lowest net present value cost in the whole life cycle, and a lower-layer optimization model calculates the operation cost through a solver so as to achieve the purpose of reducing the operation cost of operation and maintenance, electricity purchasing, electricity selling and the like. The upper layer and the lower layer depend on each other, the whole life cycle cost and the real-time operation efficiency are both considered, and the scientificity and operability of energy system planning are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to a green energy system for a railway active traction network, in particular to a planning and optimization configuration method of an active traction network energy system. BACKGROUND

[0002] Railway transportation, as an important part of modern transportation system, has a large demand for energy. Connecting a renewable energy system to the railway traction network can reduce dependence on the power grid and is beneficial to environmental protection. In the existing railway traction network energy system, only one-way power supply of the traction network is considered, that is, power is transmitted from the high-voltage main network to the traction load in stages, and the traction transformer is mainly used to supply the locomotive with power. This power supply mode is relatively simple, but it depends on centralized control.

[0003] For the layout of the renewable energy system along the railway, the geographical environment of the railway needs to be considered, and whether the railway along the line has grid-connected conditions needs to be considered. In particular, the geographical environment in the east, central and west of China is very different. Therefore, to configure an active traction network energy system with universality, it is necessary to consider three typical application scenarios of grid-connected self-generation and self-use, grid-connected surplus power on-grid and off-grid operation, support the active traction network architecture with bidirectional energy flow, and realize supply and demand balance through effective management of new energy and energy storage units. The characteristics of bidirectional flow make the active traction network not only more flexible in adjusting the power supply path when facing power grid failures and resisting external power disturbances on locomotive traction, but also reduce power consumption from the power grid and reduce electricity costs. This can maximize the economic value and make it have promotional value. Based on the above considerations, the inventors propose an energy system optimization configuration method suitable for active traction network. SUMMARY

[0004] The present application plans and configures the optimization of the active traction network energy system, and gives a universal optimization configuration method through three typical scenarios. The operation cost and the net present value of the whole life cycle are considered to plan and configure the new energy and energy storage system in the traction network, so as to ensure the economic feasibility of the railway system and maximize the realization of energy self-consistency while reducing carbon in the railway system.

[0005] To achieve the above purpose, the technical scheme adopted by the present application is as follows:

[0006] A planning method of an active traction network energy system, comprising the following contents:

[0007] Firstly, according to the application scenario of the railway, the power load demand fluctuation curve is determined, and the space of the railway is divided and the area is calculated;

[0008] Secondly, the attributes of the divided space are evaluated, and the energy-oriented active catenary energy system equipment is deployed in the corresponding available space, and the capacity constraint and configuration optimization model parameters of the equipment are given;

[0009] Finally, according to the conditions of the regional power grid where the railway is located and the power supply policy, the operation mode of the active catenary energy system is determined, and then the capacity configuration of the active catenary energy system is constructed.

[0010] An optimization configuration method of an active catenary energy system, comprising constructing a K-means++ clustering algorithm model and a double-layer optimization model, the double-layer optimization model being divided into an upper-layer optimization model and a lower-layer optimization model, and performing optimization configuration according to the following steps:

[0011] Step 1): The K-means++ clustering algorithm model is used to perform different seasonal clustering calculations on the natural resource historical data and the actual energy consumption data of the traction substation, and the data after clustering calculation is output as the input data of the lower-layer optimization model;

[0012] Step 2): The operation parameters and the planned configuration parameters of each device in the energy system are input into the upper-layer optimization model, and the iteration number, population size, adaptive crossover rate and mutation rate are set, the adaptive genetic algorithm is used to initialize the initial capacity configuration of each device in the energy system by using the upper-layer optimization model, and the initial capacity configuration result is transmitted to the lower-layer optimization model;

[0013] Step 3): The lower-layer optimization model calculates the operation scheduling cost of the energy system based on the clustered data and the initial capacity configuration data, and optimizes the operation scheduling cost, and the optimal operation scheduling cost output by the lower-layer optimization model is transmitted to the upper-layer optimization model;

[0014] Step 4): The upper-layer optimization model calculates the total net present value of the full life cycle cost of the energy system according to the received optimal operation scheduling cost, and judges whether the maximum iteration number is reached;

[0015] If the maximum iteration number is not reached, the adaptive genetic algorithm is used to generate a new population by selecting and adjusting the crossover rate and mutation rate, update the configuration capacity of each device in the energy system, and transmit the updated parameters to the lower-layer optimization model to execute step 3) again;

[0016] If the maximum iteration number is reached, the upper-layer optimization model outputs the configuration capacity of each device in the energy system and the full life cycle net present value cost.

[0017] Further, in step 1), the specific implementation method is as follows:

[0018] Step 1.1): input the 8760 time series data set of the traction substation, the data set includes wind speed, irradiance, temperature, actual energy consumption of load, and randomly initialize the first cluster center;

[0019] Step 1.2): calculate the distance of each sample point to the initialized cluster center, and then use the probability proportional sampling method to select a new data point as the next cluster center;

[0020] Step 1.3): determine the cluster count k representing the seasons of summer, winter and spring and autumn, k = 3;

[0021] Step 1.4): calculate the distance between each sample and the cluster center point, and assign the sample to the nearest center point, then calculate the average value of the feature vector of each cluster sample, and use the average value to iteratively update the cluster center until convergence;

[0022] Step 1.5): after the convergence of the cluster center, output the k cluster center data of wind speed, irradiance, temperature and load energy consumption.

[0023] Further, in the step 2), the iteration number, population size, adaptive crossover rate and mutation rate formula are as follows:

[0024]

[0025] In the above formula, η cr and η mr are the adaptive crossover rate and mutation rate respectively; n ic,n is the real-time iteration number.

[0026] Further, the energy system includes a traction substation, a photovoltaic power generation system, a wind turbine power generation system and an energy storage system, and the operation mode of the energy system is suitable for operation mode A or operation mode B or operation mode C, wherein the operation mode A is for operation under grid-connected condition, and the grid can only supply energy to the energy storage system and the traction load in one direction; the operation mode B is for operation under grid-connected condition, and the grid can supply energy to the energy storage system and the traction load, and can also absorb excess power of the photovoltaic power generation system and the wind turbine power generation system; the operation mode C is for operation under off-grid condition.

[0027] Further, the wind turbine power generation system adopts the following configuration model:

[0028]

[0029] In the above formula, P w,t is the real-time output power of the wind turbine; P wr is the rated power of the wind turbine; v is the real-time wind speed; v ci , v co and v rCut-in wind speed, cut-out wind speed and rated wind speed of the fan, respectively;

[0030] The photovoltaic power generation system adopts the following configuration model:

[0031]

[0032] In the above formula, P pv,t and P pv are the real-time output power and rated power of the photovoltaic, respectively; f pv is the performance attenuation coefficient; G c,t is the real-time solar irradiance of the working point; G stc is the solar irradiance under standard conditions; k is the power temperature coefficient; T c,t is the real-time temperature of the working point; T STC is the temperature under standard conditions;

[0033] The energy storage system adopts the following configuration model:

[0034]

[0035] In the above formula, SOC t and SOC t-1 are the SOC of the energy storage system at time t and t-1, respectively; P es,t is the real-time operating power of the energy storage system, and when P es,t <0, it indicates discharging, and when P es,t >0, it indicates charging; η c and η d are the real-time charging power and discharging power of the energy storage system, respectively; δ is the self-discharge rate; T es is the duration of the charging / discharging operation.

[0036] Further, in the step 4), the total net present value of the full life cycle cost is calculated according to the following formula:

[0037]

[0038] In the above formula, C A , C B and C C are the total net present values of the full life cycle costs under the operation mode A, the operation mode B and the operation mode C, respectively; C ic , C r and C s are the initial construction, replacement and residual value costs, respectively; C dc is the demand charge; C O_A , C O_B and C O_C are the operating expenses of the operation modes A, B and C, respectively; C c_w , C c_pvand C c_es are the unit capacity construction costs of the wind turbine, photovoltaic and energy storage system, respectively; a w , a pv , and a es are the residual value coefficients of the wind turbine, photovoltaic and energy storage system at the end of the life cycle, respectively; T is the operating life of the active traction power grid system; T RES and T ESS are the operating life of the renewable energy and energy storage system, respectively; r is the annual interest rate; d op is the annual operating hours; c bdc is the basic demand charge; η T is the utilization rate of the main transformer; E T is the capacity of the main transformer; P w , P pv , E es are the configuration capacities of the wind turbine, photovoltaic and energy storage system, respectively;

[0039] Among them, the configuration capacity constraints of the wind turbine, photovoltaic and energy storage system are as follows:

[0040]

[0041] In the above formula, S w_min , S pv_min and S es_min are the lower threshold values of the configuration capacities of the wind turbine, photovoltaic and energy storage system, respectively; S w_max , S pv_max and S es_max are the upper threshold values of the configuration capacities of the wind turbine, photovoltaic and energy storage system, respectively.

[0042] Further, in the step 3), the lower layer optimization model optimizes the operation scheduling cost of the energy system based on the following formula:

[0043]

[0044] In the above formula, C O is the daily operation cost of the energy system; C pur is the electricity purchase cost under operation modes A and B; C sale is the electricity selling cost under operation mode B; C O_w , C O_pv and C O_es are the total operation costs of the wind turbine, photovoltaic and energy storage system, respectively; β w , β pv and β es are the operation costs of the wind turbine, photovoltaic and energy storage system, respectively; P esch,t and P esdis,t are the real-time charging and discharging powers of the energy storage system; P pur,t and P sale,trespectively, are real-time power of electricity purchase and sale; c pur,t and c sale,t respectively, are electricity purchase and sale prices; C o_grid is the total grid interaction cost;

[0045] The real-time power constraints corresponding to the three operation modes are as follows:

[0046] P w,t + P pv,t + P esdis,t + P pur,t - P esch,t - P sale,t = P train,t

[0047] In the above formula, P train,t is the real-time power of the locomotive;

[0048] The power generation constraints of the wind turbine and photovoltaic system are as follows:

[0049]

[0050] In the above formula, P w_min and P pv_min are the minimum output powers of the wind turbine and photovoltaic system; P w_max and P pv_max are the maximum output powers of the wind turbine and photovoltaic system;

[0051] The real-time SOC and charging / discharging power constraints of the energy storage system are as follows:

[0052]

[0053] In the above formula, SOC min and SOC max are the lower and upper limits of SOC, taking values of 0.1 and 0.9 respectively; P es_min and P es_max represent the lower and upper threshold values of the charging and discharging powers of the energy storage system;

[0054] Under grid-connected condition, the constraints of electricity purchase and sale are as follows:

[0055]

[0056] In the above formula, I grid,t is a binary variable indicating the state of power exchange, I grid,t = 1 indicates that the energy system purchases electricity from the grid, and I grid,t = 0 indicates that the energy system sells electricity to the grid.

[0057] The energy system optimization configuration method provided by the application supports an active traction grid architecture with bidirectional energy flow, enhances the adaptability and power supply flexibility of the active traction grid system, and has universality and can be adapted to different regional grid conditions and policy requirements. The application adopts a double-layer optimization model for capacity configuration optimization, and the upper and lower optimization layers depend on each other, each layer has different objective functions and constraint sets, wherein the upper optimization model is used for capacity planning of the renewable energy system and the energy storage system, the capacity is limited by land availability and overall investment, and the upper optimization target aims to minimize the net present value cost in the whole life cycle, and the lower optimization target aims to reduce the operation cost including operation and maintenance, power purchase and sale. The double-layer optimization model adopted by the application combines intelligent algorithms and solvers, takes into account the whole life cycle cost and real-time operation efficiency, and improves the scientificity and operability of the energy system planning. BRIEF DESCRIPTION OF DRAWINGS

[0058] Figure 1 It is a topological architecture diagram of the active traction grid in the embodiment.

[0059] Figure 2 It is an energy flow schematic diagram corresponding to three operation modes.

[0060] Figure 3 It is a structure diagram of the double-layer optimization model in the embodiment.

[0061] Figure 4 It is a flowchart of the optimization configuration method of the energy system in the embodiment.

[0062] Figure 5 It is a schematic diagram of the planning and configuration research method of the active traction grid energy system. DETAILED DESCRIPTION

[0063] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application.

[0064] The embodiment discloses an active traction grid energy system, and simultaneously discloses a planning method and an optimization configuration method of the energy system. The active traction grid in the embodiment refers to a traction grid with distributed or decentralized energy inside and with control ability and non-unidirectional energy flow. The energy sources of the whole system prefer to use green energy generated by wind turbines and photovoltaic power generation, and the energy storage system is used to realize green electricity transition and emergency power supply without external power supply support. The active traction grid architecture provided in the embodiment is as shown in the drawing, which includes a traction substation, a wind turbine power generation system, a photovoltaic power generation system, an energy storage system, and whether to have a power grid support according to the region where the railway is located. Figure 1

[0065] ​The active traction network architecture given in this embodiment takes into account the differences in the geographical environment of the railway and sets three operating modes according to different energy usage scenarios, regional power grid structures, and the stability and interactivity of power grid energy supply, namely operating mode A, operating mode B, and operating mode C. Among them, operating mode A is suitable for grid-connected scenarios where the power grid cannot flow in both directions. When the renewable energy system is insufficient, the power grid can provide energy support. The energy flow in this mode is as follows: Figure 2 As shown in (a) in the figure, operation mode A is mainly suitable for traction substations and stations that use fixed electricity prices and have no flexible electricity price policy, and is suitable for use in most areas of my country. Operation mode B is operated in a grid-connected scenario. Renewable energy generation can meet the electricity demand of traction loads and energy storage systems on demand. The grid can not only provide electricity but also absorb excess electricity. The energy flow in this mode is as follows: Figure 2 As shown in (b) in the figure, operation mode B is suitable for traction substations and stations in developed regions with flexible power market trading mechanisms, such as the eastern and central regions of my country. Operation mode C is operated in an off-grid scenario, with renewable energy and energy storage systems as the main energy sources, and its energy flow is as follows: Figure 2 As shown in (c) in the figure, operating mode C is suitable for non-electrified remote areas such as plateaus, deserts and isolated islands, such as the western region of my country.

[0066] This embodiment first provides the planning method of the active traction network energy system mentioned above. It is necessary to clarify the selection and optimal configuration capacity of the infrastructure, the traffic environment where the project is implemented, and establish constraints for the configuration optimization model. According to the specific traffic scene and project elements, the optimization model parameters are configured. The overall planning method is as follows Figure 5 As shown, including the following:

[0067] First, based on the application scenario of the railway, the electricity load demand fluctuation curve is determined, and the space where the railway is located is divided and the area is calculated.

[0068] Secondly, the available railway space resources are evaluated and, based on the space attributes and energy orientation, the railway scenario assets shown in Table 2 are used to determine the feasible deployment methods, installable equipment, capacity constraints, and related parameters of the configuration optimization model.

[0069] Finally, based on the conditions of the regional power grid and the power supply policy of the railway, the operating mode of the active traction network energy system is determined, and then the capacity configuration of the active traction network energy system is constructed.

[0070] Table 2: Available space on railways and energy equipment deployment in corresponding space

[0071]

[0072] After the planning of the active traction net energy system is clear and completed, the active traction net energy system is optimized and configured, and a related optimization and configuration model is designed as follows:

[0073] Among them, the configuration model of the fan power generation system is as follows:

[0074]

[0075] In the above formula, P w,t is the real-time output power of the fan; P wr is the rated power of the fan; v is the real-time wind speed; v ci , v co and v r are the cut-in wind speed, the cut-out wind speed and the rated wind speed of the fan respectively.

[0076] The configuration model of the photovoltaic power generation system is as follows:

[0077]

[0078] In the above formula, P pv,t and P pv are the real-time output power and the rated power of the photovoltaic respectively; f pv is the performance attenuation coefficient; G c,t is the real-time solar irradiance of the working point; G stc is the solar irradiance under standard conditions; k is the power temperature coefficient; T c,t is the real-time temperature of the working point; T STC is the temperature under standard conditions.

[0079] The configuration model of the energy storage system is as follows:

[0080]

[0081] In the above formula, SOC t and SOC t-1 are the SOC of the energy storage system at time t and t-1 respectively; P es,t is the real-time operation power of the energy storage system, and when P es,t <0, it indicates discharging, and when P es,t >0, it indicates charging; η c and η d are the real-time charging power and discharging power of the energy storage system respectively; δ is the self-discharge rate; T es is the duration of the charging / discharging operation.

[0082] For the configuration optimization of the above active traction net energy system, the present application adopts the following configuration model: Figure 3The double-layer optimization model is divided into an upper-layer optimization model and a lower-layer optimization model, and the upper-layer and lower-layer optimization models depend on each other, each layer has a different optimization objective function and constraint set, wherein the upper-layer optimization is configured for renewable energy and energy storage system capacity, the capacity is subject to land availability and overall investment constraints, and the upper-layer objective aims to minimize the net present cost in the whole life cycle; the lower-layer optimization aims to reduce operating costs including operation and maintenance, electricity purchase and sale, and the like. The optimization configuration method of the present application is described in detail below. Figure 4 The optimization configuration method of the present application is described in detail below.

[0083] Step 1): Construct a K-means++ clustering algorithm model, before configuration optimization by the double-layer optimization model, the present application first uses the K-means++ clustering algorithm model to perform different seasonal clustering calculation on the natural resource historical data and actual energy consumption data of the traction substation, so as to ensure the accuracy of the solution while reducing the training time of the double-layer optimization model. The clustering data output by the K-means++ clustering algorithm model is used as the input data of the lower-layer optimization model. The specific execution mode of this step is as follows:

[0084] Step 1.1): input the 8760 time series data set of the traction substation, the data set includes wind speed, irradiance, temperature, actual load energy consumption, and randomly initialize the first cluster center;

[0085] Step 1.2): calculate the distance of each sample point to the initialized clustering center, then use the probability proportional sampling method to select a new data point as the next cluster center, and preferentially consider the point with the minimum distance;

[0086] Step 1.3): determine the cluster count k representing the season (summer, winter and transition season, transition season refers to spring and autumn), k = 3;

[0087] Step 1.4): calculate the distance between each sample and the clustering center point, and assign the sample to the nearest center point, then calculate the average value of the feature vector of each clustering sample, and use the average value to iteratively update the clustering center until convergence;

[0088] Step 1.5): after the clustering center converges, output the k cluster center data of the wind speed, irradiance, temperature and load energy consumption.

[0089] Step 2): build a double-layer optimization model, first input the operating parameters of each device in the energy system and the planning configuration parameters into the upper-layer optimization model, set the iteration number, population size, construct adaptive crossover rate and mutation rate, use the upper-layer optimization model to initialize the initial capacity configuration of each device in the energy system by using the adaptive genetic algorithm, and pass the initial capacity configuration result to the lower-layer optimization model.

[0090] Wherein, the number of iterations, population size, adaptive crossover rate and mutation rate setting formula as follows:

[0091]

[0092] In the above formula, η cr and η mr are adaptive crossover rate and mutation rate, respectively; n ic,n is the number of real-time iterations.

[0093] Step 3): the lower layer optimization model based on the clustered data and the initial capacity configuration data, through the YALMIP toolbox with Gurobi solver to carry out the lower layer optimization, using the solver to calculate the operation scheduling cost of energy system and optimize the operation scheduling cost, and then output the optimal operation scheduling cost to the upper layer optimization model.

[0094] In the lower layer optimization, different operation modes are referred to, taking operation mode C as the benchmark, operation mode C only contains operation cost without including electricity transaction related cost, while operation mode A and operation mode B both contain electricity purchase cost, and operation mode B also contains electricity sale cost. The lower layer optimization model optimizes the operation scheduling cost of energy system based on the following formula:

[0095]

[0096] In the above formula, C O is the daily operation cost of energy system; C pur is the electricity purchase cost under operation mode A and B; C sale is the electricity sale cost under operation mode B; C O_w , C O_pv and C O_es are the total operation cost of wind turbine, photovoltaic and energy storage system, respectively; β w , β pv and β es are the operation cost of wind turbine, photovoltaic and energy storage system, respectively; P esch,t and P esdis,t are the real-time charging and discharging power of energy storage system, respectively; P pur,t and P sale,t are the real-time electricity purchase and sale power, respectively; c pur,t and c sale,t are the electricity purchase and sale price, respectively; C o_grid is the total grid interaction cost.

[0097] The constraint conditions of all operation modes are set as follows:

[0098] (1) Real-time power constraint condition is as follows:

[0099] P w,t +Ppv,t +P esdis,t +P pur,t -P esch,t -P sale,t =P train,t

[0100] In the above formula, P train,t is the real-time power of the locomotive.

[0101] (2) The power generation constraints of the fan and photovoltaic power generation system are as follows:

[0102]

[0103] In the above formula, P w_min and P pv_min are the minimum output powers of the fan and photovoltaic system respectively; P w_max and P pv_max are the maximum output powers of the fan and photovoltaic system respectively.

[0104] (3) The real-time SOC and charging and discharging power constraints of the energy storage system are as follows:

[0105]

[0106] In the above formula, SOC min and SOC max are the lower and upper limits of SOC, respectively, and take values of 0.1 and 0.9; P es_min and P es_max represent the lower and upper thresholds of the charging and discharging powers of the energy storage system.

[0107] (4) In the grid-connected case, the constraints of power purchase and power sale are as follows:

[0108]

[0109] In the above formula, I grid,t is a binary variable indicating the state of power exchange, I grid,t = 1 indicates that the energy system purchases power from the grid, and I grid,t = 0 indicates that the energy system sells power to the grid.

[0110] Step 4): The upper layer optimization model calculates the total net present value of the life cycle cost of the energy system according to the received optimal operation scheduling cost, and determines whether the maximum iteration number is reached; if the maximum iteration number is not reached, the upper layer optimization model generates a new population by selecting, adjusting the crossover rate and mutation rate using the adaptive genetic algorithm, updates the configuration capacity of each device in the energy system, and passes the updated parameters to the lower layer optimization model to re-execute step 3); if the maximum iteration number has been reached, the upper layer optimization model outputs the configuration capacity of each device in the energy system and the net present value cost of the life cycle.

[0111] In step 4), the total net present value cost of the life cycle cost includes initial construction, operation, replacement and residual value cost, and the future cost is converted into present value by using the discount rate method, which is calculated according to the following formula:

[0112]

[0113] In the above formula, C A , C B and C C are the total net present value of the life cycle cost under the operation mode A, operation mode B and operation mode C, respectively; C ic , C r and C s are the initial construction, replacement and residual value cost, respectively; C dc is the demand charge; C O_A , C O_B and C O_C are the operation cost of operation mode A, B and C, respectively; C c_w , C c_pv and C c_es are the unit capacity construction cost of the fan, photovoltaic and energy storage system, respectively; α w , α pv and α es are the residual value coefficients of the fan, photovoltaic and energy storage system at the end of the life cycle, respectively; T is the operation life of the active traction network system; T RES and T ESS are the operation life of the renewable energy and energy storage system, respectively; r is the annual interest rate; d op is the annual operation hours; c bdc is the basic demand charge; η T is the utilization rate of the main transformer; E T is the capacity of the main transformer; P w , P pv and E es are the configuration capacity of the fan, photovoltaic and energy storage system, respectively.

[0114] During the upper layer optimization, the configuration capacity constraints of the fan, photovoltaic and energy storage system are as follows:

[0115]

[0116] In the above formula, S w_min , S pv_min and S es_min are respectively the lower threshold of the configuration capacity of the wind turbine, photovoltaic and energy storage system; S w_max , S pv_max and S es_max are respectively the upper threshold of the configuration capacity of the wind turbine, photovoltaic and energy storage system.

[0117] While embodiments of the application have been shown and described, it is to be understood that the application is not limited to the details of the embodiments described, since the scope of the application is defined with respect to the appended claims.

Claims

1. A planning method for an active traction network energy system, characterized in that: Includes the following: First, based on the railway's application scenario, determine the power load demand fluctuation curve, and divide the railway space and calculate its area. Secondly, the attributes of the divided space are evaluated, and the equipment of the active traction network energy system is deployed in the corresponding available space based on energy. The capacity constraints of the equipment and the configuration optimization model parameters are given. Finally, based on the conditions of the regional power grid and the power supply policy of the railway, the operating mode of the active traction network energy system is determined, and then the capacity configuration of the active traction network energy system is constructed.

2. A method for optimizing the configuration of an active traction network energy system, characterized by: This involves building a K-means++ clustering algorithm model and a two-layer optimization model. The two-layer optimization model is divided into an upper-layer optimization model and a lower-layer optimization model. The optimization configuration is performed according to the following steps: Step 1): First, the K-means++ clustering algorithm model is used to perform seasonal clustering calculations on the natural resource historical data and actual energy consumption data of the traction substation, and the clustered data are output as the input data of the lower optimization model; Step 2): Input the operating parameters and planned configuration parameters of each device in the energy system into the upper-level optimization model, set the number of iterations, population size, adaptive crossover rate, and mutation rate, and use the upper-level optimization model to initialize the initial capacity configuration of each device in the energy system using an adaptive genetic algorithm. The initial capacity configuration result is then passed to the lower-level optimization model. Step 3): Based on the clustered data and the initial capacity configuration data, the lower-level optimization model calculates the operation and scheduling cost of the energy system through the Gurobi solver and optimizes the operation and scheduling cost. The lower-level optimization model outputs the optimal operation and scheduling cost and passes it to the upper-level optimization model. Step 4): The upper optimization model calculates the total net present value of the life cycle cost of the energy system based on the received optimal operation scheduling cost, and determines whether the maximum number of iterations has been reached; If the maximum number of iterations is not reached, the upper optimization model uses an adaptive genetic algorithm to generate a new population by selecting and adjusting the crossover rate and mutation rate, updates the configuration capacity of each device in the energy system, and passes the updated parameters to the lower optimization model to re-execute step 3); If the maximum number of iterations has been reached, the upper optimization model outputs the configuration capacity of each device in the energy system and the net present value cost of the entire life cycle.

3. The method for optimizing the configuration of an active traction network energy system according to claim 2, characterized in that: In the step 1), the specific implementation method is as follows: Step 1.1): Input the 8760 time series data set of the traction substation, which includes wind speed, irradiance, temperature, and actual load energy consumption, and randomly initialize the first cluster center; Step 1.2): Calculate the distance from each sample point to the initialized cluster center, and then use the probability proportional sampling method to select a new data point as the next cluster center; Step 1.3): Determine the cluster count k representing the seasons of summer, winter, spring and autumn, where k = 3; Step 1.4): Calculate the distance between each sample and the cluster center point, and assign the sample to the nearest center point. Then calculate the average value of the feature vector of each cluster sample and use this average value to iteratively update the cluster center until convergence; Step 1.5): After the cluster centers converge, output k cluster center data of wind speed, irradiance, temperature and load energy consumption.

4. The method for optimizing the configuration of an active traction network energy system according to claim 2, characterized in that: In step 2), the formulas for the number of iterations, population size, adaptive crossover rate, and mutation rate are as follows: In the above formula, η cr and η mr are adaptive crossover rate and mutation rate respectively; n ic,n is the number of real-time iterations.

5. The method for optimizing the configuration of an active traction network energy system according to claim 2, characterized in that: The energy system includes a traction substation, a photovoltaic power generation system, a wind turbine power generation system and an energy storage system. The operating mode of the energy system is applicable to operating mode A, operating mode B or operating mode C. Among them, operating mode A is for operation in a grid-connected state, and the power grid can only supply energy to the energy storage system and traction load in one direction; operating mode B is for operation in a grid-connected state, and the power grid can not only supply energy to the energy storage system and traction load, but also absorb the excess electricity of the photovoltaic power generation system and the wind turbine power generation system; operating mode C is for operation in an off-grid state.

6. The method for optimizing the configuration of an active traction network energy system according to claim 5, characterized in that: The wind turbine power generation system adopts the following configuration model: In the above formula, P w,t is the real-time output power of the wind turbine; P wr is the rated power of the fan; v is the real-time wind speed; v ci 、v co and v r are the cut-in wind speed, cut-out wind speed and rated wind speed of the fan respectively; The photovoltaic power generation system adopts the following configuration model: In the above formula, P pv,t and P pv are the real-time output power and rated power of photovoltaic power respectively; f pv is the performance attenuation coefficient; G c,t is the real-time solar irradiance at the working point; G stc is the solar irradiance under standard conditions; k is the power temperature coefficient; T c,t is the real-time temperature of the working point; T STC is the temperature under standard conditions; The energy storage system adopts the following configuration model: In the above formula, SOC t and SOC t-1 are the energy storage system SOC at time t and t-1 respectively; P es,t is the real-time operating power of the energy storage system. es,t <0 indicates discharge, P es,t >0 means charging; η c and η d are the real-time charging power and discharging power of the energy storage system respectively; δ is the self-discharge rate; T es is the duration of the charge / discharge operation.

7. The method for optimizing the configuration of an active traction network energy system according to claim 6, characterized in that: In step 4), the total net present value of the life cycle cost is calculated according to the following formula: In the above formula, C A 、C B and C C are the total net present value of the life cycle cost under operation mode A, operation mode B, and operation mode C respectively; C ic 、C r and C s are the initial construction, replacement and residual value costs respectively; C dc is the demand electricity charge; C O_A 、C O_B and C O_C are the operating costs of operation modes A, B and C respectively; C c_w 、C c_pv and C c_es are the unit capacity construction costs of wind turbines, photovoltaics, and energy storage systems respectively; α w , α pv and α es are the residual value coefficients of wind turbines, photovoltaic systems, and energy storage systems at the end of their life cycles; T is the operating life of the active traction network system; T RES and T ESS are the operating years of renewable energy and energy storage system respectively; r is the annual interest rate; d op is the annual operating hours; c bdc As the basic demand charge; η T E is the main transformer utilization rate; T is the main transformer capacity; P w 、P pv 、E es Configure capacity for wind turbines, photovoltaics, and energy storage systems respectively; Among them, the configuration capacity constraints of wind turbines, photovoltaics, and energy storage systems are as follows: In the above formula, S w_min 、S pv_min and S es_min Configure the lower capacity thresholds for wind turbines, photovoltaics, and energy storage systems respectively; S w_max 、S pv_max and S es_max Configure capacity upper limit thresholds for wind turbines, photovoltaics, and energy storage systems respectively.

8. The method for optimizing the configuration of an active traction network energy system according to claim 7, characterized in that: In step 3), the lower optimization model optimizes the operation and scheduling cost of the energy system based on the following formula: In the above formula, C O is the daily operating cost of the energy system; C pur is the electricity purchase cost under operation modes A and B; C sale is the electricity sales cost under operation mode B; C O_w 、C O_pv and C O_es are the total operating costs of wind turbines, photovoltaic systems and energy storage systems respectively; β w , β pv and β es are the operating costs of wind turbines, photovoltaic systems, and energy storage systems respectively; P esch,t and P esdis,t are the real-time charging and discharging power of the energy storage system respectively; P pur,t and P sale,t are the real-time power purchase and power sales respectively; c pur,t and c sale,t are the electricity purchase and sales prices respectively; C o_grid is the total grid interaction cost; The real-time power constraints corresponding to the three operating modes are as follows: P w,t +P pv,t +P esdis,t +P pur,t -P esch,t -P sale,t =P train,t In the above formula, P train,t The real-time power of the locomotive; The power generation constraints of wind turbines and photovoltaic power generation systems are as follows: In the above formula, P w_min and P pv_min are the minimum output power of wind turbine and photovoltaic system respectively; P w_max and P pv_max are the maximum output power of wind turbine and photovoltaic system respectively; The real-time SOC and charge and discharge power constraints of the energy storage system are as follows: In the above formula, SOC min and SOC max are the lower and upper limits of SOC, with values ​​of 0.1 and 0.9 respectively; P es_min and P es_max Represent the lower and upper thresholds of the charging and discharging power of the energy storage system respectively; In the case of grid connection, the constraints on electricity purchase and sales are as follows: In the above formula, I grid,t A binary variable representing the power exchange state, I grid,t =1 means the energy system purchases electricity from the grid, I grid,t =0 means the energy system sells electricity to the grid.

Images